41171 Social Health Insurance for Developing Nations Edited by With contributions by William C. Hsiao Andrew Fraker Piya Hanvoravongchai R. Paul Shaw Matthew Jowett Diana Pinto Sreekanth Ramachandra THE WORLD BANK WBI DEVELOPMENT STUDIES WBI DEVELOPMENT STUDIES Social Health Insurance for Developing Nations William C. Hsiao Harvard University School of Public Health and R. Paul Shaw World Bank Institute With contributions by Andrew Fraker Piya Hanvoravongchai Matthew Jowett Diana Pinto Sreekanth Ramachandra Jointly sponsored by the World Bank Institute and Harvard University The World Bank Washington, D. C. © 2007 The International Bank for Reconstruction and Development / The World Bank 1818 H Street, NW Washington, DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org E-mail: feedback@worldbank.org All rights reserved 1 2 3 4 10 09 08 07 This volume is a product of the staff of the International Bank for Reconstruction and Development / The World Bank. The findings, interpretations, and conclusions expressed in this volume do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgement on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development / The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; telephone: 978-750-8400; fax: 978-750-4470; Internet: www.copyright.com. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street, NW, Washington, DC 20433, USA; fax: 202-522-2422; e-mail: pubrights@worldbank.org. ISBN-10: 0-8213-6949-0 ISBN-13: 978-0-8213-6949-4 eISBN: 0821369504 e-ISBN-13: 978-0-8213-6950-0 DOI: 10.1596/ 978-0-8213-6949-4 Library of Congress Cataloging-in-Publication Data has been applied for. Contents Foreword vii Acknowledgments ix Abbreviations xi 1. Introduction, Context, and Theory 1 William C. Hsiao and R. Paul Shaw Context 4 Theory and Principles 11 References 18 2. Design and Implementation of Social Health Insurance 21 William C. Hsiao Design 21 Implementation 33 Summary 39 References 40 3. Kenya: Designing Social Health Insurance 43 Andrew Fraker and William C. Hsiao Background 43 Design of the NSHIF 47 Implementation Issues 52 Conclusion 59 References 59 4. Ghana: Initiating Social Health Insurance 61 Sreekanth Ramachandra and William C. Hsiao Background 61 Evolution of Health Financing Mechanisms 65 Design Issues 71 Implementation Issues 75 Successes and Challenges 77 References 79 5. The Philippines: Extending Coverage beyond the Formal Sector 81 Matthew Jowett and William C. Hsiao Background 81 Delivery of Health Services 86 Design and Implementation Issues 90 Impact of SHI 103 References 104 iii iv Contents 6. Colombia: Social Health Insurance with Managed Competition to Improve Health Care Delivery 105 Diana Pinto and William C. Hsiao Background 105 Health Care Reform 108 Design 110 Implementation 113 Achievements 124 Challenges 124 References 131 7. Thailand: Achieving Universal Coverage with Social Health Insurance 133 Piya Hanvoravongchai and William C. Hsiao Background 133 Health Insurance Development before Universal Coverage 140 Achieving Universal Coverage 146 Lessons Learned from Universal Coverage 151 Challenges 153 Conclusion 154 References 154 8. Lessons Learned and Policy Implications 155 William C. Hsiao and R. Paul Shaw The Development Context Matters 158 The Five Country Case Studies 159 Positive Changes Attributable to SHI 159 Major Challenges Inherent in Implementation 162 Policy Implications 165 Conclusions 170 References 171 Boxes 1.1. Terms 17 8.1. Traditional Responses to Poor Health Sector Performance 157 Figures 1.1. Sources of Health Care Financing, Selected Low- and Middle-Income Nations, 2000 6 1.2. Out-of-Pocket Payments as a Percentage of Total National Health Expenditures, 2002 7 1.3. Average Tax Revenues by Country Income Level, 2000 8 1.4. Share of Total National Health Spending Derived from External Sources, Selected Countries, 2000 10 2.1. Progress Toward Universality, Selected Countries, 1990­2014 37 2.2. Hypothetical Plan to Reduce Subsidies to Public Hospitals 39 3.1. Map of Kenya 44 3.2. Administrative Structure of the NSHIF 50 4.1. Map of Ghana 62 4.2. Interaction between Administrative Entities 72 Contents v 5.1. Map of the Philippines 82 5.2. Composition of Health Expenditures, 1998­2004 85 5.3. PhilHealth's Financial Position, 1997­2004 91 5.4. Contributions by Government Employees and Government Employers, 2000­2005 92 5.5. Enrollment in the NHIP, 1972­2005 93 5.6. Changes in Membership Composition, 2003­2005 96 5.7. Estimated Coverage Rates by Member Sector, 2004 96 5.8. Cross-Subsidization across Member Categories, 2004 101 5.9. Trends in Claims by Member Category, 2002­2005 102 5.10. Average Value of Claims by Member Category, 2002­2005 102 6.1. Map of Colombia 106 6.2. Managed Competition Model for National Health Insurance 109 6.3. Administrative Structure of the Health Care System 111 6.4. Toward Universal Health Coverage, 1993­2001 113 6.5. Expansion of Insurance Coverage by Regime, 1992 and 1996­2003 114 6.6. Insurance Coverage by Urban and Rural Residence and Regime, 2003 114 6.7. Insurance Coverage by Income Quintiles, 1992­2003 115 6.8. Reasons for Not Using Health Care by Income Quintile, 1992 and 1997 124 6.9. Percentage of Respondents Who Had a Health Problem Who Were Seen by a Doctor by Enrollment Status and Area, 1997 and 2003 125 6.10. Public Hospitals' Incomes and Expenditures, 1993­99 127 7.1. Map of Thailand 134 7.2. Poverty Trend, 1988­2004 136 7.3. Health Expenditure as a Percentage of GDP, 1994­2004 137 7.4. Share of National Spending on Health by Source, 2001 138 7.5. Share of National Spending on Health by Spending Category, 2001 138 7.6. Development of Health Insurance Schemes, 1975­2002 143 7.7. Trends in Coverage by Health Insurance Schemes, Selected Years 146 8.1. Percentage of Total Public Expenditures on Health by Income Quintile Ghana, 1992 and 1998 156 Tables 1.1. Government and External Donors' Expenditures on Health, Selected Sub-Saharan African Countries, 1995 and 2000 9 1.2. Factors Contributing Positively to an Enabling Environment for SHI 12 1.3. Country Characteristics, Selected Countries, 2000­2004 13 1.4. Selected Characteristics of SHI Compared with Other Forms of Health Insurance 16 2.1. Working Population by Employment Status and Income, Uganda, 1999­2000 23 2.2. Design Issues, Selected Countries, as of 2005 28 3.1. Background Statistics, Kenya 45 3.2. Shares of Health Expenditures by Provider, 2000­2002 46 3.3. The NSHIF Gradual Implementation Scenario, 2004­12 47 3.4. Projected NSHIF Expenditures and Revenues, 2004­10 49 3.5. Sources of Total Health Expenditures, Actual in 2001 and Projected under the NSHIF 54 3.6. Sources of NSHIF Funding 54 vi Contents 3.7. Implementation Scenarios, 2004­12 57 3.8. Average Annual Contributions by Public and Private Sector Employees 58 3.9. Required Government Contribution to Balance the Fund, Selected Years 58 4.1. Background Statistics, Ghana 63 4.2. Chronological Development of Health Financing 66 4.3. Distribution of Membership, Four Largest MHOs, 2002 68 4.4. Employment by Category, Selected Years 77 5.1. Background Statistics, Philippines 83 5.2. Milestones in the Development of the NHIP 85 5.3. Department of Health Licensed Hospital Facilities by Ownership and Level, 2003 87 5.4. Hospital Bed Distribution in Department of Health Licensed Hospitals by Ownership and Level, 2003 87 5.5. PhilHealth Accredited Hospital Facilities by Ownership and Level, 2005 88 5.6. Hospital Bed Distribution in PhilHealth Accredited Hospitals by Ownership and Level, 2005 88 5.7. Facility Accreditation for New PhilHealth Benefit Packages, December 2004 89 5.8. Health Professionals Registered by the Department of Health, 2004 90 5.9. PhilHealth Financial Statement, March 31, 2004, and March 31, 2005 92 5.10 PhilHealth Member Benefits and Cost Ceilings, 1999 99 6.1. Background Statistics, Colombia 107 6.2. Scenarios for Expanding Coverage by the Subsidized Regime as of December 2004 116 6.3. Comparison of the Contributory and Subsidized Regime Benefits Packages as of December 2004 117 6.4. Payment Method Used by Health Plans in the Contributory Regime 121 6.5. Solidarity and Guarantees Fund Balances, 1998­2002 122 6.6. Total Resources and Sources of Funding for the Subsidized Regime, 1995­2002 123 6.7. Sources of Public Hospital's Sales Revenues by Level of Care, 2003 128 6.8. Ratio of Total Health Expenditures to Total Income from Premiums 2003 and 2004 130 7.1. Background Statistics, Thailand 135 7.2. Numbers and Use of Public and Private Health Facilities 2003 139 7.3. The Four Main SHI Schemes before Universal Coverage, 1999 140 7.4. Insurance Schemes and Utilization of Services, 1996 145 7.5. The Three SHI Schemes after Universal Coverage, 2004 149 7.6. Breakdown of the Universal Coverage Scheme Capitation Rate and Payment Methods, 2004 152 8.1. Selected Characteristics of Country Groupings by Income Level, 2000 159 Foreword In dozens of developing countries, special technical groups are busy advising min- isters of health, ministers of finance, vice presidents, and presidents on the fea- sibility of social health insurance (SHI) as a way to mobilize revenues for health, enhance the health sector's performance, and provide universal coverage. Yet evi- dence and guides on "how-to" design and implement SHI have been largely based on the relatively rich, developed countries. Such countries' experience may not be good guides for low-income developing countries because they differ sharply in terms of fiscal capacity, per capita expenditures on health, size of their formal sec- tors, dependency ratios, administrative capability, and diversification of provider markets. In contrast, evidence on the design and implementation of SHI in develop- ing countries is hard to come by, which is precisely the rationale for this book. The volume includes a review of design and implementation issues that chal- lenge SHI in low- and middle-income countries and case studies on Colombia, Ghana, Kenya, the Philippines, and Thailand that shed light on the trials and tribu- lations of implementing SHI in contexts far less hospitable than those in rich coun- tries. Accordingly, the case studies provide a road map of design options, aims and intentions, midcourse revisions, and successes and pitfalls. The volume concludes by presenting lessons learned and policy implications. Perhaps the most important message of this volume is that SHI should not be seen as a magic bullet that will solve all the woes of health care financing and pro- vision in developing countries. It clearly has the potential to make a positive con- tribution, but success comes slowly, because of the drawbacks and risks involved. Our hope is that by contributing to awareness of these issues, this volume will help policy makers in developing nations chose the right approaches, and then, once chosen, to implement them correctly. Frannie A. Léautier Vice President and Head World Bank Institute vii Acknowledgments This volume evolved as part of the research agenda in support of the Flagship Learning Program on Health Sector Reform and Sustainable Financing developed jointly by the World Bank Institute and the Harvard School of Public Health. The authors are grateful to the Irish Trust Fund given to the World Bank Institute for financial support and for contributions by collaborating authors on the five country case studies in this monograph. ix Abbreviations CEO chief executive officer CSMBS civil servants medical benefits scheme (Thailand) DOTS directly observed treatment short-course GDP gross domestic product IPP Individual Payment Program (Philippines) LGU local government unit MHO mutual health organization (Ghana) NGO nongovernmental organization NHIA National Health Insurance Act (Ghana) NHIF National Hospital Insurance Fund (Kenya) NHIP National Health Insurance Program (Philippines) NSHIF National Social Health Insurance Fund (Kenya) NHSO National Health Security Office (Thailand) PhilHealth Philippine Health Insurance Corporation PRO PhilHealth Regional Offices PPP purchasing power parity SGF Solidarity and Guarantees Fund (Colombia) SHI social health insurance SISBEN sistema de identificación de beneficiarios (system for identifying beneficiaries) (Colombia) SRA subsidized regime administrator (Colombia) SSS social security scheme (Thailand) VHCS voluntary health card scheme (Thailand) WHO World Health Organization xi 1 Introduction, Context, and Theory William C. Hsiao and R. Paul Shaw Good health is necessary for well-being. Even billionaires cannot enjoy life when their health is poor. Good health is also required for economic and social devel- opment (WHO 2000). Workers have to be healthy to work and children have to be healthy to attend school and partake in other activities, yet most developing nations lag far behind what they could achieve in terms of their populations' health. At the same time, poor health has another critical impact: it causes poverty, in that large health expenditures can bankrupt families. Studies show that health expenditures are a primary cause of impoverishment (Fu 1999). In developing nations, the principle causes of poor health are inadequate pre- vention and lack of reasonable access to basic health care, along with poor nutrition and unclean water, whereas health-related impoverishment results from a lack of risk pooling and insurance (WHO 2000). Underfunding of health care is central to both of these negative outcomes. Moreover, many countries compound these prob- lems by making inefficient use of the resources they do have for health care and risk pooling. Most low- and middle-income nations face difficulties in funding health care. While nations declare similar admirable goals to provide their citizens with equal access to a reasonable quality of health care and to prevent health-caused impov- erishment, the reality is starkly different. These exalted aims are not backed up with adequate public funds or a rational financing system. As a result, poor health, impoverishment, and disparity are prevalent. Many nations are now hoping that formally mandated social health insurance (SHI), involving payroll taxes, will pro- vide a solution. A wave of SHI initiatives has swept across Africa, Asia, and Latin America. In May 2005, the World Health Assembly passed a policy resolution for the World Health Organization (WHO) whereby WHO would use SHI as the strategy for mobilizing more resources for health, pooling risk, providing more equitable access to health care for the poor, and delivering better quality health care (WHO 2005a). The WHO is encouraging its member states to move ahead with SHI and will pro- vide technical support to help nations develop SHI. In addition to its capacity to mobilize additional funds for health care, SHI is also being touted by several international aid agencies, including the World Bank, 1 2 Introduction, Context, and Theory the WHO, and the German Agency for Technical Cooperation, as a policy instru- ment that could help facilitate or stimulate four desirable elements of health sector reform, namely: · when low-income nations do not have adequate tax revenues to fund health care of a reasonable quality for everyone, SHI targets public funds to subsi- dize premiums for the poor rather than financing and providing universal health care for all; · freeing up public funds so they can be targeted to public health goods and services; · shifting public subsidies from the supply side to the demand side to improve the efficiency and quality of health care. This separates the responsibilities for collecting and managing SHI financing from the responsibilities for pro- viding health care to patients, whereby services are contracted from pro- viders that are separate entities. Providers are required to be accountable to patients for the quality of services; · using the capacity of nongovernmental organizations (NGOs) and private providers to improve access by the insured to health care by means of contracting. This monograph examines the principles, design, and practices of SHI for low- and middle-income nations and the necessary conditions for its viability and sus- tainability, with a focus on design and implementation issues. In relation to our recommendations on planning and implementing SHI, we rely on a simple theory of insurance, on logic, and on experience from around the world. A nation must overcome many barriers to establish successful, universal SHI, and the intent of this monograph is to help readers understand these barriers and design measures to overcome them. This volume presents five country case studies to provide evidence and greater detail on key issues that arise at different stages of implementation in low-income countries. They have been selected to reflect on a continuum and timeline of oper- ational stages, beginning with the initial design and legislation of SHI, the first phase of implementation, the expansion to cover larger segments of the population, and on up to completion, whereby SHI becomes the predominant form of health care financing in a country. Accordingly, Kenya has been selected for illustration of the design stage, Ghana for initiation, the Philippines for extension of population coverage, Colombia for SHI and reform of health care delivery, and Thailand for universal coverage and reform of health care delivery. This sequencing and implied timeline of case studies allows us to reflect on two questions. First, where can a country expect to be in relation to designing and implementing SHI in, say, 10 years? Second, as countries gain experience with SHI, what can they expect to offer or achieve in terms of variations in benefit design, who administers SHI, and how providers are contracted and paid? In addition to providing an informative road map of the challenges facing SHI in countries at dif- ferent stages of development, the country case studies also illustrate the immense changes SHI can bring to a country's health system in terms of financing sources and agents, payment of public and private providers, organization and regulation of health care, behavioral change among both providers and patients, and rede- fined roles of government. William C. Hsiao and R. Paul Shaw 3 The Kenya case illustrates a country at the design and legislation stage. The country has passed a law to establish SHI motivated in part by the government's desire to overcome the organizational deficiencies of its current Hospital Insur- ance Fund. Various financing and membership scenarios have been worked out, all underpinned by good intentions; however, political consultation during passage of the SHI law was inadequate, meaning that stakeholders' involvement and support has been highly uneven. Moreover, the president has yet to sign implementation of SHI into law, because closer analysis of the costs, SHI premiums, and related tax implications suggested that the proposed expansion path of SHI would be finan- cially unsustainable. In addition, international aid agencies and donors, such as WHO, the German Agency for Technical Cooperation, and the World Bank, have been cautious about Kenya's capacity to sustain its proposed SHI plan. Unless the problems can be overcome, the country may not realize the good intentions associ- ated with the design of SHI. We selected Ghana as a case study, because legislation was signed into law during 2003­4 and implementation has begun. When it launched its initial implementation plan and budget, the government aimed to cover up to 25 percent of the population. To this end, SHI has adopted an ingenious approach by incorporating existing com- munity financing schemes. However, whether Ghana can afford its generous mini- mum benefits package and how effective the district-level, community-based health insurance schemes are in enrolling eligible people remains to be seen. Another important feature of SHI in Ghana is the decision to give considerable manage- ment authority to districts and to allow eligible people to enroll in either a publicly managed district health insurance plan or a private insurance plan. This reflects the government's commitment to emphasize a district-based approach to health system development, rather than a top-down, centralized approach. We selected the Philippines, because SHI is in the middle stage of development, with 10 years of implementation. The Philippines' SHI agency, the Philippines Health Insurance Corporation (PhilHealth), a quasi-government agency that is one step removed from government ownership and control, manages SHI. PhilHealth is a single payer, acting mostly as a traditional health insurer, paying claims rather than protecting the insured from price gouging. Coverage reached 81 percent in 2004, with the government providing large subsidies to increase enrollment by the poor during the election year. Since then the government had reduced subsidies for the indigent, and enrollment has fallen to 63 percent. A major problem PhilHealth faces is a high evasion rate of as much as 70 percent among small employers, along with major barriers to enrolling nonpoor farmers and informal sector workers. In addition, providers bill patients for the balance of what PhilHealth does not pay. As a result, the insured are not receiving the risk protection intended by SHI. More- over, one study found that payments by PhilHealth to private providers tended to generate profits for the private providers rather than improve health services (Gertler and Solon 2002). We chose Colombia because it is in the middle stages of SHI, with 12 years of implementation experience. It is well known for its decision to use a solidarity fund to cross-subsidize membership by the poor, whereby a portion of contributions for SHI from urban and formal sector workers, who are relatively well-off, are used to cross-subsidize contributions by rural and informal workers. Today, 70 percent of Colombians are covered by SHI, although the benefit packages for the poor and 4 Introduction, Context, and Theory near-poor are substantially less comprehensive than the package for employed workers. Colombia also yields important insights for other countries in relation to its efforts to reform its health care delivery system by imbedded managed compe- tition in its SHI. On the one hand, managed competition does not appear to have yielded the expected results regarding improved efficiency and quality of health service delivery, and on the other hand, adverse selection has also emerged as a serious problem. We selected Thailand to illustrate how a desirable configuration of influences can be mobilized to create the political force to yield almost universal coverage in seemingly record time in a middle-income country. This desirable configura- tion included almost 20 years' experience with SHI for formal sector and civil ser- vice workers, more than 10 years' experience with voluntary risk pooling for rural households, and accumulated financial surpluses in the SHI fund. The political influences centered on the prime minister, a billionaire without strong popular support in the countryside, who decided to run for election on a platform of prom- ising universal insurance for rural and farm households. Thus following his elec- tion in 2001, Thailand attained universal coverage with its SHI by committing gen- eral tax revenues to pay the premiums for all the poor, near-poor, self-employed, and informal sector workers. Today approximately 75 percent of the population have their premiums paid for this way. Moreover, SHI acts as an active and pru- dent purchaser by paying providers based on a capitation method,1 which has improved the efficiency and quality of health service delivery. A remaining chal- lenge is that civil servants and formal sector workers have more generous benefits packages than others. Whether the new government and new prime minister will continue to commit enough funds to sustain universal coverage and/or equalize benefits remains to be seen. Context SHI needs to be understood in context. SHI is a financing approach for mobilizing funds and pooling risks. The newly mobilized funds should be allocated for the poor and near-poor to improve their financial access to health care. SHI may be a solution for a critical part of a nation's systemic health care problem, but is not nec- essarily a solution for the whole problem. Poor health prevails in many developing nations. The average infant mortal- ity rate in many African countries still exceeds 100 per 1,000 live births, compared with 4 per 1,000 live births in advanced economies. Asian nations usually do better, but the results are still poor: China has an infant mortality rate of 30 per 1,000 live births, and the figure stands at 31 per 1,000 in Indonesia and 62 per 1,000 in India (World Bank 2005). However, large disparities in health status persist between the poorest and richest in most countries. For example, the infant mortality rate is 6 per 1,000 live births in urban China, but 56 per 1,000 in poor rural regions of China and immunization rates are still below 80 percent for many countries such as Nigeria and Uganda. Obviously we can do better. 1. A fixed rate of payment per person for services delivered over a fixed period. William C. Hsiao and R. Paul Shaw 5 What explains these unsatisfactory outcomes in developing countries? In addi- tion to underfunding for health, studies have found at least four other reasons, namely: · Poorly targeted public resources tend to favor the rich. · Many countries are unable to manage their public health services efficiently and effectively. In other words, they are unable to transform money into effi- cient and good quality health services. · Public sector primary care services do not match rural people's demand in terms of their location and organization. · Health risks are not appropriately pooled, thus the poor, the low income, the elderly, and the less healthy are excluded from insurance. In relation to many countries' inability to manage their public health services efficiently and effectively, public facilities operate under bureaucratic rules and managers have little power to make financial and personnel decisions. Often oper- ational funds do not reach public facilities on a timely basis, resulting in low staff productivity and facilities that regularly run out of drugs and supplies (Foster 1993; Mills 1995; World Bank 1993). Moreover, even when facilities are built and staffed and funds are spent, they are not located where everyone has access to them and/or do not provide the services that people demand and value. As a result, these facilities are underused (Bitran 1995; Gilson 1995; Zere, Mcintyre, and Addison 2001). In rural areas, qualified prac- titioners often do not want to work at the subdistrict or village level, assuming that public services are even available at this level. Frequently, physicians simply do not show up or do not show up regularly and/or provide poor customer service. How- ever, people want primary care within an hour by foot (Diop, Yazbeck, and Bitran 1995; Hjortsberg and Mwikisa 2002; Liu, Rao, and Hu 2002). This means that ser- vices for rural residents have to be at the village level, but governments persistently establish primary care centers at the subdistrict level, which is far away for most rural residents. As a result, when villagers become sick, most of them use their meager incomes to pay local traditional healers, private practitioners, and drug peddlers. When villagers become seriously ill, they go to public and charity hospi- tals, resulting in overcrowding. Studies document that most developing countries have allocated their govern- ment resources for health to public hospitals in urban areas (WHO 2000; World Bank 1993). These public facilities, especially tertiary-level hospital services, are used mostly by more affluent urban residents. As a result, public funds have been spent disproportionately for the rich (Castro-Leal and others 1999). Another cause of the access barriers involves risk pooling. Most developing countries do not rely on the insurance mechanism to pool health risks. When they do, the risks are pooled only for civil servants, and perhaps for workers in the for- mal sector. These people are employed and tend to be more affluent than other seg- ments of society, and the poor and the less healthy are unable to benefit from these insurance pools (Dror and Jacquier 1999). Mobilizing additional funding for health care only offers a partial remedy to the systemic health problem. When a country embarks on an SHI strategy, there is no assurance that more and better health care will be delivered. The government must also plan how the funds can be transformed into effective services. Each country 6 Introduction, Context, and Theory differs in terms of organization of public and private providers, payment systems, regulation, and conditions affecting the disease burden. For example, India seems to spend close to a reasonable amount on health, that is, 6 percent of gross domestic product (GDP), but its health system is unable to provide effective services for the poor rural population (Peters and others 2002). In contrast, Sri Lanka spends a mod- est amount, 3.7 percent of GDP, and has produced enviable results in health status and risk protection (WHO 2005b). Thus we have more confidence that additional public funding by Sri Lanka could yield significant gains, while we cannot say the same about India unless it adopts companion reforms. Sources of Funds for Health Care Developed countries often use SHI to mobilize funds and pool risks, but low- and middle-income countries rarely use this approach and rely mostly on general revenues and direct out-of-pocket payments as sources of health care financing. Figure 1.1 shows the sources of financing for selected developing countries. It illustrates that general revenues and out-of-pocket payments are the dominant sources of financing. SHI plays only a minor role in low-income nations, but as their national incomes grow, the share of health care financed by SHI increases. Figure 1.1. Sources of Health Care Financing, Selected Low- and Middle-Income Nations, 2000 100 80 60 financing of 40 percentage 20 0 Mali Kenya China India Rica Peru Tanzania Zambia Malawi Nigeria Uganda Thailand MalaysiaIndonesia VietnamCambodiaColombia Bolivia Philippines Costa NicaraguaSalvador Ecuador Guatemala El Mozambique out-of-pocket other private sources general revenues private insurance social insurance Source: WHO 2002b (Statistical Annex). William C. Hsiao and R. Paul Shaw 7 Until recently, many countries only paid attention to the amount the govern- ment spent for health care, failing to recognize that perhaps public spending accounted for only a modest part of the whole. Patients' out-of-pocket payments represent a significant portion of health care spending, and these payments are inequitable, placing a greater financial burden on the less healthy and on poor peo- ple and deterring patients from seeking necessary health care. Equally important, individual out-of-pocket payments do not pool risks. Figure 1.2 shows the extent of out-of-pocket payments in selected countries in Africa, Asia, and Latin America. While out-of-pocket payments as a share of total national health expenditures are generally larger in poor rather than in rich coun- tries, wide variations are apparent among low-income countries. In countries of Sub-Saharan Africa, where average per capita income was US$546 in 2002­4, the data in figure 1.2 suggest that out-of-pocket payments account for, on average, about 41 percent of total national health expenditures. In South Asia, where per capita income was US$521 in 2002­4, the share of out-of-pocket payments averages about 55 percent. In Latin America, with an average per capita income of US$3,631 in 2002­4, the out-of-pocket share averages about 44 percent. Moreover, in low- income countries, out-of-pocket payments represent a larger share of total house- hold health expenditures among the poorest households than among better-off households. These figures imply that almost half of total national health expendi- tures in low-income countries exchange hands in a highly disorganized fashion, Figure 1.2. Out-of-Pocket Payments as a Percentage of Total National Health Expenditures, 2002 80 77.5 70.7 70 67.3 63.9 62.0 60 59.0 56.6 53.1 51.9 expenditure 50 48.7 47.5 48.8 45.3 43.6 42.8 health 39.8 40 37.7 37.3 35.5 34.3 32.7 total of 30 25.1 23.0 20 10.6 percentage 10 0 PeruRica Nigeria Ghana KenyaMali India Uganda Zambia Tanzania Malawi China Cambodia Vietnam IndonesiaMalaysia Philippines Thailand EcuadorNicaragua Bolivia Salvador El GuatemalaCosta Mozambique Source: WHO 2002b (Statistical Annex). 8 Introduction, Context, and Theory often when the poorest people are the most in need of health care, yet the least able to pay for it. Underfunding and Government Capacity to Spend More The literature has extensively documented that health spending as a share of GDP varies significantly between developing countries (see, for example, WHO 2002a). Nonetheless, most developing countries are underfunding health, that is, the level of resources spent is measurably below the level needed. One obvious cause for underfunding is that governments allocate too few dollars to health. However, the amount of tax funds a nation can spend on health is limited by the government's ability to collect tax revenues. Another consideration involves the nation's political economy, which influences how much of the budget is allocated to health services and how much to competing priorities. SMALL AND NARROW TAX BASE. Most developing countries have a narrow tax base because of their small industrial bases and large shadow economies charac- terized by labor working in the informal sector. Collecting taxes from the shadow economy is extremely difficult. Moreover, most developing nations do not have the infrastructure and administrative capacity needed to collect taxes effectively. These factors limit countries' ability to generate large amounts of tax revenue. As a nation develops economically, its tax base increases, because both its shadow economy shrinks and its administrative capacity improves. Tax revenues rise as a result, and the country has more resources to support a larger public sec- tor. Figure 1.3 compares the size of the public sector that is facilitated by tax rev- enues for countries at different income levels. Figure 1.3. Average Tax Revenues by Country Income Level, 2000 40 GDP) countries 30.9 of 30 all of centage 22.3 19.4 evenuesr (per 20 tax oup 14.0 gr total 10 income in average 0 low-income lower-middle- upper-middle- high-income countries income income countries (US$9,360) (US$760­ (US$3,031­ US$3,030) US$9,360) Source: IMF 2002. Note: Figures in parentheses indicate GDP per capita. William C. Hsiao and R. Paul Shaw 9 BUDGETALLOCATION. ManyinternationalhealthexpertsandNGOshaveargued that governments should reallocate their budgets and spend more on health, but health has to compete with other programs for government resources. The politi- cal economy of a nation determines the share of general revenues to be spent on health. Most developing nations allocate only 6 to 10 percent of their government budgets to health (World Bank 2005). More important, governments have not allo- cated adequate public funding for basic health care for the poor,2 who, along with low-income people, usually have little political voice and influence on government decisions. While many African nations have signed the Abuja Declaration whereby they pledge to allocate 15 percent of their government budgets to health, few nations have actually achieved this goal. Table 1.1 compares the shares of government bud- gets allocated to health and the shares of health budgets that come from interna- tional donors in selected African countries. Between 1995 and 2000, the unweighted average of government expenditures on health as a percentage of total government expenditures grew by only 7.8 percent. In contrast, external resources for health as a percentage of government expenditures on health grew by almost 40 percent. DONOR DEPENDENCY. As figure 1.4 shows, the share of total national health expenditures that comes from donors is high in many Sub-Saharan African nations. These funds are channeled through the government and can account for more than 25 percent of the public health budget. Donor financing raises two issues. First, to what extent are donors driving a nation's health priorities and are donors' priorities Table 1.1. Government and External Donors' Expenditures on Health, Selected Sub-Saharan African Countries, 1995 and 2000 Government expenditures on External resources for health health as a percentage of total as a percentage of government government expenditures expenditures on health Country 1995 2000 1995 2000 Eritrea 4.1 4.0 32.8 60.7 Ethiopia 5.8 5.5 17.9 35.6 Ghana 8.3 7.9 15.8 24.1 Kenya 6.6 8.1 32.3 38.3 Lesotho 9.6 10.8 5.5 6.8 Malawi 11.3 14.6 50.2 86.7 Nigeria 1.7 3.0 33.4 32.2 Tanzania -- 8.1 -- 50.0 Uganda 9.2 9.5 84.0 96.0 Zambia 11.5 11.2 22.0 24.8 Unweighted average 7.6 8.2 32.6 45.5 2. The industrial nations (other than the United States) use general revenues or com- pulsory social insurance to pay for and provide health care for their citizens working in the nonformal sector. 10 Introduction, Context, and Theory Figure 1.4. Share of Total National Health Spending Derived from External Sources, Selected Countries, 2000 40 39.3 37.6 se 35 30 28.8 26.9 expenditur 25 health 20 18.6 18.5 18.2 total of 15 10 9.8 9.3 centage 7.0 6.1 per 4.9 5 4.6 4.4 2.8 1.8 1.8 1.0 1.30.9 0.2 0.1 0 0.7 0 0 u Per Rica MalawiTanzania Uganda Zambia GhanaMaliKenya Nigeria CambodiaVietnam India China Bolivia PhilippinesIndonesiaThailandMalaysia Nicaragua Guatemala CostaEcuadorColombia Salvador El Mozambique Source: WHO 2002b (Statistical Annex). compatible with domestic priorities? Second, how stable is donor financing, and will health programs be sustainable if donors withdraw their support? Prevailing System of Financing and Health Care Provision in Developing Nations SHI can mobilize additional funds for health care, but more funds do not neces- sarily mean more and better health care. Companion reforms must take place if the additional funds are to be transformed into effective and efficacious services. To assess what companion reforms are necessary, a nation has to know where its national health system stands. Only then can the government plan where to go and how to get there. In this subsection we describe the common characteristics of health care systems in most developing countries to illustrate where the starting place is likely to be. Public health, prevention, maternal and child health services, and HIV/AIDS programs usually receive inadequate priority and funding from domestic govern- ments (WHO 2005b). When international donors give priority to these programs and support them with funds, most governments have to establish vertical pro- grams to deliver the specified services in addition to the national programs they manage. Each vertical program creates its own bureaucracy, clinics, and supply systems. The programs often overlap with each other, competing for the limited number of trained health personnel and fighting over the few available vehicles and other resources. In addition to impairing other preventive and primary care William C. Hsiao and R. Paul Shaw 11 services, the sustainability of these vertical programs is a serious concern because of their dependence on unstable external funding. Beyond preventive and special vertical programs, the government takes respon- sibility, at least on paper, for organizing, managing, and delivering primary, sec- ondary, and tertiary curative services to all citizens. However, governments in developing countries can seldom fund these services adequately and rarely have the capability to deliver these services efficiently. Consequently, the financing and provision of curative health services is segmented into three tiers according to patients' ability to pay (Stiglitz 1999). In the first tier, affluent households demand high-quality services, which are provided mostly by private sector physicians and small private hospitals that charge high fees. The affluent patients pay directly out-of-pocket for these services; however, they obtain tertiary (highly specialized and expensive) services from public teaching hospitals, because such services require large capital investments that the private sector usually does not make. In the second tier, middle-income households finance and obtain their health services differently from the affluent. Many developing nations have insurance programs for civil servants, and some large employers, such as banks, also cover their employees under insurance programs. These insurance programs, some of which may be social insurance, select and contract with the better public and pri- vate facilities to provide services. These facilities generally charge the insurance plan on a fee-for-service basis.3 If a developing country has adopted SHI, the SHI plan often operates its own clinics and medical facilities that offer higher-quality services than the public facilities. In the third tier, public clinics and hospitals serve the vast majority of the pop- ulation: poor and low-income households. In Africa, charity hospitals also play an important role in serving this population, charging them on a fee-for-service basis, but reducing the fees for the poor. Although public health services are almost free, waiting lines are often long, health centers are located far away from patients, physicians may not be on duty, clinic hours may be inconvenient, facilities may be dilapidated and crowded, drugs and other supplies may not be available, and pro- viders tend to be unfriendly. Because of these conditions, when people suffer from illnesses that are not life-threatening, they often resort to self-care or seek care from indigenous practitioners. Household expenditure surveys consistently find that poor and low-income households spend a significant portion of their incomes on drugs and indigenous medicine (see, for example, Ha, Berman, and Larsen 2002). Only when they are seriously ill, requiring hospital care, are they compelled to rely on public hospitals. When a nation starts out with a three tiered system where people have access to health services of vastly different quality, the critical issues are (a) how to reor- ganize the segregated public and private health care delivery systems under SHI, (b) how to integrate vertical programs into the general health system, (c) how to use resources and modern management to improve the efficiency and quality of preventive and medical services, and (d) how to assure that all those insured have equal access to similar quality health services. 3. A fee-for-service basis refers to a payment per visit or activity. 12 Introduction, Context, and Theory Theory and Principles SHI has been developing for more than a century following its establishment in Germany by Bismarck in 1883 (Saltman and Dubois 2004). Worldwide, so far 27 countries have established the principle of universal coverage via SHI (Carrin and James 2005). This process took 127 years to achieve in Germany, 118 in Belgium, 79 in Austria, 72 in Luxembourg, 48 in Costa Rica, 36 in Japan, and 26 in the Republic of Korea. Carrin and James (2005) show that it took 40 years in Austria (from 1890 to 1930) for population coverage to grow from 7 to 60 percent, and then another 35 years (from 1930 to 1965) to reach 96 percent. More relevant to the countries discussed in this vol- ume, it took 20 years for SHI to reach population coverage of 17 percent in Costa Rica (from 1941 to 1961), another 5 years to double coverage to 34 percent (1966), another 12 years to again double coverage to 74 percent (1978), and then another 13 years to attain 83 percent coverage (1991). Above all, a country's level of economic development and its economic struc- ture influence how many people can be covered and how rapidly SHI can expand toward universal coverage. Among richer nations, for example, several structural features of the economy, as illustrated in table 1.2, tend to contribute positively to the enabling environment for SHI. Conversely, low- and middle-income countries differ considerably on various criteria known to be important to the enabling environment for SHI. Based on the information in table 1.3, we would therefore expect that a low per capita income, a small formal sector (often proxied by a small urban sector), a high prevalence of Table 1.2. Factors Contributing Positively to an Enabling Environment for SHI Structural feature Contribution Large formal sector employment · Ease of administering mandated payroll tax on employers and/or employees · Ease of locating employers and collecting premiums High wages and salaries · Reduced economic burden of payroll tax · Opportunity to finance broader benefit entitlements Low poverty rate · Reduced need to subsidize membership of poor households Small family and/or household size · Reduced need for worker contributions to cover large number of dependents Efficiently functioning provider networks · Improved access by members to providers · Greater choice of providers · Possibility of quality-based competition among providers Strong human resource capacities · Available skills to manage SHI and monitor and evaluate quality Strong administrative support · Banking, accounting, actuarial, and legal support available Government capacity to regulate · Greater capacity to regulate for quality and manage grievance procedures Table 1.3. Country Characteristics, Selected Countries, 2000­2004 Percentage of the Percentage of health Population Per capita gross population living expenditures Country SHI Stage (millions) national product (US$)Poverty rate (%) in urban areas out-of-pocket Dependency ratio 13 Kenya Design 31 360 50 41 56 0.80 Ghana Initiation 20 270 39 46 59 0.90 Philippines Extension 80 1,020 37 62 61 0.70 Colombia SHI + managed care 44 1,830 50 77 17 0.60 Thailand Universal coverage 62 2,000 17 32 30 0.40 Source: United Nations statistical databases (http://unstats.un.org/unsd/demographic/default.htm); World Bank statistical databases (http://devdata.worldbank.org/hnpstats/). Note: The figures shown are rough averages. 14 Introduction, Context, and Theory poverty, and a high dependency ratio would represent immense challenges to the initiation and scaling up of SHI in Ghana and Kenya. In contrast, we would expect countries with higher per capita incomes, larger formal sectors, less poverty, and lower dependency ratios, such as Colombia and Thailand, to offer a more receptive environment for initiating and expanding SHI. No textbook model is available for the design, implementation, and expansion path for SHI to achieve universal coverage in low- or middle-income countries. As noted earlier, the Thailand case is exceptional insofar as historical risk-pooling tra- ditions combined with a unique set of political conditions to push the country's SHI toward a policy of universal coverage when per capita incomes were only about US$2,500. Theory of SHI The uncertainty of illness underpins the theory of SHI (Arrow 1963; Rothschild and Stiglitz 1976). Each year a relatively small number of people suffer from serious ill- ness and disability. Their medical problems can result in large medical expenses that most people cannot afford, but faced with life and death decisions, people will tend to seek expensive medical services even though the costs may bankrupt patients and their families. Consequently, most people want to be insured against such risks because they are risk averse. At the same time, some people may not demand insurance because they believe that illnesses and accidents will spare them or they simply ignore the risks of potentially impoverishing their families should they experience catastrophic financial loss. Such an irrational choice could create serious social problems. Moreover, people are also selfish. If health insurance is voluntary, young, healthy people will not want to pool their low health risk with high-risk people such as the elderly and the chronically ill. This leads to adverse selection, a critical problem of voluntary group insurance. Meanwhile, poor and low-income households cannot afford the insurance premiums, and thus have to be subsidized. For all these reasons, nations that wish to implement universal health insurance must look beyond a purely voluntary system. SHI pools low- and high-risk people, avoids adverse selection and people's fail- ure to address risks, and allows enrollees to contribute based on their ability to pay. SHI creates an exchange in a public market whereby the insured pay a desig- nated amount (the premium) and in return receive a set of benefits. The premium does not come from the government's budget (general revenues), but instead from a compulsory premium that is usually assessed as a percentage of workers' salaries. The free-standing nature of SHI financial operations makes the system transparent and accountable in terms of how much people pay and what they are paying for. In short, SHI is a modern socioeconomic program that promotes equity and creates solidarity. It pools risks and redistributes income between the rich and the poor, the healthy and the less healthy, and the old and the young. As SHI evolved, it took on an additional function: prepayment. Many people might be able to afford basic health services and drugs, and these could be rationed by price, that is, by having people pay out-of-pocket. However, price rationing could deter many people from seeking early diagnosis and treatment until their conditions develop into serious, acute illnesses. Studies find that prevention and primary care are more cost-effective in relation to improving health than second- ary and tertiary medical services, but people often do not demand enough primary William C. Hsiao and R. Paul Shaw 15 care because their illnesses are not obvious or acute (Fries and others 1993; Leaf 1993; Somers 1984). Studies of voluntary purchases of health insurance found that people prefer insurance plans that not only insure them against catastrophic medical expenses, but allow them to prepay for smaller health expenses, because this increases the chances that those insured would receive some payout from their plans. The early history of private insurance in the United States showed that most people purchased prepayment plans in preference to insurance against catastrophic medical expenses (Somers and Somers 1961). For these reasons, modern SHI is more than an insurance plan that insures only against large medical expenses. Instead, SHI systems are also designed to serve as prepayment plans for less expensive health care such as pre- ventive services and primary care, as prepayment encourages people to use more of these cost-effective services. However, prepayment and insurance creates moral hazard, in that people may demand unnecessary services and drugs (Pauly 1974). Copayment and coinsurance are designed to reduce the risk of moral hazard. Definition of SHI SHI has three distinct characteristics. First, social insurance is compulsory, which is the major feature that distinguishes SHI from voluntary private insurance. Under social insurance, everyone in the contributory regime group must enroll and pay the specified premium or contribution. As noted earlier, the contribution is most often specified as a percentage of wages, or, as referred to by the economic litera- ture, a dedicated payroll tax. For the poor and certain special categories of the pop- ulation, such as the elderly and children, the government may pay the premiums on their behalf. The latter are referred to as enrollees in the subsidized regime. The second characteristic pertains to eligibility. Citizens only become entitled to receive benefits when they have paid the required premium. In other words, SHI is not necessarily universal, unlike health insurance financed by general revenues, whereby every citizen is covered, such as the Canadian system. The third characteristic is that SHI premiums and benefits are described in a social compact, which is usually set out in the legislation that establishes the eco- nomic exchange between the two parties: enrollees and the social insurance plan. Thus the legislative process creates an implicit bargain between the SHI and those it covers. Citizens agree to pay a certain amount with some confidence that their contributions will be used fairly and effectively to fund health care for all those who are part of the system. The benefits package specifies in writing the benefits the insured are entitled to in exchange for payment of the premium. The contribu- tion rate and benefits are secure, and not subject to annual budgetary decisions by the government that tax-funded systems such as that in Canada face. If the SHI fails to deliver the promised benefits specified in the benefits package, the enrollees can seek legal remedies. Another way of characterizing SHI is to crudely distinguish it from other forms of health insurance as shown in table 1.4. Note that a system whereby man- datory contributions from a defined membership, combined with the earmark- ing of funds for health, lends itself to more rigorous assessment of the financing required to honor commitments (namely, the benefit entitlements promised to members) than national insurance, which promises universal coverage "free" to all citizens, but does not guarantee that tax revenues will be sufficient or will be 16 Introduction, Context, and Theory Table 1.4. Selected Characteristics of SHI Compared with Other Forms of Health Insurance Funds Type of Nature of earmarked for insurance Financing source contributions health Membership SHI Employer and/or Mandatory Yes Contributing employee contributions members and from wages and usually their salaries dependents Private Out-of-pocket Voluntary Yes Contributing insurance payments of premiums members and usually their dependents Community Out-of-pocket Voluntary Yes Contributing prepayment payments of premiums members and health usually their "insurance" dependents National Government general Funded No All citizens insurance revenues and other mostly from taxes tax revenues allocated to health. The reality of national insurance is that competing priorities may deplete the promised funding for health, for example, natural disasters, wars, or shifting political agendas. Such problems are inherent to the dilemma regard- ing health financing through general revenue taxation in many low- and middle- income countries. SHI often evolves in parallel with other forms of insurance, that is, multiple types of insurance often provide coverage or supplementary coverage for groups with different benefits packages. Sometimes this occurs in a highly disjointed fash- ion, resulting in administrative duplication, overlaps, and inefficiencies. At a later date, governments may face the challenge of somehow unifying the benefit entitle- ments and membership under a single system as illustrated in chapter 7. Recent developments have modified the definition of SHI. Developing countries embark on SHI aiming for universal coverage, but realize that achieving this in the near future is impossible. Nonetheless, the SHI scheme provides a comprehen- sive framework whereby everyone can eventually be insured. At the start, some people will enroll voluntarily because the SHI agency lacks the ability to compel all affluent people to pay. The voluntary groups usually consist of nonpoor farm- ers and workers in the informal sectors. In the long run, when a nation reaches a more advanced stage of socioeconomic development and most of its citizens are employed in the formal sector, the SHI could become universal, compelling all non- poor citizens to enroll. Major Preconditions for SHI A successful launch of SHI requires several major preconditions. William C. Hsiao and R. Paul Shaw 17 Box 1.1. Terms In this volume, and throughout the literature, there tends to be considerable overlap in the use of terms involving payment or reimbursement of providers for health services and commodities, such as user fees, fee-for-service, copayments, and cash-and-carry. Our use of these terms implies the following in developing countries: · User fees are charges for health services that tend to be paid for out-of-pocket by clients to providers at the time of service. In developing countries, the con- cept of "user fees" usually invokes the notion that patient charges have been introduced at public facilities, where services may have been provided "free" of charge in the past. In such cases, the user fees rarely represent full cost recov- ery, but rather try to retrieve some portion of the cost so as to contribute to the recurrent costs of public health facilities. · Fee-for-service are charges for health services that tend to be paid for out-of- pocket by clients to private or NGO providers at the time of service, or that are paid, retrospectively, to the providers by health insurance on behalf of clients. Fee-for-service as a form of reimbursement is usually regarded as a formal pay- ment mechanism that represents full cost recovery or full cost recovery plus a profit margin. · Copayments are partial charges requested from clients at the time of service by private, NGO, or public providers, who will be principally reimbursed for the services they provide by health insurance entities. Copayments are thus a form of cost sharing between health insurance entities and clients, with the amount of cost sharing pre-agreed as part of the health insurance policy. For example, a health insurance entity may agree with its client members that clients will pay 10 to 20 percent of charges out-of-pocket, while the lion's share will be reim- bursed by the insurance entity. · Cash-and-carry are charges for health care commodities, principally drugs, paid for out-of-pocket by clients. Cash-and-carry schemes for drugs in publicly financed health systems have often been seen as a way of immediately replen- ishing the recurrent funds needed to fully stock drug supplies rather than waiting on uncertain commitments of public funds. The charges involved may represent full or partial cost recovery. INCENTIVE FOR PEOPLE TO PAY PREMIUMS. People must be motivated to accept and pay for SHI, even in compulsory systems. People have the incentive to prepay only if they currently have to pay for their health services. If adequate public sec- tor services of good quality are provided free or nearly free, why would people who use these services want to enroll and pay for SHI? People will not want to pay for SHI unless user fees (box 1.1) are high, if patients have to purchase drugs and supplies, or if public services are so poor that many patients pay out-of-pocket for private providers. A comparison of the Ghanaian and Tanzanian experiences can be instructive. Ghana shifted to the "cash-and-carry" (user fee) system in 1999, and patients had to pay fairly high user fees. Consequently, voluntary prepayment plans such as the community-based mutual health organizations (MHOs) flourished, growing from 4 MHO funds in 1999 to 157 by 2002. In 2003, Ghana was able to pass legislation to establish SHI nationwide, relying on the MHOs as a building block. By contrast, Tanzania does not have high user fees. Since 1996, Tanzania has tried to attract and enroll its population into its district-based insurance, the community health funds. The government subsidizes 50 percent of the premium, regardless of income level, 18 Introduction, Context, and Theory yet the enrollment rate remains low, ranging from 5 to 20 percent of the eligible population, and those who enroll tend to be the elderly and the sick. CERTIFICATION OF QUALIFIED PROVIDERS. Developing nations have tended to pay little attention to the safety and quality of health services rendered in the private sector, other than establishing minimum standards such as licensing requirements. Following initial licensing, the actual safety and quality of health services remain largely unmonitored and unregulated. In rural areas, drug peddlers and indigenous doctors have free reign, because regulations are not enforced. Moreover, govern- ments rarely require private facilities to be transparent in relation to their financial operations or to adopt modern financial and medical record systems. Under such conditions, the quality of private sector health services is highly variable, and detect- ing fraud and price gouging when SHI pays for claims is difficult. Publicly provided health services are also problematic. Governments manage public facilities by means of bureaucratic rules that tend not to encompass modern accounting, financial, and clinical information systems. The average clinical qual- ity of public facilities might be better than that of private facilities, but is neverthe- less highly variable. These deficiencies have to be remedied before or concurrently with SHI to gain sustained public support, perform its role of assuring a reasonable quality of health care, and sustain its operations financially. The SHI administration should purchase health care for its insured prudently. A prudent purchaser has to ensure that services and drugs meet certain stan- dards. Equally important, SHI has to be able to control fraudulent claims and sup- plier-induced demand for unnecessary services; "inside" dealings between doc- tors, pharmacies, and testing laboratories; and so on. Conditions in the market for health services often require SHI to set safety, quality, financial, and audit stan- dards beyond what currently exist so that SHI can be a responsible and prudent purchaser. Under such circumstances, SHI has to develop and implement new standards and enforcement mechanisms to assure the safety and clinical quality of health care, as well as standard medical records and accounting systems, and adequate inspection and auditing of providers. Establishing standards and a sys- tem for enforcing them must be high priorities before SHI can be implemented. RAPID ECONOMIC GROWTH. As noted previously, rapid economic growth is an important consideration in sustaining an SHI program and in expanding it to achieve universal coverage. Health care costs rise rapidly because of inflation, ris- ing expectations, and expensive new drugs and technology. Unless wage rates are also rising rapidly, premiums would have to be increased frequently. Meanwhile, governments need rising revenues to subsidize the growth in premiums for the poor and to expand coverage. Moreover, rapid economic growth has positive effects on SHI enrollment in that it (a) can lift people out of poverty, meaning that more people can afford to pay their premiums; (b) can bring more workers into the formal sector, which increases the number of people in the contributory regime; (c) can raise the government's general revenues, meaning that the government can subsidize more of the poor; and (d) tends to increase the government's administrative capacity to collect taxes and insurance premiums. Rapid economic growth will therefore enable a nation to move toward universal coverage. William C. Hsiao and R. Paul Shaw 19 References Arrow, K. W. 1963. "Uncertainty and the Welfare Economics of Medical Care." American Eco- nomic Review 53 (5): 940­73. Bitran, R. 1995. "Efficiency and Quality in the Public and Private Sectors in Senegal." Health Policy and Planning 10 (3): 271­83. Carrin, G., and C. James. 2005. "Social Health Insurance: Key Factors Affecting the Transi- tion towards Universal Coverage." International Social Security Review 58 (1): 45­64. Castro-Leal, F., J. Dayton, L. Demery, and K. Mehra. 1999. "Public Social Spending in Africa: Do the Poor Benefit?" World Bank Research Observer 14 (1): 49­72. Diop, F., A. Yazbeck, and R. Bitran. 1995. "The Impact of Alternative Cost Recovery Schemes on Access and Equity in Niger." Health Policy and Planning 10 (3): 223­40. Dror, D. M., and C. Jacquier. 1999. "Micro-Insurance: Extending Health Insurance to the Excluded." International Social Security Review 52 (1): 71­97. Foster, S. 1993. "Economic Aspects of the Production and Use of Pharmaceuticals in Devel- oping Countries." In Health Economics Research in Developing Countries, ed. L. Mills, 288­313. Oxford, U.K.: Oxford University Press. Fries, J. F., C. E. Koop, C. E. Beadle, P. P. Cooper, M. J. England, R. F. Greaves, J. J. Sokolov, and D. Wright. 1993. "Reducing Health Care Costs by Reducing the Need and Demand for Medical Services: The Health Project Consortium." New England Journal of Medicine 329 (5): 321­25. Fu, W. 1999. Health Care for China's Rural Poor. Washington, DC: World Bank. Gertler, P., and O. Solon. 2002. "Who Benefits from Social Health Insurance in Developing Countries?" Working paper, University of California Berkeley. Gilson, L. 1995. "Management and Health Care Reform in Sub-Saharan Africa." Social Sci- ence and Medicine 40 (5): 695­710. Government of Tanzania. 2003. "Health Sector Public Expenditure Review." Government of Tanzania, Dar Es Saalaam. Unpublished report. Ha, N. T. H., P. Berman, and U. Larsen. 2002. "Household Utilization and Expenditure on Private and Public Health Services in Vietnam." Health Policy and Planning 17 (1): 61­70. Hjortsberg, C. A., and C. N. Mwikisa. 2002. "Cost of Access to Health Services in Zambia." Health Policy and Planning 17 (1): 71­77. IMF (International Monetary Fund). 2002. Mobilization of Domestic Resources for Health. Wash- ington, DC: IMF. Leaf, A. 1993. "Preventive Medicine for Our Ailing Health Care System." Journal of the Amer- ican Medical Association 269 (5): 616­18. Liu, Y., K. Rao, and S. Hu. 2002. Towards Establishment of a Health Protection System in Rural China. Beijing: Asian Development Bank. Mills, A. I. 1995. Improving the Efficiency of Public Sector Health Services in Developing Countries: Bureaucratic Versus Market Approaches. Public Health and Policy Departmental Publica- tion Series. London: London School of Hygiene and Tropical Medicine. Pauly, M. V. 1974. "Over Insurance and Public Provision of Insurance: The Roles of Moral Hazard and Adverse Selection." Quarterly Journal of Economics 88 (1): 44­62. Peters, D. H., A. S. Yazbeck, R. Sharma, G. N. V. Ramana, L. Pritchett, and A. Wagstaff. 2002. Better Health Systems for India's Poor: Findings, Analysis, and Options. Washington DC: World Bank. Rothschild, M., and J. E. Stiglitz. 1976. "Equilibrium in Competitive Insurance Markets: An Essay on the Economics of Imperfect Information." Quarterly Journal of Economics 90 (4): 629­49. 20 Introduction, Context, and Theory Saltman, R. B., and H. F. W. Dubois. 2004. "The Historical and Social Base of Social Health Insurance Systems." In Social Health Insurance Systems in Western Europe, ed. R. B. Salt- man, R. Busse, and J. Figueras, 21­32. Maidenhead, U.K.: Open University Press. Somers, A. 1984. "Why Not Try Preventing Illness as a Way of Controlling Medicare Costs?" New England Journal of Medicine 311 (3): 853­56. Somers, H. H., and A. R. Somers. 1961. Doctors, Patients, and Health Insurance. Washington, DC: The Brookings Institution. Stiglitz, J. 1999. "Incentives and Institutions in the Provision of Health Care in Developing Countries: Toward an Efficient and Equitable Health Care Strategy." Paper presented at the International Health Economics Association, Rotterdam, June 7. WHO (World Health Organization). 2000. World Health Report 1999: Health Systems-- Improving Performance. Geneva: WHO. ------. 2002a. Mobilization of Domestic Resources for Health: Report of Working Group 3 of the Commission on Macroeconomics and Health. Geneva: WHO. ------. 2002b. World Health Report. Geneva: WHO. ------. 2005a. Sustainable Health Financing, Universal Coverage, and Social Health Insurance. Agenda Item 13.16, 58th World Health Assembly. Geneva: WHO. ------. 2005b. World Health Report 2005: Make Every Mother and Child Count. Geneva: WHO. World Bank. 1993. Better Health in Africa. Washington, DC: World Bank, Africa Technical Department. ------. 2005. World Development Indicators 2005. Washington, DC: World Bank. Zere, E., D. Mcintyre, and T. Addison. 2001. "Technical Efficiency and Productivity of Public Sector Hospitals in Three South African Provinces." South African Journal of Economics 69 (2): 336­58. 2 Design and Implementation of Social Health Insurance William C. Hsiao SHI must be designed in a way that is not only sufficiently inclusive in terms of benefits, but so that it is affordable and sustainable over time. Moreover, it must be implemented in ways that do not exceed the capacity of health systems and that can be regulated to assure quality. Design SHI is a tool for achieving several goals: mobilizing more funds for health, promot- ing equal access to reasonable health care for the poor, pooling health risks and pre- venting impoverishment, and improving the efficiency and quality of health care. The design of SHI essentially involves maximizing social benefits under financial and political constraints. The policy makers have to decide on at least seven inter- connected groups of major questions in order to pursue the goals of SHI. For exam- ple, policy makers have to trade off between the goal of covering as many poor as possible with the goal of offering them a comprehensive benefits package. Mean- while, the cost of the benefits package is also determined by how the SHI contracts for and pays providers. The following are the seven major groups of questions. The first three deal with determining the composition of the contributory, poor, and nonpoor self-employed populations. The remaining four questions deal with deter- mining the benefits package, assessing the fiscal capacity to cover the benefits and expand enrollment to achieve universal coverage, determining SHI governance, and improving the delivery system for better quality and efficiency of health care. · How many people can be enrolled under the contributory regime, who should these people be, and how can premiums be collected from them? The SHI premiums the insured pay are the new source of financing that provides additional funds for health while pooling their risks. · How should "poor" be defined, how many of the poor should be subsidized, how should the subsidy be targeted, how much will this cost, and how will it be financed? Answers to these questions determine the government budget required for funding SHI for the poor, how many poor will gain equal finan- cial access to health care, and to what extent health gains will be produced. 21 22 Design and Implementation of Social Health Insurance · How can nonpoor self-employed and informal sector workers be enrolled and how can premiums be collected from them? Decisions on these issues influence whether a nation will be able to achieve universal SHI and pool health risks widely, and also have the potential to reduce fraudulent claims, because people who are not covered often "borrow" membership cards of those covered so that they can obtain health services. · What is the benefits package for each group and how much will it cost? These decisions determine the premium rates for the contributory regime popula- tion, the government budget needed to fund the poor, and people's access to health care and insurance protection. · What is the nation's fiscal capacity to fund the poor and near-poor to achieve universal coverage, and what is the projected timetable for this? Such plan- ning disciplines decision makers in the adoption of a long-term strategy for SHI and considers the steps that have to be taken to achieve universality. · How should SHI be governed? Should the SHI agency be a public agency, a quasi-public agency, or a private nonprofit entity? What is the best and most viable administrative structure for SHI? Decisions on these matters influence the efficiency and effectiveness of SHI operations. · How can SHI improve health care delivery? How should providers be con- tracted and paid? Decisions pertaining to these issues influence the effi- ciency, cost, and quality of health care, and in turn the population's health, the premium rates, and the government's budget for subsidizing the poor. Determining Eligible Population Groups The success of an SHI system depends largely on its ability to enroll and collect premiums from the population and the government's ability to subsidize premi- ums for the poor. The law defines who is eligible to enroll and pay a premium and who will be subsidized. SHI for developing countries usually divides the popula- tion into three groups: compulsory enrollment and payment, voluntary enrollment and payment, and fully or partially subsidized. The premium and benefits package for each group may differ. This section addresses the identification and coverage of these three population groups. The government can compel large employers in the formal sector to enroll and pay premiums for their workers and their workers' families. The nonpoor self-employed, including farmers, and informal sector workers may have to be treated differently, as enforcing compulsory premium payments by this group is difficult, and developing nations often have to offer such people voluntary enrollment. The poor should be fully subsidized and the near-poor may be par- tially subsidized. We use Uganda to illustrate the potential coverage of SHI for the four catego- ries of population. As conveyed in table 2.1, only about 10 percent of the Ugandan working population, or 800,000 people, work for the government or firms with 10 or more employees; the poor account for close to 50 percent of the population; and the near-poor account for an additional 20 percent. These statistics indicate that, at best, about 10 percent of households would pay for social insurance, but enrolling that 10 percent of households into social insur- ance could significantly benefit the poor. This 10 percent is likely to live in cities and consume close to 35 percent of government tax-funded public health services. William C. Hsiao 23 Table 2.1. Working Population by Employment Status and Income, Uganda, 1999­2000 (thousands of people) Income level High Middle Low income Total economically Employment income income or poor active population Formal sector Government 25 232 n. a. 257 Private Employers with 10 or more employees 110 358 100 568 Employers with fewer than 10 employees 30 200 398 628 Informal sector and self-employed 80 200 537 817 Farmers and hunters 250 750 4,760 5,760 Total 495 1,740 5,795 8,030 Under a well-designed SHI system, these households would pay for their own health expenses through premiums, and the money the government previously spent for them could be used to pay premiums for the poor. POPULATION COVERED UNDER THE CONTRIBUTORY REGIME. Enrollment and premium collection can be implemented relatively easily for workers employed by formally registered and large enterprises such as banks and manufacturing corporations, which collectively are often referred to as the formal sector. Such employers have to maintain accurate and reliable personnel and wage payment records for business purposes and the SHI plan can rely on these records to col- lect premiums, otherwise significant portions of these workers and their employ- ers may evade premium payment. However, adverse selection remains a critical problem even for formal sector employees. To overcome this, SHI has to create incentives for workers to enroll by requiring employers to pay a share of the SHI premium (private group insurance also does the same to reduce adverse selec- tion). Colombia, Ghana, Kenya, the Philippines, and Thailand all designed their SHI so that employers pay half or more of the premium. Evasion has been a serious problem for the employees of small employers. A significant portion of this population is unwilling to pay, particularly high-income employees whose contributions are large when premiums are assessed as a per- centage of their wages. Experience worldwide has shown that evasion is the major challenge to successful implementation of SHI. For example, a study found that 10 years after the implementation of SHI, 35 percent of Colombians who should pay under the contribution system are still able to evade doing so (Bitran and Associ- ates, Econometrita, and Superior School of Public Health 2002). Compulsory contributions to social insurance do have economic implications, for instance, will such contributions have a negative impact on the labor market and on economic growth? Employers have a certain willingness to pay for vari- ous kinds of workers and they may not care whether payment takes the form of 24 Design and Implementation of Social Health Insurance wages or fringe benefits. How much of the employers' share of SHI contributions can be passed back to workers by paying them lower cash wages depends on labor market conditions, including the strength of labor unions. For countries with weak labor unions, empirical studies reveal that in the medium-term, workers will pay the largest share of health insurance premiums, either directly or in the form of lower wages (Gruber and Krueger 1991). If premiums do not affect total labor costs, then the impact of social insurance on the labor market and on economic develop- ment should be minimal. However, workers may reduce their labor supply if they value cash wages more than the health insurance premium deducted from their paychecks. At the same time, if employers cannot shift their premium costs back to workers, they reduce the demand for labor. These negative effects on the labor mar- ket could reduce economic growth. POPULATION COVERED UNDER THE SUBSIDIZED REGIME. The poor cannot afford premium payments, thus the government has to fully subsidize the poor and par- tially subsidize the near-poor. All nations have provided a government subsidy to the poor for their premiums when they established universal SHI plans. In low- income countries, the poor and the near-poor account for a large share of the total population,forexample,50percentofthepopulationinKenya,49percentinColom- bia, 40 percent in Ghana, 37 percent in the Philippines, and 12 percent in Thailand live in poverty. Subsidizing more than one-third of the population requires a huge budget. Consequently, developing countries design their SHI systems to gradually phase in the poor, in the hope that their economies will grow over time and fewer people will need to be subsidized, as happened in newly industrialized economies such as Korea (Yu and Anderson 1992) and Taiwan, China (Lu and Hsiao 2003). In order to target the subsidy to the poor, countries must first define who is poor and near-poor and then find a way to identify them fairly and accurately. Identify- ing the poor usually requires an income test, and accurate assessment of those who are eligible for the subsidy entails complex procedures and detailed income data. ThedesignersofSHIhavetofindabalancebetweenadministrativecomplexity(and its associated costs) and preventing cheating. Sometimes corruption also creeps in when local officials use their power to award subsidies to their relatives, friends, or political supporters. Ghana and the Philippines delegated the determination of eligibility for the subsidy to district governments, but with strict oversight from the SHI agency. Colombia relied on its elaborate and systematic scheme known as the system for identifying beneficiaries (sistema de identificación de beneficiarios or SISBEN), which periodically investigates and classifies every household into one of six income levels. Instead of income testing, some countries subsidize easily identifiable groups the majority of whose members are poor. Examples include farmers in certain poor regions,residentsof poor districts, theelderly, orphans, andthedisabled.Forexam- ple, Ghana subsidizes the elderly and children. This approach might be preferable, because it reduces leakage of the subsidy and incurs lower administrative costs. The amount of subsidy required varies depending on the premium. The pre- mium for the poor is largely determined by the benefits package offered to them and the production efficiency of health care. Rationally, the basic benefits package should be designed using actuarial cost estimates prepared to ascertain the budget required to fund the package. Balancing the design of the benefits package with its William C. Hsiao 25 costs has been a critical political problem in establishing SHI. Political leaders want to promise the most but pay the least. As a result, nations often design a generous basic benefits package and include it in the SHI legislation, but without stipulat- ing its costs and budget, which perhaps may not even have been determined. Con- sequently, the legislation could be an empty promise that cannot be implemented because of a lack of funds. Colombia, Ghana, and Kenya are good illustrations of this political legislative process. Kenya underwent a painful process to reconcile the benefits promised and the budget required to fund them. Often a nation has to compromise and provide the poor with fewer benefits than anticipated or promised. Colombia passed leg- islation that intended to give everyone the same benefits package, but during the implementation stage, the government had to face realities about the costs of such a uniform scheme and compromised with a three-tiered benefit structure. COVERAGE OF THE NONPOOR SELF-EMPLOYED AND WORKERS IN THE INFORMAL SECTOR. Extending SHI to cover the nonpoor, the self-employed, and employees in the informal sector presents the greatest challenge to SHI, as they do not work for organizations where SHI premiums could be deducted from their salaries. The self-employed include farmers, fishermen, hunters, shopkeepers, and day laborers. Employees in the informal sector include maids working in people's homes, waiters in small eateries, and hired help in small shops. Their household income is above the poverty line and they should pay the premium. While such people are not poor, they are likely to be less able to pay for SHI because their incomes are lower, on average, than those of formal sector employees. Often the government has to sub- sidize these nonpoor to provide the incentive for them to enroll, which means that the government must commit a substantial amount to subsidize nonformal sector workers and their families for SHI to become universal. Enrolling nonpoor farmers and vendors in rural areas is particularly difficult, and nations have tried different approaches to enroll them. Community-based pre- payment schemes are being tried in China, Colombia, Ghana, India, and Tanzania and seem to hold some promise, particularly in China and Ghana. Determining the Level of Benefits A key policy decision in designing SHI involves what services the benefits package covers. Costs are directly related to the comprehensiveness of the benefits package. This issue immediately raises the question of what is affordable for different popu- lation groups, for example, self-employed workers may not be able to afford the same benefits package that can be financed by employer and employee contribu- tions. The issue often becomes a choice between comprehensive benefits but fewer people covered versus less extensive benefits but more people covered. When a country adopts SHI, its government is often unwilling or unable to allo- cate enough tax funds to finance the same benefits package for the poor as for oth- ers. As a result, the poor and the self-employed often get a much smaller benefits package than those who are employed in the formal sector. As the five cases illus- trate, all five nations except the Philippines have different benefits packages for dif- ferent groups, with the poor getting much less. In the Philippines, SHI for the poor became a key presidential election strategy to solicit votes from the poor, so full 26 Design and Implementation of Social Health Insurance benefits were extended to them. Indeed, the poor are the only group with outpa- tient benefits. In designing an affordable benefits package, the question arises as to what services should be included. The current academic literature argues that cost-effectiveness studies should be the basis for selecting services to be included or excluded (Gold and others 1996). In reality, a benefits package cannot be based on current cost- effectiveness studies, because this field is still at a primitive stage of development. Current cost-effectiveness studies only consider one effectiveness criterion--health gains--and totally ignore protection against financial risks. Meanwhile, the SHI benefits package has to achieve two social purposes: health gains and protection against impoverishment from catastrophic medical expenses. Assessing the Fiscal Requirements for Funding SHI and Achieving Universal Coverage The major technical issue pertaining to the viability and sustainability of SHI relates to its costs and whether sufficient funding can be provided for it. While the goal may be to establish SHI to achieve universal, equitable access to reasonable health care, the tax funds required to finance it could be prohibitively high. Plan- ning a sustainable SHI requires several rounds of analysis, each of which requires a careful specification of those eligible to be covered, a detailed delineation of the actual benefits package and an actuarial analysis of its costs, and an assessment of how the costs will be financed. This process is when noble visions, however wor- thy, face a reality check. The Kenya case illustrates the fiscal challenge facing the government. The legis- lation intends to provide universal SHI. Employees in the formal sector will pay a percentage of their wages to cover the full cost of their premiums and others will payaflat-ratepremium.Thegovernmentplanstoimposean11percentvalueadded tax to fully subsidize the poor, who account for around 30 percent of Kenyans. Even though the president proposed legislation to establish SHI and parliament passed it, the president has delayed signing the legislation into law because he is uncertain that the contribution rate will be sufficient to fund SHI in the long term. Ghana established an SHI program in 2003 and issued regulations clarifying the intent of the law in 2004. Ghana's strategy differs from that of Kenya. Ghana plans to enroll 20 percent of the total population in three years and 50 percent of its citizens in 15 years. These seem to be realistic goals, but whether the Ghanaian community-based based insurance scheme can overcome the difficulties of enroll- ing the poor, nonpoor farmers, and informal sector workers to achieve the 50 per- cent goal remains to be seen. The Philippines passed a law and established a universal SHI program in 1995, intending to achieve universality by 2005, but 40 percent of the population still remains uncovered. Most of these are the near-poor, the nonpoor self-employed, and workers in the informal sector. The government has now stated that its goal is to achieve universality by 2010. Colombia implemented SHI in l993 with the intent of achieving universal- ity within a few years by compromising the benefits package for the poor, who account for 40 percent of the population, that is, giving the poor half the benefits that employed workers received. Nevertheless, 11 years later, only 67 percent of the William C. Hsiao 27 population was enrolled. The near-poor and informal sector workers have largely remained uncovered by SHI. Thailand has just achieved universal coverage. The successful candidate for prime minister in 2001 put universal SHI at the top of his election promises. Once elected, the prime minister allocated sufficient government funds to pay the pre- mium for all the poor, the near-poor, and the nonpoor informal sector workers. Nonetheless, their benefits package is less than those employed in the formal sector receive, which leaves Thailand with a multitiered system. Realistically, developing nations may achieve universal coverage in two or three decades if their economies grow rapidly. SHI can first cover civil servants, formal sector employees, and the poor who are being fully subsidized, then move step-by- step to include other groups until universal coverage has been achieved. World- wide experience indicates that SHI can be effectively implemented in developing countries for formal sector workers plus those who are largely subsidized by the government budget, such as the poor and the near-poor. Enrolling and collect- ing premiums from nonpoor workers in the informal sector and farmers are the major hurdles for universal coverage. It seems that universal coverage can only be attempted with a chance of success when a nation's economy has grown to approxi- mately US$3,000 per capita per year (as exemplified by Costa Rica and Thailand). Table 2.2 summarizes how the five countries discussed in this volume designed their SHI systems and the populations covered under the three regimes: contrib- utory, fully subsidized, and voluntary enrollment of nonpoor self-employed and informal sector workers. The cost of SHI is largely determined by the design of its benefits package. Establishing Governance for SHI SHI shifts some power from the supply to the demand side to balance the influence of patients versus providers. SHI empowers patients to demand satisfactory health services, and the SHI agency should act as a wise and prudent purchaser for the insured. However, in terms of how SHI should be structured to best represent the interests of the insured, several issues have to be addressed at the design stage. SHI governance, narrowly defined, means the structure and processes of the control mechanisms used to hold the SHI agency accountable to beneficiaries and funders (that is, the government and employers) of the scheme (Savedoff 2005). Governance also pertains to the integrity of management. According to economic, political, and organizational theories, the following critical choices have to be made in relation to SHI governance, taking into account a country's political structure and institutions, climate of law enforcement, and power of interest groups: · OwnershipoftheSHIagency.Ownershipcanbepublic,quasi-public,orprivate nonprofit. Ownership drives an organization's motives and purpose. Public organizations are controlled by the government and closely tied to politics and political influence, while private nonprofit organizations are more inde- pendent and insulated from political interference. · Organizational structure. Organization can be analyzed using principal-agent theory. Beneficiaries and premium payers act as the principals who select and contract the board of directors as their agent to represent and pursue the principals' interests. In turn, the board contracts with the chief executive Table 2.2. Design Issues, Selected Countries, as of 2005 Design issue Kenya Ghana Philippines Colombia Thailand Percentage of the 20 (in current hospital 20 60 67 100 population covered insurance fund) by SHI Percentage of the 20 -- 30 30 35 population who contribute Percentage of the 50 40 37 50 12 28 population that is poor Percentage of the Not yet implemented Elderly, children, plus 23 33 65 population that is fully a few poor (limited subsidized to 0.5% of each health insurance plan) Percentage of the 80 -- -- Included in the 30% -- population who are who contribute (an nonpoor self-employed additional 4% receive or informal sector partial subsidization) employees Table 2.2. Design Issues, Selected Countries, as of 2005 (continued) Design issue Kenya Ghana Philippines Colombia Thailand Percentage of the None -- 7 -- -- nonpoor self-employed who are enrolled Benefits package Not decided Intended to be Universal Tiered Tiered universal or tiered universal Fiscal capacity for Fiscal worries stalled Plans to enroll 50% in Most of the uncovered Despite benefit Achieved universality universality the SHI plan before 15 years, which will are near-poor and compromises and large with a lesser benefits 29 it started, as it would require full subsidies informal workers; allocated funds, most package for those require large new for the poor and partial whether they can be near-poor and informal subsidized taxation subsidies for farmers covered remains to be workers are uncovered and informal sector seen workers Administrative structure One new Competition between One fund One fund Several funds, trying to independent fund community-based merge organization and private plans 30 Design and Implementation of Social Health Insurance officer (CEO) as the agent to pursue the goals set by the board. Hence the composition and election of board members and the determination of how they are accountable to beneficiaries and premium payers become the para- mount concern in relation to organizational design. · Management structure. The management team must be given discretion in financial and personnel decisions so that it has the power to use financial and human resources to achieve the goals of the SHI agency. For example, if employees of the SHI agency are civil servants, then civil service rules on hiring, promotion, and firing would tie the management team's hands. · Government supervision. The state has to regulate and monitor the operations and performance of the SHI agency for two reasons, first, to assure that SHI serves society's interests and the original purposes for which it was estab- lished, and second, to supervise the SHI agency because the government is usually a principal funder of SHI. When a nation decides to establish an SHI agency as a public or quasi-public organization, it has at least three choices as to which ministry should be respon- sible for SHI: the ministry of health; the ministry of labor and social security; or a new independent ministry, the SHI agency. Most developing nations have placed their new SHI agency under the ministry of health, but with other ministers also serving on the supervisory board. Each structure has potential problems. Several studies of SHI agencies orga- nized under the ministry of health found discouraging performance, as the min- istries were dominated by medical professionals who were more concerned about the welfare of the supply side than of the demand side and the additional revenues generated by the SHI often largely benefited the suppliers of medical services. For example, a study of Colombia (General Comptroller of the Republic 2002) found that the salary of health staff increased by more than 40 percent in real terms dur- ing the initial years of SHI. Similarly, a study of the Philippines (Gertler and Solon 2002) found that 86 percent of the increased funding for health that became avail- able due to SHI financing went to providers as profits or higher salaries. The ministry of labor and social security may not be able to manage SHI any better, however. SHI managed by the ministry of labor tends to view insurance simply as a payment mechanism, rather than as a prudent, organized purchaser for the insured. SHI usually focuses on the fund's balance by controlling what the SHI will pay and shifts the remaining liability to patients. Providers are not prevented from "balance billing" and "extra billing" patients. These SHI systems tend to have low loss ratios and accumulate large surpluses in the SHI fund. In other words, the premiums are not necessarily used for the benefit of the insured. When the ministry of health manages SHI, the ministry has to undergo a major transformation, because SHI changes the ministry's role from a financier and opera- tor of public sector services to one that sets policies and regulates all providers. Gen- erally, ministries of health have focused on financing and managing public facilities and their staff, while paying less attention to patients' well-being. Some ministries of health tend to protect the interests of public providers and take a laissez-faire atti- tude toward private sector providers. If a ministry of health starts at such a point, it will require a major change in its corporate culture for the ministry to represent William C. Hsiao 31 patients' interests effectively. Moreover, to become an effective purchaser requires the ministry to reorganize its functions and operations. As explained earlier, when developing countries adopt SHI, they generally have to divide the population into different eligible groups, each of which may have a different benefits package, premium and subsidy rate, and enrollment procedures. Hence the question arises whether a nation should have just one fund or sepa- rate funds and administrations for each group. This issue often generates furious bureaucratic competition between the ministry of health and the ministry of labor and social security over who controls the fund or funds. The decision between single or multiple funds influences how well a nation can provide equitable SHI when it approaches universality. Having separate funds for each group with their own administrative organizations creates political and bureaucratic barriers to universal SHI with equal access to reasonable health care for all. The major challenge currently facing Thailand is how to merge various funds once universal coverage has been achieved, because each fund has its own benefits package, payment methods and rates for providers, and bureaucratic rules. Even more important, each fund has an established, political, vested interest group with strong supporters. Merging funds to create a truly universal plan takes huge political capital. Germany, the grandfather of SHI, has been trying to gradually merge its funds, but still has almost 300 funds (Busse and Riesberg 2004). Korea spent a dozen years mobilizing strong public support and a presidential commit- ment before it was able to merge its many funds (Kwon 2003). In addition to creating barriers to universality, having multiple funds creates duplication in administration and complicates the administrative work for provid- ers, who have to satisfy many different rules. Colombia, Kenya, and the Philippines learned lessons from the advanced economies and created a single SHI fund with one administration managing overall policy, strategy, actions, and subsidy pro- grams. Separate departments and financial accounts were set up to administer the different group plans under a single SHI administration. Some academics, however, argue that many competing funds, both public and private, would give people more choices of insurance benefits packages, thereby enhancing economic welfare (Zweifel and Pauly 2005). In addition, they argue that competition among insurers would promote efficiency in producing health ser- vices. These arguments overlook two facts: low-income countries lack the human resources and knowledge to start even one fund operation properly and a competi- tive insurance market becomes crippled by serious adverse selection and risk selec- tion. Even more important, advocates of multiple competing insurance funds have ignored the high sales and administrative costs associated with such an approach. The United States, a nation that relies on this approach, spends more than 25 per- cent of its total national health expenditure for administration, sales, and market- ing(Woolhandler,Campbell,andHimmelstein2003). Taiwan (China), with asingle, universal SHI, spends less than 5 percent (Lu and Hsiao 2003). As for the theory that competing insurance plans promote greater efficiency in producing health services, the evidence is mixed. Switzerland, a nation that used this strategy in the past, spent the highest percentage of its GDP (11.5 percent) on health than any other nation except the United States (OECD 2005). The United States, another country with competing insurers, did achieve some savings for a 32 Design and Implementation of Social Health Insurance decade under its managed care reform; however, it had a surplus of hospital beds, laboratories, and physicians that enabled competing managed care plans to bargain for lower prices and to control utilization, producing one-time savings (Light 1999; Sullivan 2000). Developing countries tend not to have an oversupply situation. Some developing nations have set up one fund to pool risks nationwide, but use multiple, competing insurance plans to serve as intermediaries for enrolling peo- ple and improving health services delivery. This approach is discussed in the next section. Improving the Delivery of Health Services The designers of SHI have to consider how to promote the efficiency and quality of health care under SHI. Essentially, they have three strategies to chose from: cre- ate competition, use rational payment methods, and decentralize. For each strategy, the designer has to consider the role of the private sector and how to create a level playing field between the public and private sectors to harness the private sector's resources and enterprising spirit. Private providers (charity and for-profit) play a major role in delivering health services and these resources have to be harnessed and managed for people's benefit. COMPETITION. Competition is not an end in itself, but a means to enhance the efficiency and quality of health services. It can be introduced at two levels: insurance or provider. Nations with a single fund can create competition at the insurance plan level by having the SHI agency pay a risk-adjusted premium to competing insurance plans. In theory, the agency will selectively contract the high- est-quality providers with the lowest charges (Enthoven 1988). Such contracting would put pressure on providers to improve their efficiency and quality. Colombia and Ghana follow this strategy. Colombia uses managed competition to establish a level playing field between public and private sector insurance plans. Ghana is initiating a system of community-based insurance plans and also allows private plans to compete, but the latter can only enroll better-off people. However, Ghana establishes a playing field between community-based and private insurance plans that is not level by subsidizing the former. At this early stage, we do not have suf- ficient evidence to assess the performance of the Ghanaian system. Competition can also be created directly at the provider level without competing insurance plans. Money follows patients in an SHI system. A single fund can com- bine three measures to establish competition among providers. The first measure offers the insured a choice of public and private providers and patients pay the same out-of-pocket payments. The second measure involves reforming the fee-for-service payment system to create better incentives for providers. The third measure reduces the direct subsidy given to public facilities so they have to compete for patients. SELECTION AND CONTRACTING. Instead of being a passive payer of claims, the SHI agency should be a collective purchaser on behalf of the insured. As a pru- dent purchaser, the SHI agency has to think and act strategically and intelligently when purchasing health services. It should focus on outcomes and outputs instead of inputs; it should rely on competition whenever possible to select providers and set prices; and it should establish what quality of services to buy, how to buy them, and from whom (Preker 2005). William C. Hsiao 33 A prudent purchaser will identify those providers whose health services are of good and bad quality and selectively contract with the former hospitals and clinics. To identify good and bad hospitals, the SHI agency may establish quality indicators and profile providers accordingly. For example, the indicators could include each hospital's risk-adjusted mortality rates, surgical complication rates, hospital infec- tion rates, and hospital readmission rates after surgery. Some studies show that profiling hospitals and making the information public has measurable effects on hospitals' performance (Hibbard, Stockard, and Tusler 2003). The SHI agency may exercise its power as a monopsony to counteract the superior market power of medical providers. It can negotiate with hospitals and physicians for lower rates and purchase drugs in bulk through public bidding. However, this monopsony power can be abused by paying too low a price, which could cause service quality to suffer. Hence rules to determine the reasonable costs of production for services have to be instituted. Alternatively, when several providers exist in the same service area, payment rates can be set by means of competitive bidding. PAYMENT. Payment methods and levels can have an important effect on the volume, quality, and cost-effectiveness of health services provided. Fee-for-service payment fosters supply-induced demand and health cost inflation. The capitation payment method encourages efficiency, as do any other package payment rates set prospectively, for example, for diagnostic-related groups (McGuire 2000). Pro- spectively set package rates give providers an incentive to become more efficient, because they can retain savings to use at their discretion. Thailand's capitation pay- mentsystemenhancedefficiencyinhealthcaredelivery.Incontrast,theSHIagency in the Philippines operates as a passive financial intermediary that pays claims on a fee-for-services basis, and evidence indicates that it has not improved the efficiency of health services delivery. Implementation Passing an SHI law and implementing it are two related, but distinct, affairs. Pass- ing a law may be easy, but effective implementation is a challenge. Often, in an attempt to escape difficult trade-offs and unpleasant compromises, laws prom- ise lofty goals, but the resources needed to carry them out are unavailable. Also laws may not incorporate a practical strategy for achieving their goals and may not consider the feasibility of implementing the laws. Even if the major questions are addressed during the design phase, the details still need to be worked out during implementation, when the state has to balance resources and commitments, plac- ing a higher value on practicality than ideals. Undertaking Administrative Organization and Reform All five nations discussed in this volume decided to place the SHI agency under the ministry of health as either a public (Colombia and Ghana) or quasi-public (Kenya, the Philippines, and Thailand) agency. The government will exert strong control and manage SHI for two reasons: the desire for political control and because it is the principal funder. Business, labor, and public representatives serve on the agen- cy's board, but the extent of their influence is unclear in most countries. 34 Design and Implementation of Social Health Insurance Implementing SHI effectively requires a capable and efficient organization. Corruption and patronage are major concerns when establishing the SHI agency. Health insurance funds can accumulate a large surplus. The Kenyan Hospital Insurance Fund offers a good example of how corrupt officials can divert a large insurance surplus into their own bank accounts. To prevent a recurrence, Kenya is trying to structure the board of the new National Hospital Insurance Fund so that most members represent employers, workers, and local governments. Colom- bia, the Philippines, and Thailand have similar structures. The minister of health serves as the board's chair, and the board reports to and is directly accountable to the president. Nevertheless, the president appoints the board members, thus cor- ruption remains as a possibility unless other checks and balances are put in place, such as independent outside audits, professional actuarial certification, and regular legislative oversight hearings. The mission and corporate culture of the SHI agency greatly influence who ben- efits from SHI. Insurance mobilizes money, but it has to be transformed into effec- tive and efficient services. The Philippines designated PhilHealth to be a financial intermediary, that is, to enroll people, collect premiums, and pay claims. Hence PhilHealth operates like a traditional private insurance company whose primary concernissolvency.PhilHealthestablishedfeeswithpaymentceilings,andonaver- age, PhilHealth's fee schedule covers only 30 to 40 percent of providers' charges, and patients have to pay the remainder. PhilHealth is solvent and has accumulated a large surplus, with US$1 billion in assets. As previously mentioned, a study found that more than 80 percent of PhilHealth payments became "profit" forhospitalsand clinics, while only 14 percent of was transformed into more or better health services for patients (Gertler and Solon 2002). In contrast, Thailand designated its National Health Security Office (NHSO) as a prudent purchaser of health services for the insured. The NHSO selects primary contractors, such as general and provincial hospitals, and pays them a capitation rate for most of the services covered. The NHSO also limits what providers can charge patients. A primary contractor organizes a network of providers by selec- tively contracting clinics and small hospitals to deliver prevention and primary care. Patients choose their primary contractor and obtain most of their health care from one network. Under its system, the NHSO improves the efficiency and quality of health services. Another administrative organizational implementation issue is deciding which agency should collect the SHI premium. The tax collection agency would be a logi- cal organization, but as Colombia discovered, this agency may not agree to handle this new responsibility. Another possibility is the social pension insurance agency, as SHI is often established in conjunction with a social pension plan. However, this agency may also not want to become involved in SHI, because pension contribu- tions can be as high as 15 to 25 percent of taxable wages, and adding another 4 per- cent or more for SHI would exacerbate evasion problems. SHI is a sophisticated and complicated financing method, therefore managing and operating an SHI system requires many well-trained administrators; financial, computer, and management information specialists; actuaries and accountants; and policy planners and evaluators. Building up the necessary human resources and computer and information systems takes years. Ghana encountered difficul- ties in implementing its SHI because of human resource problems. Colombia tried William C. Hsiao 35 to rely on private insurance companies and community-based organizations to enroll people, collect premiums, and pay claims, but still has a long way to go after a decade. In comparison, Thailand was able to achieve universal coverage because it had invested in its human resources for policy planning and analysis for almost two decades. Enrolling People, Collecting Premiums, and Targeting Subsidies Enrolling workers in the formal sector is relatively easy, but identifying the poor who will receive a subsidy and enrolling nonpoor informal workers are major challenges. CONTRIBUTORY REGIME. Worldwide experience shows that enrolling and col- lecting premiums from civil servants, employees of state enterprises, and workers employed in private companies with more than 10 employees is feasible. Evasion is a problem, but a modest one. However, evasion is a serious problem when enrolling workers employed by smaller private companies, as some companies may totally evade contributing, while others may underreport salaries or shift some portion of salaries into the allowance category. These latter behaviors reduce absolute salary levels that are taxed to fund SHI premiums. Colombia estimates that it collects only about 65 percent of the amount of premiums that should be collected. As for the portion of the population that can be enrolled by a mandatory con- tributory regime, Ghana and Kenya estimated that they could enroll 8 percent and 20 percent, respectively, of their populations. The Philippines has 40 percent of the population covered under its contributory plan, Colombia has 35 percent, and the figure is 20 percent for Thailand. SUBSIDIZED REGIME. Enrolling people who are fully subsidized poses no serious problem; however, if the eligibility criterion is based on income, then income test- ing is difficult to implement and often has many leakages. Colombia conducted a national income study, SISBEN, which investigated every household's income and assets periodically, then classified each family into an income category. It then used SISBEN categories to determine who is eligible for subsidy. Ghana, the Philippines, and Thailand all relied on local governments to assess who met the indigent crite- ria for subsidy, but these systems are susceptible to fraud. Thailand found fraudu- lent practices by local government officials who used subsidy certificates to reward their political supporters, friends, and relatives. Commitments have to be matched with resources. Low- and middle-income countries seldom commit enough resources to subsidize all the poor to enroll in SHI.Ghana,initsfirstyear,subsidizedcloseto21percentofitspopulationtoenroll, of which approximately 10 percent were children, 3 percent were elderly, and 9 per- cent were adult indigents, even though 50 percent of its population are classified as poor households (Atim and others 2001). The Philippines had subsidized almost all poor Filipinos in 2004 while the president ran an election campaign promising insurance coverage for the poor. A year later, however, only 20 percent of Filipi- nos were subsidized, even though 35 percent were classified as indigent. Colombia fully subsidized 30 percent of its population, while 50 percent of households were classified as poor. Universal coverage requires resources to subsidize the poor, but 36 Design and Implementation of Social Health Insurance unless the government allocates the necessary funds, universal coverage remains an illusion. NONPOORSELF-EMPLOYEDANDINFORMALSECTORWORKERS. Asidefromthenec- essary funding to help the poor, enrolling the nonpoor self-employed and informal sector workers poses the greatest barrier to universal coverage. These people usu- ally account for 25 to 40 percent of the total population in low- and middle-income countries. Nations usually try to have them enroll and pay voluntarily. Thailand tried many voluntary approaches using various types of health card schemes for 20 years and found that adverse selection posed serious problems: patients diag- nosed with chronic diseases and pregnant women eagerly enrolled, while healthier people did not. Abuse was also a serious issue: relatives and close friends borrowed insured people's cards to obtain free health care. Finally, in 2002, Thailand decided to use general revenues to cover all those who were not insured under insurance provided for civil servants and formal sector workers. Such an approach can cer- tainly achieve universal coverage, but as noted later, the Ministry of Finance is con- cerned about the future fiscal burden of maintaining this system. Ghana is trying a different approach to cover the nonpoor self-employed and informal sector workers by relying on community-based insurance plans at the district level to enroll them. However, Hsiao (2003) points out that people often do not trust the district-level schemes sufficiently to be willing to pay a premium. Moreover, district-level schemes do not engage active participation by people in managing the schemes and the benefits are not likely to be attractive. Atim and others (2001) find that district-level funds in Sub-Saharan Africa have done badly in attracting people to enroll, but that subdistrict funds do better, and that com- munity funds do even better. Like other countries, Colombia encountered serious difficulties in enrolling the nonpoor self-employed. Recently, the government tried to overcome the barrier by providing some subsidy for the near-poor and developed a less generous benefits package so that the premium contribution would be lower. In the Philippines, PhilHealth engages in innovative efforts to enroll nonpoor informal sector workers and the self-employed through organizations they might belong to, such as guilds, cooperatives, and microfinance organizations. Such an approach could reduce adverse selection and enrollment costs. How successful it can be in achieving universal coverage remains to be seen. Traveling the Long Road to Universality For a low-income nation to develop and expand an SHI program to cover all its citizens takes decades. Figure 2.1 compares the progress made toward universality by five nations. The dotted lines show the declared intentions of Ghana, Kenya, and the Philippines. Improving the Efficiency and Quality of Health Care Countries can employ four tools to improve the efficiency and quality of health care: contracting, creating proper incentives with the payment method and rate, restricting the essential drug list, and reducing the supply-side subsidy. William C. Hsiao 37 Figure 2.1. Progress Toward Universality, Selected Countries, 1990­2014 100 80 covered Thailand 60 Colombia population the 40 Kenya of Philippines Ghana 20 percentage 0 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016 year Source: Colombia: Ministry of Social Protection data; Ghana: 2003 National Health Insurance Act; Kenya: WHO and German Agency for Technical Cooperation 2004; Philippines: PhilHealth data; Thailand: WHO 2004. CONTRACTING. The SHI agency should select and contract providers based on quality of health care and prices; however, as a national program, it cannot set the standards so high that many existing providers are excluded, because this would lead to a shortage of suppliers. In addition, depriving suppliers of their livelihood and causing closures and layoffs would be politically difficult. As the five country cases illustrate, the SHI agency sets minimum quality standards and certifies which providers are qualified to serve SHI enrollees. SHI agencies usually start with existing government accreditation and certification standards and modify them. Most of the standards pertain to inputs such as practi- tioners' medical qualifications, hospital staffing ratios, and basic hygiene and safety conditions. Most developing nations do not set outcome standards. The Philippines seems to have done the most to accredit providers with measurably higher quality standards than the usual licensing and accreditation requirements set by the min- istry of health. Developing nations' SHI systems have not adopted quality enhance- ment approaches such as pay for performance, as the information and monitoring systems are too onerous. SHI agencies are able to encourage more efficiency and better quality through selective contracting than by merely setting standards. The contracting process sets the price the SHI agency will pay. When SHI uses bundled payment methods such as capitation, per admission, or diagnostic-related group to pay providers, those providers with higher costs would not want to contract with SHI unless they can bring down their costs to the payment level. Hence the high-cost providers are pressured to reduce their costs. Thai and many Colombian health plans use this approach effectively. In contrast, PhilHealth largely pay providers on a fee-for- service basis and allows providers to do balance billing. Such a contracting 38 Design and Implementation of Social Health Insurance approach has little impact on promoting greater efficiency and quality of health care. PAYMENT METHOD AND RATES. The payment system creates the incentive struc- ture that affects how health services are organized, their quantity and quality, and the production process. Ghana delegates decisions for payment methods and rates to district MHOs and private insurers, but the districts are unlikely to have the technical expertise and data to choose better payment methods and establish reasonable rates. The Philippines continues to pay providers on a fee-for-service basis for most health care, which encourages the inflation of health care costs and promotes supplier-induced demand. Colombia allows its numerous health plans to decide on their payment methods, and most have adopted either a capitation pay- ment method or a mix of capitation and fee-for-service methods. Thailand chose a capitation payment method and was able to improve the efficiency of its health care delivery system. DRUG LIST AND REGULATION OF DRUGS. The SHI benefits package has to spec- ify the health services and drugs that will be covered. Almost all SHI programs have specified a drug list for which the SHI will reimburse providers. The essential drugs have been fairly well defined, but not the drugs used in hospitals to treat complicated cases. New, expensive drugs emerge continuously, most of which have marginal benefits, however small. Most developing nations lack the procedures and the capacity to make rational decisions about which new, expensive drugs should be covered under SHI. In developing nations, pharmaceutical companies are given wide latitude in advertising to patients (Lexchin 1996). The companies also lobby the SHI agencies to include their products in the reimbursable drug list. None of the five country discussed in this volume could regulate direct advertising and lobbying activities. REDUCTION OF SUPPLY-SIDE SUBSIDIES. Under SHI, public facilities could receive double payments for patients they have treated. Public facilities obtain their rev- enues for providing medical services mostly from government budgets. Under SHI, these facilities will also receive payments from the insurance plan for the insured patients they have treated. Understandably, the public facilities would welcome the increase in revenues and would give various reasons why they need the addi- tional revenues. However, this would be a misallocation of a nation's resources: the additional revenues become economic rents for the staff or are used to invest in expensive, high-tech medical equipment that may benefit few patients. An obvious solution is to reduce the supply subsidy when public facilities increasingly receive revenues from the insurance plan over time (figure 2.2). Implementing such a policy can, however, encounter serious political difficul- ties. Colombia had planned to synchronize the reduction of public hospital sub- sidies with an expansion in the number of people being covered by SHI, that is, the savings from the supply subsidy were going to be shifted toward subsidizing more poor and near-poor people. However, the political influence of public hos- pitals and their labor unions have forestalled the reduction in the supply subsidy. Consequently, the government lacks the funding to expand its subsidy to the near- poor. Meanwhile, Colombian public facilities have dramatically increased their William C. Hsiao 39 Figure 2.2. Hypothetical Plan to Reduce Subsidies to Public Hospitals 100 evenuesr 80 payment by SHI 60 hospitals' public 40 government budget of 20 centage per 0 0 1 2 3 4 5 6 7 8 9 10 year Source: Author. staff compensation and invested in new capital (General Comptroller of the Repub- lic 2002). Similarly, the previously mentioned study (Gertler and Solon 2002) finds that the insurance payments made by PhilHealth had largely gone for economic rents for the providers and that only a modest portion had gone to improve access to better services. In short, when introducing SHI, companion measures must be taken to simultaneously reduce supply-side subsidies, otherwise patients may ben- efit little from the new SHI. Summary The health sectors of developing countries face several common problems: public underfunding of health care, inequity in financing, poor resource allocation, mis- management, and inadequate manpower. As a result, patients in most developing countries have to pay for large shares of their health expenses out-of-pocket (fig- ure 1.2), which creates an access barrier and impoverishes many families. At the same time, the affluent population uses a disproportionate share of public health services. Poor quality of health services and inefficiencies are prevalent in public health facilities. Clearly, many health systems need to be reformed. Many health planners see SHI as a magic solution to most of these problems, but the reality is quite different. Nonetheless, SHI does seem to have several strengths. It · can mobilize new, stable funds for health; · can pool risks widely and provide insurance protection; · enables the government to target new public funds to the poor; · has the potential to shift existing public resources to the poor by having for- mal sector employees and their families pay for the "free" public health ser- vices they currently use; 40 Design and Implementation of Social Health Insurance · canbeastrategytoreformthehealthcaredeliverysystemtoproducehigher- quality and more efficient services; · can rely on both public and private facilities for health services provision. The key question in designing SHI is how to match high-minded goals with scarce resources. Subsidizing premiums for the poor and near-poor requires large additional government funds. Most low-income countries do not have sufficient tax revenues and are unable to reallocate enough funding from other sectors to the health sector. Donor funds can cushion the costs of SHI during the transition period, but they are not a stable and sustainable long-term funding source. SHI is a financing instrument that can reduce financial barriers for people to access health care; however, the physical supply may not be available. Most devel- oping nations have a two- or three-tiered health care system, and the supply of health care for people in the bottom tiers must be improved if they are to have equal access to reasonable health care. Currently, most rural residents do not have primary care providers located nearby, drug distribution is sparse, and counter- feit drugs are prevalent. Governments have to invest in new facilities and human resources in underserved areas. To assure good quality and efficient health care and protect the insured, the SHI agency has to select and contract qualified public and private providers; establish rational payment systems; prevent providers from using their market power to gouge patients; and coordinate prevention, primary care, and tertiary health ser- vices. These roles and responsibilities require a sound organization and sophisti- cated knowledge and skills. Building up the organization, human resources, and information systems to manage and administer SHI well takes years. For SHI to become universal takes decades. As a result, countries must operate a dual system during the years of development, when a significant portion of the population will not be insured. These people must be provided with a health safety net to protect them from impoverishment and assure them some access to basic health care. This implies that the government has to institute a public assistance program or continue to provide somewhat free health services to the uninsured. Designing and operating such a dual system is a complex and difficult affair. In addition, developing nations must have rapid economic growth to attain universal coverage within a few decades. References Atim, Cris, Steven Grey, Patrick Apoya, Sylvia Anie, and Moses Aikens. 2001. A Survey of Health Financing Schemes in Ghana. Bethesda, MD: Partners for Health Reform, Abt Asso- ciates. http://www.phrplus.org/Pubs/Tech013_fin.pdf. BitranandAssociates,Econometrita,andSuperiorSchoolofPublicHealth(BitrányAsociados- Econometría and Escuela Superior de Administracion Publication). 2002. Propuestas de reestructuración de los procesos, estrategias y organismos públicos y privados encargados de la afiliación, pago y recaudo de aportes al SGSSS. Health System Reform Support Program of the Ministry of Social Protection (Programa de Apoyo a la Reforma. Proyecto 03-99). Bogota: Ministry of Health. Busse, R., and A. Riesberg. 2004. Health Care Systems in Transition: Germany. Copenhagen: European Observatory on Health Care Systems. Enthoven,A.C.1988."ManagedCompetition:AnAgendaforAction."HealthAffairs7(3):24­47. William C. Hsiao 41 General Comptroller of the Republic (Contraloría General de la República). 2002. Situación de la Red Hospitalaria Nacional. Fiscal Management Financial Report 1. Bogota: General Comptroller of the Republic. Gertler, P., and O. Solon. 2002. "Who Benefits from Social Health Insurance in Developing Countries?" Working paper, University of California Berkeley. Gold, M., J. Siegel, L. Russell, and M. Weinstein. 1996. Cost-Effectiveness in Health and Medi- cine. New York: Oxford University Press. Gruber, J., and A. Krueger. 1991. "The Incidence of Mandated Employer-Provided Insurance: Lessons from Workers' Compensation Insurance." In Tax Policy and the Economy, vol. 5, ed. D. Bradford, 111­44. Cambridge, MA: MIT Press. Hibbard, J. H., J. Stockard, and M. Tusler. 2003. "Does Publicizing Hospital Performance Stimulate Quality Improvement Efforts?" Health Affairs 22 (2): 84­94. Hsiao, W. C. 2003. "Report on a Feasibility Study of Social Insurance," Ministry of Health, Uganda. Unpublished. Kwon, S. 2003. "Healthcare Financing Reform and the New Single Payer System in the Republic of Korea: Social Solidarity or Efficiency?" International Social Security Review 56 (1): 75­94. Lexchin, J. 1996. Deception by Design: Pharmaceutical Promotion in the Third World. Penang, Malaysia: Consumers International. Light, D. W. 1999. "Good Managed Care Needs Universal Health Insurance." Annals of Inter- nal Medicine 130 (8): 686­89. Lu, J. F., and W. C. Hsiao. 2003. "Does Universal Health Insurance Make Health Care Unaf- fordable? Lessons from Taiwan." Health Affairs (Millwood) 22 (3): 77­88. McGuire, T. G. 2000. "Physician Agency." In The Handbook of Health Economics, vol. 1A, ed. A. J. Culyer and J. P. Newhouse, 461­536. Amsterdam: Elsevier. OECD (Organisation for Economic Co-operation and Development). 2005. OECD Health Database. http://www.oecd.org/topicstatsportal/0,2647,en_2825_495642_1_1_1_1_1,00. html. Preker, A. 2005. Spending Wisely: Buying Health Services for the Poor. Washington, DC.: World Bank. Savedoff, W. 2005. "Private Health Insurance: Implications for Developing Countries." Bul- letin of the World Health Organization. 83 (2): 127­34. Sullivan, K. 2000. "On the `Efficiency' of Managed Care Plans." Health Affairs 19 (4): 139­48. WHO (World Health Organization). 2004. Regional Overview of Social Health Insurance in Southeast Asia. New Delhi: Regional Office for Southeast Asia. WHO and German Agency for Technical Cooperation. 2004. National Social Health Insurance Financial Projections and Future Bilateral/Multilateral Cooperation. Joint WHO-GTZ Mission to Kenya. Mission 6, June 21­26. Nairobi: WHO Country Office. Woolhandler, S., T. Campbell, and D. U. Himmelstein. 2003. "Costs of Health Care Adminis- tration in the United States and Canada." New England Journal of Medicine 349 (8): 768­75. Yu, S.-H., and G. F. Anderson. 1992. "Achieving Universal Health Insurance in Korea: A Model for Other Developing Countries?" Health Policy and Planning 20 (3): 289­99. Zweifel, P., and M. Pauly. 2005. "Private Voluntary Health Insurance in Developing Coun- tries." Presentation at the "Wharton Impact Conference on Private Voluntary Health Insurance in Developing Countries," University of Pennsylvania, Philadelphia, March 15­16. 3 Kenya: Designing Social Health Insurance Andrew Fraker and William C. Hsiao Kenya is a low-income country in Sub-Saharan Africa. It currently has an SHI program, but it just covers hospital expenses and only one-fifth of the population is enrolled. This case study examines the design and implementation issues of Kenya's proposed National Social Health Insurance Fund (NSHIF), which has been sidelined because of financial sustainability concerns. The proposed scheme would offer comprehensive benefits, and the government would eventually attempt to extend coverage to all Kenyans. Background Kenya lies on the equator in East Africa, bordered by Somalia, Ethiopia, Sudan, Uganda, Tanzania, and the Indian Ocean (figure 3.1). Formerly part of British East Africa, Kenya gained independence as a republic in 1963. Table 3.1 provides basic statistics regarding Kenya's demography, economy, health status, and health system. Economy Almost 80 percent of Kenyans live in rural areas, working mostly as farmers. The average income in Kenya is higher than in neighboring Ethiopia, Somalia, and Tan- zania, but lower than in Sudan and Uganda (World Bank 2006). Half the population lives below the national poverty line. Kenya is one of the most corrupt countries in the world, which makes health system reforms at the national level challenging, because people are afraid to let government officials manage their prepayments. Health Life expectancy and infant mortality are slightly better in Kenya than in the rest of Sub-Saharan Africa, but both have worsened in the past two decades. Health out- comes had improved dramatically since the end of colonial rule, but life expectancy is now back to the same level it was in 1962. Communicable diseases cause most ill- nesses and deaths. About one-third of outpatient visits are related to malaria (WHO 2002). 43 44 Kenya: Designing Social Health Insurance Figure 3.1. Map of Kenya Source: World Bank. Health Financing and Expenditure At $50 per person per year, health spending is lower in Kenya than in any of the other country cases in this volume. The low level of funding employs few doctors: about 1 for every 7,000 people. In addition, risk pooling is insufficient and out- Andrew Fraker and William C. Hsiao 45 Table 3.1. Background Statistics, Kenya Category Statistic Source Demographics (2005) Population (millions) 34.3 World Bank 2006 Growth rate (%) 2.3 World Bank 2006 Urban population (%) 20.7 World Bank 2006 Economy GDP per capita, purchasing power parity (2005) 1,165.0 World Bank 2006 GDP annual growth rate, per capita, real purchasing power parity (past decade) ­0.2 World Bank 2006 Percentage of the population below the national poverty line (2000) 50.0 CIA 2006 Percentage of the population living on less than $2/day (1997) 56.1 World Bank 2006 Percentage of the population living on less than $1/day (1997) 22.8 World Bank 2006 Agricultural employment (percentage of the population) 75.0 World Bank 2006 Corruption (rank out of 163 countries, 2005)a 142 Transparency International 2006 Health status Life expectancy (years) (2004) 48.3 World Bank 2006 Infant mortality/1,000 live births (2004) 78.5 World Bank 2006 HIV/AIDS prevalence (percentage of the population age 15­49) (2005) 6.1 World Bank 2006 Health system Health spending (percentage of GDP) (2003) 4.3 World Bank 2006 Health spending (US$ per capita, purchasing power parity) (2003) 50 World Bank 2006 Physicians/1,000 population (2004) 0.14 World Bank 2006 Hospital beds/1,000 population -- Health System financing (percentage of total health expenditure) (2003) SHI 9.1 WHO 2006 General government revenue 30.0 WHO 2006 Out-of-pocket 47.4 WHO 2006 Private insurance 10.9 WHO 2006 Other private 2.6 WHO 2006 of-pocket spending accounts for nearly half of total health expenditure. General government revenue is the next largest share of health financing, and is used to subsidize public providers. Public and Private Provision of Health Services Publicly owned facilities provide services and goods accounting for 60 percent of total health expenditures, with privately owned facilities accounting for the 46 Kenya: Designing Social Health Insurance remainder.1 Of total health expenditures, 55.3 percent were for care given at hospi- tals, with clinics and dispensaries accounting for 35.1 percent of total health expen- ditures (table 3.2). Health Insurance Kenya has had compulsory SHI for hospital services for more than 40 years; how- ever, only about 7 million Kenyans, or 20 percent of the population, are covered, namely, those employed by the government or by some large formal sector employ- ers. The National Hospital Insurance Fund (NHIF) legislation was passed in 1998, and requires employees with incomes above a certain threshold to make monthly contributions to the fund. The NHIF suffers from poor management and corruption: only 22 percent of the fund is actually used to pay for benefits, with 25 percent going to administra- tive costs and 53 percent to investment projects, such as lavish new headquarters (Kenya Private Sector Alliance 2004). A large portion of the fund's accumulated reserves was lost because of corruption. The NHIF was intended to cover all hos- pital costs for members, but only pays for the cost of staying in a hospital, not the costs of medical treatments or drugs, thus NHIF beneficiaries must pay all of these costs out-of-pocket. In 2001, the NHIF only financed 4.1 percent of total health expenditures. Contributions are deducted from the paychecks of workers in the formal sec- tor. The contribution is based only on the salary portion of workers' incomes and does not take into account "allowances," which can make up half or more of most government employees' incomes. The 23 branches of the NHIF are responsible for identifying employers and employees and enforcing contributions. As of August 2003, only one of the branches had a computer network link to the central database, Table 3.2. Shares of Health Expenditures by Provider, 2000­2002 Provider Percent Hospitals 55.3 Public 38.8 Private for-profit 11.9 Private nonprofit 3.1 Mental health 1.4 Public specialty 0.1 Private specialty 0 Dispensaries and clinics 35.1 Private clinics 10.5 Public dispensaries and clinics 10.1 Private drug stores 7.4 Traditional healers 1.7 Private nonprofit dispensaries and clinics 5.4 Other 9.6 1. One percent of expenditures went to providers whose ownership was not specified in government accounts. Andrew Fraker and William C. Hsiao 47 and none received information from the tax authorities, with which all businesses must register. Reviewing claims usually means sending someone to the hospital to make sure that a patient is actually in the bed. The claims are then sent to the NHIF's headquarters, which remits payment back to the regional hospital. This is usually a slow process. Design of the NSHIF In June 2004, the government proposed the NSHIF to give people access to high- quality hospital care and to pool risks among the rich and the poor, the young and the old, and the healthy and the sick. Everyone with an income above a certain amount will make compulsory contributions and every citizen will receive hospi- tal care without paying user fees. The current system depends heavily on out-of-pocket payments. Even public hospitals require user fees, and few of the poor receive waivers. This means that sick people often forgo care or become burdened with debt. The new system seeks to convert out-of-pocket payments into prepaid funding. Financing Requirements and Sustainability The NSHIF provides clear directives for coverage, benefits, and expected sources of revenue. COVERAGE. The NSHIF Bill calls for insurance coverage for inpatient and outpa- tient hospital care for all Kenyans and residents, regardless of their age, health, eco- nomic status, or social status. While everyone will be legally obligated to register for membership, the government acknowledges that enrolling the entire popula- tion will not be easy. Its gradual implementation scenario projects that covering all formal sector employees will take five years and covering the large self-employed group, consisting mainly of poor informal economy workers and farmers, will take nine years (table 3.3). BENEFITS PACKAGE. The NSHIF will cover both inpatient and outpatient hos- pital services. The comprehensive package includes surgical, medical, and dental procedures; laboratory and diagnostic tests; drugs and medical equipment; physio- therapy; doctors' fees; and room and board (Republic of Kenya 2004). The NSHIF's board of directors will decide on specifics within these categories, and substantial omissions might be necessary to make the package affordable. Preventive and pro- motive services and drugs, such as the national vaccination programs, are not cov- Table 3.3. The NSHIF Gradual Implementation Scenario, 2004­12 (percent) Category of employees 2004 2005 2006 2007 2008 2009 2010 2011 2012 Self-employed 25 30 40 50 60 70 80 90 100 Formal sector employees 80 85 90 95 100 100 100 100 100 48 Kenya: Designing Social Health Insurance ered by the NSHIF, but are provided by the Ministry of Health. Anything outside the benefits package and not provided by the Ministry of Health can be purchased out-of-pocket or with prepaid private insurance. SUPPLEMENTARY INSURANCE. The private sector will be allowed to offer supple- mentary insurance to high-income earners who want coverage beyond the NSHIF benefits package. For example, the government anticipates that some people will want supplementary insurance to cover more comfortable hospital accommodations. SUBSIDIZATION. The government has not defined who is too poor to pay the pre- mium or how it will determine this. With such a large proportion of poor people, the primary challenge to universal insurance is raising the funds to pay insurance premiums for the poor. Deciding who to subsidize is therefore an essential part of the design of SHI, and the Ministry of Health is trying to delay this step until the implementation phase. REVENUES AND EXPENDITURES. The plan is to fund the NSHIF from five sources: contributions by private employers and employees, contributions by the self- employed, payroll deductions for civil servants, general government revenues for the poor, and other sources such as donations. The bill does not specify the contri- bution rate for formal private sector employees. Originally the government plan was to collect contributions from employers and employees, with employers cover- ing double what employees have to pay, but more recent information indicates that the split will be 50:50. The actual rates will not be defined until the bill becomes law and the fund's board makes a decision. Paying the premiums of the poor will be so difficult that the government is apparently willing to use any of the NSHIF funding sources for this. A dedicated value added tax or contributions out of general government revenues will not be enough. Some of the paying self-employed will also need to be cross-subsidized by formal sector employees, as their expected premiums of K Sh 400 to K Sh 450 per year will not be enough to pay even for the average number of outpatient visits at the lowest provider level. Balancing the fund will be a major challenge. To estimate the costs of the NSHIF, the government first attempted to determine the costs of different services, basing its figures on private sector charges or on NHIF payments plus user fees. The first method yielded estimates of K Sh 4,000 per inpatient stay and K Sh 310 per dispensary consultation. Outpatient visits cost K Sh 410 each using the first method, but only K Sh 100 using the second, so the govern- ment decided that a payment of K Sh 150 to K Sh 200 per visit should be sufficient. The second method was used for district hospitals and yielded a cost of about K Sh 2,300 per inpatient stay. The government acknowledges the need to discuss pay- ment levels with providers. It also estimated usage rates for each level of service. Based on existing enrollment assumptions as well as cost and utilization esti- mates, the government projected surpluses for the first five of seven years of opera- tion (table 3.4). Subsidization of Public Hospitals The NSHIF will replace government subsidies as public hospitals' primary income source. In 2001, K Sh 10.3 billion of the public hospitals' K Sh 18.2 billion in revenues Andrew Fraker and William C. Hsiao 49 Table 3.4. Projected NSHIF Expenditures and Revenues, 2004­10 Category 2004 2005 2006 2007 2008 2009 2010 Expenditures (K Sh billions) 10.86 19.42 25.95 33.55 42.37 51.89 62.84 Administration and reserves as a percentage of expenditures 7 7 7 7 7 7 7 Health care costs as a percentage of expenditures 93 93 93 93 93 93 93 Revenues (K Sh billions) 24.26 31.94 36.13 40.81 46.05 50.45 55.30 Contributions as a percentage of revenues 51 60 62 64 66 67 67 Government subsidies as a percentage of revenues 45 37 35 33 31 31 30 Donor contributions as a percentage of revenues 4 3 3 3 3 3 3 Balance (K Sh billions) 13.40 12.52 10.18 7.26 3.68 ­1.44 ­7.54 Balance as a percentage of expenditures 123 64 39 22 9 ­3 ­12 (56 percent) came from the Ministry of Health. During the transition to the NSHIF, the ministry will continue to cover hospitals' personnel and infrastructure costs. National hospitals are supposed to become financially autonomous from the min- istry within 2 years, provincial hospitals within 4 years, district hospitals gradually over 6 to 10 years, and health centers and dispensaries within 10 years. After that, the NSHIF payments will have to cover investment and personnel costs, or else the hospitals will become underfunded. The government has not yet addressed this issue, and tight budgets will lead to fighting over responsibilities if no clear plan for the transition is forthcoming. Administrative Structure To overcome the corruption and administrative shortcomings of the NHIF, the NSHIF is to be an independent, autonomous, statutory body with a corporate per- sonality, with an administrative structure as shown in figure 3.2; however the Min- istry of Health will have some oversight. In the extreme case that the minister of health decides that board members are not operating in the best interests of fund members or are not conforming to the rules laid down by the NSHIF Bill, the min- ister can take over operations. The decision makers will be the approximately 20 members of the Board of Trustees, including a chair appointed by the president and a CEO appointed by the minister of health. The composition of the board members is designed to reduce the potential for corruption. Each of Kenya's eight provinces will elect a member, pref- erably with a background in medicine, finance, institutional management, or law. Five board members will come from the government, namely, the permanent secre- tary of the Ministry of Finance, the permanent secretary of the Ministry of Health, the attorney general, the permanent secretary or director of personnel manage- ment, and the director of medical services. Five to seven members will come from 50 Kenya: Designing Social Health Insurance Figure 3.2. Administrative Structure of the NSHIF Board of Trustees Investigation/ Chief Executive Public Relations Antifraud/ Officer and Education Theft Unit Information Finance and Human Quality and Management Administration Resources Standards Internal Legal Enforcement Enforcement Research Audit Services and and and Compliance Compliance Development Source: Ministry of Health 2004. the private sector, of which the Federation of Kenyan Employers will nominate one, the Central Organization of Trade Unions will nominate one, the Kenya National Union of Teachers will nominate one, one will come from the Association of Kenyan Insurers, one will represent nonprofit health care providers, one will come from the Kenya Medical Association, and one will come from the NGO Council. Of the last four members, the minister of health will appoint two. The board has four important tasks to complete before the insurance program can begin. First, it must define the benefits package, including the drug formulary. This needs to be completed in conjunction with its second task, determining the level of member contributions to finance the expected costs. The board will need to make difficult decisions about what to omit from the benefits package to make it afford- able. The third and fourth tasks to be completed before the program can begin are to certify and contract health care providers. The board's operational duties will include receiving contributions, presenting the minister of health with revenue and expendi- ture estimates for approval, making payments out of the fund, managing the fund, ensuring compliance by providers, and determining changing health care needs. The Office of the CEO will run the day-to-day operations of the fund. Each of the four departments under the Office of the CEO will be decentralized to districts and lower levels if necessary. Initially, most operations will be centralized, but over time responsibilities could be devolved to districts. One idea is to allow premium collection and provider payment at the local level to reduce the length of time pro- viders must wait for payment. The existing NHIF is corrupt, thus the major concern is to structure the NSHIF so as to prevent corruption and convince the public that people's contributions will go toward providing health care, not fancy administrative buildings or board members' Swiss bank accounts. The government acknowledges that fraud within the fund is seen as one of the biggest concerns that could greatly undermine the sustainability of the fund. To allay the public's concerns, the NSHIF is designed with the following safeguards to prevent corruption: Andrew Fraker and William C. Hsiao 51 · Election of board members. Of the 20 board members, 8 will be elected and will have to run for reelection after three years, making them publicly account- able. They can only hold office for two terms. However, people have already raised concerns that they will not be adequately represented, as 60 percent of the members will be appointed, including the two most powerful members: the chair and the CEO, even though the original bill called for the CEO to be elected by the board. · Rules to constrain the board's behavior. When discussing investments or con- tracts in which a board member has a financial interest, he or she must declare that interest, not participate in the discussion, and abstain from vot- ing. Another rule to constrain corruption is the 5 percent limit on admin- istrative costs and the 3 percent limit on reserves: reducing the size of accumulated assets reduces the potential for corruption. The types of invest- ments the board is allowed to make are also explicitly defined. · Investigation/Antifraud/Theft Unit. One thing that stands out in figure 3.2 is the prominence of this unit, which will be responsible for all types of fraud, including insurance fraud, which is of particular concern to the government. Fraud can occur at three levels: internally by fund management, including administrators, the CEO, and the board; externally by the contracted service providers; and externally by the beneficiaries. The unit reports to the CEO and can investigate any matter brought to its attention by any other divi- sion or by an individual. If investigating the Office of the CEO, it will report directly to the board. How independent this unit will be administratively or financially is not clear. Apparently it will not be able to investigate fraud at the board level, and a completely independent unit that reports to the Minis- try of Health and legislators might be necessary. · Internal Audit Unit. This unit falls under the auspices of the Finance and Administration Department, but is free to report directly to the CEO when necessary. · Enforcement and compliance divisions. Both the Finance and Administration and the Quality and Standards departments have such divisions to prevent fraud by contracted providers and by beneficiaries. Section 14 of the bill allows enforcement officers appointed by the board to enter any contracted medical facilities or place where such officers have "reasonable grounds to believe that any persons are employed or self-employed." Officers have the right to question anyone and demand documents. Anyone failing to answer questions or provide documents or obstructing entrance is subject to a fine of up to K Sh 50,000 or up to three years in prison. Officers can use these pow- ers to investigate many types of fraud, including falsified reimbursement claims, falsified membership cards, provision of health services of a quality below the contracted level, and failure to make contributions for employees. However, Kenya has a weak judicial system, so the risk of wrongdoers actu- ally being punished could be low. Payment Methods and Rates The NSHIF Bill specifies that the program will contract with providers at various levels, including dispensaries, private ambulatory care (curative), health centers (mostly outpatient), district hospitals, provincial hospitals (secondary), and national 52 Kenya: Designing Social Health Insurance hospitals (tertiary). The minister of health declared that the fund will not cover care in the 11 private hospitals that cater to the rich, but will contract with all other pro- viders that meet the Bill's quality criteria. All possible payment methods are on the table and are subject to negotiation between the board and providers. The expec- tation is that fee-for-service will be used initially until cost accounting measures improve. Rates are also subject to consultation and negotiation with providers. Implementation Issues The NSHIF Bill passed parliament in June 2004 in a record 30 minutes, but after the finance minister declared that the NSHIF was not financially feasible, President Mwai Kibaki refused to sign the bill into law (Maliti 2005). President Kibaki sent it back to parliament, wanting it to pass a bill with a plan for phased implementation. Interest groups that had had little input into the drafting of the bill are now trying to shape it to their advantage. The issues discussed in the following subsections are strategies for implementing the program and are constantly evolving. Working with the Ministry of Health The NSHIF is set up to be independent from the government, but the government is not comfortable with sacrificing all control, thus the NSHIF must coordinate with the Ministry of Health and, in some cases, answer to it. The Health Insurance Act will empower the ministry to regulate, supervise, and coordinate all health insur- ance schemes, including the NSHIF. The NSHIF Bill gives the ministry the power to take any necessary actions if the board is not complying with the law or otherwise not acting in beneficiaries' best interests. The ministry will also need to regulate traditional medicine and eliminate health maintenance organizations, which will have to turn into insurers or health care providers, but not both. For health services covered by the NSHIF to reach most of the population, new infrastructure is required and existing infrastructure needs to be improved. The Ministry of Health will need to ensure the availability of facilities in each prov- ince and district, possibly using temporary mobile care facilities in rural areas. In addition to assessing infrastructure needs, the ministry will accredit and regulate facilities, including the necessary level of education of various types of employees. It will also determine which services from the benefits package will be available at different provider levels, as well as the referral system between levels. Figuring Out Enrollment and Premium Collection Three groups pose three different sets of challenges to implementing the NSHIF, namely, formal sector employees, informal sector workers who can afford to pay, and informal sector workers who are too poor to pay. FORMAL SECTOR. The government predicts that in the first year it will enroll 80 percent of formal sector workers, including both private and public sector employ- ees. In Nairobi, most workers are already enrolled in the NHIF, but compliance is much lower outside the capital, and the NSHIF will have to share information with the tax authorities more extensively than the NHIF ever did. To improve compli- ance, as noted earlier, the bill provides for enforcement officials who can legally Andrew Fraker and William C. Hsiao 53 enter any suspected place of employment to ensure that all employees are enrolled and that both employers and employees are making regular contributions. Employers might try to rehire their salaried employees as contractors to avoid paying NSHIF contributions. To avoid this, the NSHIF will need to define the dif- ference between contractors and employees, or perhaps try to set contributions by the self-employed equal to the sum of employer and employee contributions to eliminate the financial incentive to shift compensation. INFORMAL SECTOR. Most workers are not employed in the formal sector. Some of these will need partial subsidization and a large proportion will need to be fully subsidized, but many, especially self-employed professionals, will be able to pay the full premium. The government will need to contract existing organizations that might be more efficient than NSHIF branch offices in dividing the self-employed into those who can pay and those who are too poor, and then enrolling them. To this end, the government has compiled a list of organizations that might be helpful in identifying and enrolling these populations, including cooperatives, artisans' associations, women's and youth groups, occupational groups, village post offices and banks, utility companies, churches, community-based organizations, and other NGOs. The government will also need to identify poor formal sector workers and the unemployed poor. Identifying which informal workers can pay the NSHIF premium will be dif- ficult. Most of the poor are farmers or are otherwise self-employed.2 Many of these receive a large part of their incomes from subsistence farming, making it difficult to determine their incomes and whether they can afford the premium. The gov- ernment hopes to raise about 25 percent of the necessary funding from the self- employed. Its revenue estimates assume that 75 percent of the self-employed will pay their own premiums; however, a major hurdle to implementing the NSHIF will be dividing the self-employed into those families that can pay their own premiums and those that the government will have to pay for. Realizing that enrolling the informal sector, especially in rural areas, will take a long time, the Ministry of Health decided to provide care at dispensaries and medi- cal centers (which provide outpatient and limited inpatient care) for free, aside from a small registration fee. This will cost the government K Sh 4.1 billion over its seven- year projection (WHO and German Agency for Technical Cooperation 2004). This is intended to be an interim measure to provide care for the poor until they become members of the NSHIF, but it might actually slow enrollment in the NSHIF, as those in the informal economy who might have been willing to pay premiums can now meet most of their health care needs for free, making NSHIF membership less attrac- tive. This issue could be resolved if the government enforced NSHIF membership: those receiving free care could be registered and subjected to a means test. Assuming that the NSHIF is able to separate informal workers into those who can pay and those who cannot, means that testing will be required of many local employees, which will add to administrative costs. If President Kibaki ever signs the NSHIF Bill into law, it will most likely be phased in, starting with public and private formal sectors workers, followed by informal workers on a voluntary basis. As enrollment will be voluntary, it will be subject to adverse selection, because only those most in need of care will enroll initially. 2. The self-employed include farmers. 54 Kenya: Designing Social Health Insurance Another concern is corruption in relation to issuing membership certificates. Millions of people are not poor enough to qualify for full subsidization, but the K Sh 400 per head per year premium will present a huge burden for them. This cre- ates an immense opportunity for corruption in the form of bribery. Also preventing people from forging membership certificates may be problematic, especially as the NSHIF will not be able to afford sophisticated identification cards. Furthermore, poverty is dynamic, in that people pass in and out of poverty all the time, there- fore the NSHIF will need a system whereby it can frequently reassess who needs subsidization. Paying for the Poor: Costs and Financing Table 3.5 shows the financing sources for total health care expenditure in 2001 and how these sources would change under a fully implemented NSHIF. As the table indicates, the NSHIF would drastically alter the sources of financing by reducing out-of-pocket payments and increasing public funding. In 2001, more than half of health expenditure was paid out-of-pocket. Table 3.6 breaks down the sources of funding for the NSHIF. Table 3.5. Sources of Total Health Expenditures, Actual in 2001 and Projected under the NSHIF (percentage of total) Source 2001 With the NSHIF Private 65 22 Out-of-pocket 53 13 Employers 6 4 Private insurance 4 3 Nonprofit institutions 2 2 Public 35 78 Tax revenues 21 21 Public worker benefits 10 n. a. NHIF 4 n. a. NSHIF n. a. 57 Table 3.6. Sources of NSHIF Funding Source K Sh billions Share (%) Contributions by formal sector employees and employers 12 29 Earmarked value added tax 11 27 Contributions by the self-employed 10 24 Previous public worker benefitsa 7 17 Other (donations, etc.) 1 3 Total 41 100 Source: Ministry of Health 2004. Note: The total is equivalent to about US$500 million. a. These are monies previously collected to provide public worker benefits that are transferred over. Andrew Fraker and William C. Hsiao 55 All the sources shown in table 3.6 are problematic. The original plan was for employers to contribute twice the amount that employees would contribute, that is, K Sh 8 billion and K Sh 4 billion per year, respectively. The NHIF only raised K Sh 2.8 billion in 2001, almost exclusively from employee contributions. There- fore the NSHIF represents a large new tax on employers, and also forbids them from lowering salaries to finance it. Raising the required K Sh 12 billion in the face of strong industrial opposition will be politically difficult. The government has already given up on an earmarked value added tax, instead favoring a general commitment to finance contributions for the poor. As the poor live mostly in rural areas and will take longer to enroll, this is not a problem in the short run. However, future governments will find it hard to increase existing taxes or introduce new taxes, and overfunding the system now and investing the surplus to fund projected future deficits might be preferable, but fears of corruption led to the prohibition of large surpluses, thus the fund's reserves are not allowed to exceed three months of projected needs. As mentioned previously, collecting contributions from the self-employed will be difficult and identifying who can afford the annual premium will be challenging. As for the actual amount to charge, a task force came up with a list of suggested con- tributions by province that ranged from annual individual payments of K Sh 1,175 for the Coast province to K Sh 200 for Nyanza, with an estimated annual average of K Sh 425. Whether the government intends to impose different rates in different regions of the country, which would be more equitable, or if it intends to round the national average contribution to K Sh 400 or K Sh 450 (about US$5 or US$6) and charge all the self-employed the same rate, which would be administratively easier, is not clear. The government also needs to establish contribution rates for self-employed professionals earning high incomes. The payroll harmonization has struck a dissonant chord with the Kenya National Union of Teachers and other civil servants. The union objects to convert- ing members' medical allowances to mandatory contributions. These would be included as part of the "previous public worker benefits" shown in Table 3.6 that would be shifted to SHI. Apparently teachers were paid medical allowances regard- less of whether or not they needed care. Another important financing issue was whether contributions should be per head or per family. Both methods could be structured to raise the same amount of funds, but the per head system was chosen because it would help eliminate pos- sibilities for fraud. Regarding contributions for dependents, one suggestion is to reduce the payments for dependents of the self-employed from K Sh 400 per year to K Sh 100. The hope is that this would convince more self-employed near the pov- erty line to enroll, and would only cost the government K Sh 0.7 billion to K Sh 1.3 billion in 2006, assuming that 45 percent of the self-employed enrolled (WHO and German Agency for Technical Cooperation 2004). FINANCING THE POOR. Kenya's demographics present a major challenge to a financially sustainable SHI program. Kenya has a young population: 66 percent are dependents, many of whom probably work in the informal economy. Of the 34 percent of the population that is working or retired, 83 percent are self-employed. Only 10 percent work for formal employers registered with the government and 7 percent are public employees. 56 Kenya: Designing Social Health Insurance The national definition of poverty places about half the population below the poverty line. The poor will not be able to pay their own premiums, so originally the government planned to pay for them out of earmarked taxes; however its pro- jections showed that only 27 percent of revenues would have come from the value added tax. This means that either some of the poor will have to make contributions, which is unreasonable or, more likely, the contributions of formal sector employees will have to cross-subsidize the poor. FEASIBILITY. The government's projections presented earlier relied on some key assumptions that make them unrealistic. Even under its optimistic assumptions, the system runs large deficits in the long term. Perhaps the faultiest assumption is that 75 percent of the self-employed will be able to pay K Sh 450 per year for themselves and the same amount for each dependent. Another crucial assumption is that the government will be able to negotiate low rates with providers, such that limited SHI funds can be stretched further. Table 3.7 presents two enrollment scenarios, moderate and rapid. Calculations were made to determine the amount the government will have to contribute to make the fund balance, given three different levels of contributions by employees and employers (both private and public). The possible contributions by employers and employees as a percentage of salary are as follows: 6.00 percent/3.00 percent, 4.66 percent/2.33 percent, and 2.90 percent/2.90 percent. The system cannot depend on the nonpoor self-employed making substantial contributions in the fund's early years. Assumptions were made about the average costs of inpatient and outpatient treatments at primary, secondary, and tertiary level of care. However, what rates the NHSIF will negotiate with providers at different levels remains to be seen, and much of this will depend on how quickly providers become financially autono- mous from the Ministry of Health, which is supposed to cover personnel and infra- structure costs during the transition. Other assumptions behind the calculations include three outpatient visits per person per year and 0.26 inpatient days per per- son per year, but actual numbers could be much higher, because most people were previously uninsured and about 25 percent did not seek health care for financial reasons. Administrative costs were assumed to fall to 5 percent and investments to 3 percent within five years, as written into the bill, but these figures are extremely optimistic. Table 3.8 presents the required government contribution given three different levels of combined employee and employer contributions (as a percentage of sal- ary). The government contribution is key to the feasibility of the entire system. If the amount is too high, the government will not be able to pay for it even with new taxes or massive borrowing and the system will not be financial sustainable. Table 3.9 shows the required government contributions under the moderate and rapid enrollment scenarios. In the early years, a low formal sector contribution might be feasible; however, with 60 percent of the self-employed enrolled by 2013, even with a high contribution rate (9 percent), the government will have to pay about K Sh 32 billion (about US$400 million) to cover the financing gap. As noted earlier, ideally, the fund would invest reserves in the early years to pay for future needs when large numbers of the poor enroll, but because of the fear of corruption, this is not possible. Table 3.7. Implementation Scenarios, 2004­12 (percentage of each group enrolled) Proportion of Category population (%) 2004 2005 2006 2007 2008 2009 2010 2011 2012 Moderate Dependents of the self-employed 52.0 15.0 25.0 37.5 40.0 42.5 45.0 47.5 50.0 52.5 Self-employed 28.0 15.0 25.0 37.5 40.0 42.5 45.0 47.5 50.0 52.5 Public employees 2.5 80.0 85.0 90.0 95.0 100.0 100.0 100.0 100.0 100.0 Private employees 3.5 60.0 62.5 65.0 67.5 70.0 72.5 75.0 77.5 80.0 Pensioners 1.0 80.0 85.0 90.0 95.0 100.0 100.0 100.0 100.0 100.0 57 Other dependents 13.0 68.0 72.0 75.0 79.0 82.0 84.0 85.0 87.0 88.0 Percentage of the total population that would be enrolled 100.0 26.0 34.0 45.0 48.0 51.0 53.0 55.0 57.0 62.0 Rapid Dependents of the self-employed 52.0 37.5 40.0 45.0 50.0 55.0 60.0 65.0 70.0 75.0 Self-employed 28.0 37.5 40.0 45.0 50.0 55.0 60.0 65.0 70.0 75.0 Public employees 2.5 80.0 85.0 90.0 95.0 100.0 100.0 100.0 100.0 100.0 Private employees 3.5 80.0 85.0 90.0 95.0 100.0 100.0 100.0 100.0 100.0 Pensioners 1.0 80.0 85.0 90.0 95.0 100.0 100.0 100.0 100.0 100.0 Other dependents 13.0 80.0 85.0 90.0 95.0 100.0 100.0 100.0 100.0 100.0 Percentage of the total population that would be enrolled 100.0 46.0 49.0 54.0 59.0 64.0 68.0 72.0 76.0 80.0 58 Kenya: Designing Social Health Insurance Table 3.8. Average Annual Contributions by Public and Private Sector Employees Total employer and employee Contribution contribution rate (%) Percentage of salary (employer and employee) 5.8 7.0 9.0 Contribution from person making an average annual public sector salary of K Sh 60,000/year 3,480 4,200 5,400 Contribution from person making an average public sector salary of K Sh 140,000/year 8,120 9,800 12,600 Source: WHO and German Agency for Technical Cooperation 2004. Table 3.9. Required Government Contribution to Balance the Fund, Selected Years (K Sh billions, 2004 constant prices) Total employee and employer contribution (%) 2004 2005 2006 2009 2013 Moderate scenario 5.8 2.1 5.6 11.3 16.1 30.3 7.0 0.5 3.9 9.4 13.7 27.3 9.0 0.0 1.0 6.3 9.8 22.4 Rapid scenario 5.8 9.0 9.9 12.0 19.4 41.2 7.0 7.0 7.8 9.6 16.4 37.7 9.0 3.7 4.1 5.6 11.4 31.9 The moderate implementation scenario is much more affordable and more real- istic, but will also require high contribution rates by the formal sector and high government contributions in the long term, which could be a problem if the gov- ernment refuses to earmark part of the value added tax. The difficulties inher- ent in enrolling the rural poor could give the NSHIF the time it needs to become established and to gain grassroots support. If too many poor enroll too quickly, the system will collapse from underfunding. The reality is that the fund's long-term feasibility depends either on not enrolling those who cannot pay for themselves or having the small group of urban rich pay for millions of rural poor, both out of their contributions and by means of consumption taxes; however, the rich have already called for contribution ceilings of K Sh 5,000 per month (or about US$770 per year). If this ceiling is granted, the government's contribution out of general revenues would have to be enormous. Paying Providers: Methods and Rates Another essential step to implementing the NSHIF is deciding on provider payment methods and rates. The payment system has serious implications for the cost of the program during the first year, and is even more important in relation to incentives to control the growth of expenditures and provide high-quality care. The govern- ment has considered the advantages and disadvantages of many payment systems, Andrew Fraker and William C. Hsiao 59 but all methods still seem to be on the table. It notes that fee-for-service payment can encourage excessive use and involves high administrative costs in relation to checking claims. Per case payments (per visit, per admission, per bed day, per diagnostic-related group) are simpler to administer, but may not avoid all exces- sive use. The government also noted that budgets and capitation-based systems are good in terms of ease of administration and forecasting costs, but might lead to underprovision. They intended to investigate using payments based on case, bed- day, admission, or diagnostic-related group. Initially, maintaining the current fee-for-service payment system might make sense until cost accounting procedures improve. The government might then con- sider flat-rate renumeration per inpatient day, reducing rates after seven or so days to discourage overuse. The government is counting on negotiating low provider payments to improve the financial viability of the fund, but doing so might exacer- bate the brain drain of health professionals. Conclusion Kenya's parliament passed the NSHIF Bill in June 2004, but the president did not sign it into law because of its apparent lack of sustainability without large new taxes or borrowing. In addition, the legislation was voted on only a month after the bill was written and passed in a record 30 minutes, so stakeholders were not as involved as they would have liked. Now that the bill has stalled, they have been able to raise their voices and demand concessions: lower tax rates, lower contribu- tion ceilings, refusal to give up the old benefits system, representation on the board, input into provider payment methods and rates, and so on. Too many concessions will make the NSHIF even less sustainable. Major implementation issues remain. The government needs to define who is poor and cannot afford to pay the premium and plan to certify this population. It then needs to clearly outline how it plans to pay for the poor. Will the government cover their contributions entirely? Will it depend partly on cross-subsidies from contributions by the formal sector? Will the poor receive a lesser benefits package? If the government chooses the second option, which seems financially necessary, it needs to be upfront about this, or else employers and employees will think that their contributions are being squandered. The other major concern has been pre- venting corruption, and despite significant consideration of this issue when design- ing the NSHIF structure, it will always be a challenge. If the public fears its money is being siphoned away by board members or administrators, voluntary compli- ance rates will fall and the fund will need to hire more enforcement officers, and it will never meet its administrative cost target of 5 percent. References CIA (Central Intelligence Agency). 2006. The World Factbook. McLean, VA: CIA. http://www. cia.gov/cia/publications/factbook/. Kenya Private Sector Alliance. 2004. Report on the National Workshop on the Proposed National Social Health Insurance Fund (NSHIF). Nairobi: Kenya Private Sector Alliance. Maliti, John. 2005. "The Flip-Side of Free Health Care." Inter-Region Economic Network News- letter 3 (January). 60 Kenya: Designing Social Health Insurance Ministry of Health. 2002. Kenya National Health Accounts 2001­2002. Nairobi: Government of Kenya. ------. 2004. Sessional Paper No. 2 of 2004: On National Social Health Insurance in Kenya. Nai- robi: Government of Kenya. Republic of Kenya. 2004. Kenya Gazette Supplement No. 29 (Bill No. 10): The National Social Health Insurance Fund Bill. Nairobi: Government of Kenya. Transparency International. 2006. Corruption Perception Index. London, Pluto Press. WHO (World Health Organization). 2002. WHO Country Cooperation Strategy: Republic of Kenya. 2002­2005. Brazzaville: WHO. ------. 2006. World Health Report 2006. Geneva: WHO. WHO and German Agency for Technical Cooperation. 2004. National Social Health Insurance Financial Projections and Future Bilateral/Multilateral Cooperation. Joint WHO-GTZ Mission to Kenya. Mission 6: June 21­26. Nairobi: WHO Country Office. World Bank. 2006. World Development Indicators 2006. Washington: World Bank. 4 Ghana: Initiating Social Health Insurance Sreekanth Ramachandra and William C. Hsiao Ghana, a low-income agrarian country, recently adopted and implemented SHI. The financing is centralized at the national level, but the administration is decen- tralized to the district level. This case deals with what low-income countries must consider in designing and implementing SHI. Background Formed from the merger of the British colony of the Gold Coast and the Togoland trust territory, in 1957, Ghana became the first Sub-Saharan country in colonial Africa to gain its independence. A long series of coups resulted in suspension of the constitution in 1981 and a ban on political parties. A new constitution that restored multiparty politics was approved in 1992. Recently, the international community has praised Ghana as a democratic nation where a fair and orderly election was held in 2004, with John Kufuor winning a second term. Ghana is located in Western Africa, bordering the Gulf of Guinea, between Côte d'Ivoire and Togo (figure 4.1). Its natural resources include gold, timber, industrial diamonds, bauxite, manganese, fish, rubber, petroleum, silver, salt, and limestone. Sixty-three percent of the population is Christian, 21 percent adhere to indigenous beliefs, and 16 percent is Muslim. As concerns ethnic groups, in 1998, 98.5 percent were black and 1.5 percent were of other races. The major tribes are Akan (44 per- cent of the total population), Moshi-Dagomba (16 percent), Ewe (13 percent), Ga (8 percent), Gurma (3 percent), and Yoruba (1 percent). English is the official language, and African languages, including Akan, Moshi-Dagomba, Ewe, and Ga, are also spoken.Table4.1providesbasicstatisticsregardingGhana'sdemography,economy, health status, and health system. Economy Ghana's income per capita is above the average in Sub-Saharan Africa and grew faster than that in the rest of this region over the past decade. However, Ghana has a large poor population. The urban population is relatively large, though most city dwellers live in slums. Well endowed with natural resources, and with twice the per capita output of poorer countries in West Africa, Ghana still remains heavily dependent on inter- 61 62 Ghana: Initiating Social Health Insurance Figure 4.1. Map of Ghana Source: World Bank. national financial and technical assistance. Gold, timber, and cocoa production are major sources of foreign exchange. The domestic economy continues to revolve around subsistence agriculture, which accounts for 35 percent of GDP and employs 60 percent of the workforce, mainly small landholders. With the current focus on how health status influences income, which in turn facilitates health, the huge economic inequity (Gini coefficient of 41) and the extent Sreekanth Ramachandra and William C. Hsiao 63 Table 4.1. Background Statistics, Ghana Category Statistic Source Demographics (2005) Population (millions) 22.1 World Bank 2006 Growth rate (%) 2.0 World Bank 2006 Urban population (%) 47.8 World Bank 2006 Economy GDP per capita, purchasing power parity (2005) 2,402.0 World Bank 2006 GDP annual growth rate, per capita, real purchasing power parity (past decade) 2.4 World Bank 2006 Percentage of the population below the national poverty line (1999) 39.5 CIA 2006 Percentage of the population living on less than $2/day (1999) 75.0 World Bank 2006 Percentage of the population living on less than $1/day (1999) 45.1 World Bank 2006 Agricultural employment (percentage of the population) (2005) 60.0 World Bank 2006 Corruption (rank out of 163 countries, 2005)a 70 Transparency International 2006 Health status Life expectancy (years) (2004) 57.2 World Bank 2006 Infant mortality/1,000 live births (2004) 68.0 World Bank 2006 HIV/AIDS prevalence (percentage of the population age 15­49) (2005) 2.3 World Bank 2006 Health system Health spending (percentage of GDP) (2003) 4.5 World Bank 2006 Health spending (US$ per capita, purchasing power parity) (2003) 108 World Bank 2006 Physicians/1,000 population (2004) 0.15 World Bank 2006 Hospital beds/1,000 population (2005) -- World Bank 2006 Health system financing (percentage of total health expenditure) Public (2003) 31.8 WHO 2006 Private (2003) 61.2 WHO 2006 of poverty are reasons for worry. While in 2000, the richest 20 percent of the popu- lation earned 46.7 percent of the total income, the poorest 20 percent earned a mere 5.6 percent. While overall poverty declined from 51.7 percent in 1992 to 39.5 per- cent in 1999 and extreme poverty declined from 36.5 percent to 26.8 percent during the same period (Government of Ghana 2003),1 the reduction was concentrated in 1. Using a lower poverty line (extreme poverty) of ¢700,000 per adult per year and an upper poverty line (overall poverty) of ¢900,000 per adult per year (US$1 = 9,603). The ¢ poverty lines are based on what is needed to meet the nutritional requirements of house- hold members. 64 Ghana: Initiating Social Health Insurance Accra and in rural and urban forest localities. In the urban savannah, the popula- tion defined as poor actually increased. Health While Ghana's health indicators compare favorably with those in other Sub-Saharan African countries, they still lag far behind world averages. Large improvements in life expectancy and infant mortality occurred between 1960 and 1995, but both indicators have stagnated since that time. Inequities between urban and rural areas are cause for concern, for example, in northern Ghana, the mortality rate among children under five is 171 per 1,000 live births, triple that in the capital region (UNDP 2003). The maternal mortality rate figures are also high, and insufficient prenatal care, unsupervised deliveries, and inadequate postnatal care (more than 50 percent of mothers receive no postnatal care) are among the important under- lying causes of high levels of maternal deaths. Malnutrition among women, high fertility rates, and harmful traditional practices also contribute to this high rate of maternal mortality. The 1998 demographic and health survey reported that 83 per- cent of pregnant women received prenatal care from a trained health worker. An encouraging trend is that by 2001, this figure had increased to 96 percent. However, only some 30 percent of pregnant women received prenatal care from a doctor. The high prevalence of HIV/AIDS is a problem. In 2005, an estimated 310,000 people were living with HIV/AIDS and 29,000 died during the year (UNAIDS 2006). Malaria remains a leading cause of death and is the largest cause of outpatient hospital visits. Little progress has been made in reversing its incidence. In 2003, malaria accounted for 45 percent of all outpatient cases and more than 33 percent of all inpatient cases in health facilities. Twenty-two percent of all deaths among children under five are due to malaria. The prevalence of malaria-related fever is 34 to 38 percent in the impoverished northern, upper east, and Volta regions, com- pared with a national average of 27 percent (UNDP 2003). Even though Ghana has received grants from the Global Fund to Fight AIDS, Tuberculosis, and Malaria, a severe shortage of doctors and nurses has dampened its success. Estimates sug- gest that two-thirds of the doctors and nurses trained in Ghana have left the coun- try for better economic prospects overseas. Rural areas are particularly hard hit, and many northern regions have no access to doctors at all (http://www.oneworld .ca/guides/ghana/development). Health Financing and Expenditure Total health care expenditure was around 4.5 percent of GDP in 2003. The majority of health spending was financed privately, mostly out-of-pocket payments. Most public money is channeled though the Ministry of Health. The regular budget of the Ministry of Health accounted for about 5.8 percent of general revenues, which is low in terms of public spending on health by interna- tional standards, of which the amount coming from donors (28.2 percent) is high by international standards. The government has pledged to increase public spend- ing (domestic and foreign aid) to 15 percent of general government revenues in the next few years. With the introduction of the new health insurance system, a formal mechanism for pooling revenues and spreading risks across population groups, from rich to poor and across the lifecycle, is expected. Sreekanth Ramachandra and William C. Hsiao 65 Public and Private Provision of Health Services Hospitals in affluent regions tend to be publicly owned. In the most affluent region, Greater Accra, all 11 hospitals are public. Throughout the country, 95 out of 204 hos- pitals are public. Almost 60 percent of health clinics are public, and the distribution of public and private ownership varies widely by region. For example, 14.5 percent of clinics are public in Greater Accra, but 86.0 percent are public in Volta. Evolution of Health Financing Mechanisms Table 4.2 sets out the development of health financing mechanisms in Ghana from a publicly financed national health service to the current national health insurance scheme. Mutual Health Insurance Organizations The voluntary mutual health insurance movement started in Ghana during the early 1990s with encouragement from the Ministry of Health and support from donors such as the Danish International Development Agency and the U.S. Agency for International Development (Aiken 2003; Apoya and others 2001; Atim 2000; Atim, Grey, and Apoya 2002). Such community initiatives began to bridge the large gap in social protection between people covered by formal schemes and those with no protection against the costs of illness or who were exposed to the impoverishing effects of user charges (Arhin-Tenkorang 1995).2 The movement has recently taken off, increasing from 4 MHOs in 1999 to 47 in 2001 and 159 in 2002 (Atim, Grey, and Apoya 2002). These MHOs are spread across 67 districts in all 10 regions of the country. The MHOs show richness in variety and innovation of design and man- agement styles. Some, such as the Dodowa community health insurance scheme, aredistrictwide,whileothersarebasedondifferenttypesofcollectivegroups,such as occupation (like the Ashanti Region Civil Servants Medical Insurance Scheme), religion (for instance, the Koforidua Diocese Mutual Health Scheme), or gender (for example, the Manhyia Susu Health Scheme). Data from a survey conducted in November 2002 by Partners for Health Reform Plus indicates that approximately 89 percent of MHO membership was concen- trated in three regions: Ashanti (38.7 percent), Brong Ahafo (20.1 percent), and the Northern region (30.5 percent). Of the total number of group members, 62 percent were female and 62 percent were farmers. The survey also reports that the existing MHO market is heavily concentrated. The four largest MHOs have 71 percent of total members (table 4.3). These figures also imply that only about 1.1 percent of the population is enrolled in an MHO, with the remainder depending on the cash-and- carry system. Membership in individual schemes ranges from 1,000, as in the Tano Commu- nity Health Insurance Scheme, to more than 60,000, as in the Ashanti Region Civil Servants Medical Insurance Scheme. In some cases, such as the Koforidua Diocese 2. This is similar to the role of communities in health financing in other African coun- tries (Arhin-Tenkorang 2001; Huber, Hohmann, and Reinhard 2003; Musau 1999; Waelkens and Criel 2004) and elsewhere (Baeza, Montenegro, and Nunez 2002; Bennett, Creese, and Monasch 1998; Jakab and Krishnan 2001; Preker and others 2001). Table 4.2. Chronological Development of Health Financing Year Event Rationale Features Source of financing Result 1957 Introduction of a Driven by early economic · Everyone entitled to free General revenues Not sustainable: with the decline national health performance, natural health care in economic performance, service modeled after resources, and strong · Health care delivery the scheme proved to be too the British system export base through a network of expensive publicly owned facilities 1985 Copayments To avoid a collapse in · Copayments for services General revenues and Out-of-pocket user fees charged introduced publicly funded services · Health care delivery user fees from partial to full cost recovery through a network of publicly owned facilities 66 1992 Cash-and-carry · To increase funds for · Full cost recovery for General revenues and Outpatient visits dropped by system instituted providers drugs user fees 66% (Dzikunu and Thorup 2003) · To make fee recovery · Reduced fees for legal children and primary · To restrict unnecessary care facilities use Early 1990s Voluntary mutual · Heavy cash-and-carry · Subsidization of the · Donors such as the · Reduced gap between those health insurance burden vulnerable by the Danish International covered and those not covered organization · Lack of social protection better off Development Agency (Arhin-Tenkorang 1995) movement mechanisms · Social protection against and the U.S. Agency · Paved the way for spread · Lack of government the impoverishing costs for International of MHOs oversight of the informal of illness Development sector · Community Table 4.2. Chronological Development of Health Financing (continued) Year Event Rationale Features Source of financing Result (table continues on following page) Early 2000s Profusion of MHOs · Trend in other African · Spread across 67 out of · Donors such as the · Financial protection and health nations 138 districts in 10 regions Danish International services access for poor · Success of initial MHOs · Diverse in management Development Agency · Model for covering larger parts in Ghana styles and benefits and the U.S. Agency of the population · Encouragement by the · Based on district or for International 67 Ministry of Health occupation or religion Development or gender · Community 2003 National mandatory · Relative success of the · Abolish cash and carry · Health insurance levy · National Health Insurance health insurance MHOs · Introduce mandatory · 2.5% of SSNITa state Council set up reform · Agenda of the ruling health insurance budget transfers · Interim administrative government (re-election · Expand coverage · Returns on investment arrangements introduced platform) through MHOs in all made by the National · Move toward 90 district mutual districts HealthInsuranceCouncil health insurance schemes · Voluntary contributions 68 Ghana: Initiating Social Health Insurance Table 4.3. Distribution of Membership, Four Largest MHOs, 2002 Percentage of total Number of number of people MHO members insured by MHOs Stage Ashanti Region Civil Servants Medical Refund Scheme 60,300 27 Functional Nkoranza Community Health Insurance Scheme 44,000 20 Fully operational Damongo Health Insurance Scheme 32,000 14 Fully functional Tiyumtaabi Welfare Association 21,200 10 Fully functional Total 157,500 71 Source: Atim, Grey, and Apoya 2002. Mutual Health Scheme, enrollment is compulsory for all church members, but in others, like the Dodowa Community Health Insurance Scheme, enrollment is vol- untary, and in the first year was barely 5 percent of the eligible population. Most schemes are sponsored and managed by providers. For example, in the St. Rose's Secondary School Health Insurance Scheme in Akwatia, St. Dominic's Hospital, a mission hospital and the largest district hospital in the area with spe- cialist staff, initiated the scheme in January 2000 and is also the service provider, that is, the service provider owns the scheme. One of the largest schemes, the Nkoranza Health Insurance Scheme, started out as a community-based scheme with the service provider being St. Theresa's Hospital, the only referral hospital in the district, but the scheme was housed in the hospital, and finally ended up being run by the service provider. Health benefits can generally be accessed one year after joining any scheme. In most cases, benefits are related to ailments requiring hospitalization for at least one day. For example in the Nkoranza Health Insurance Scheme, the benefits relate to admissions for medical, surgical, obstetric, and emergency care. Risk coverage is limited to hospital admission and deliveries are covered only when they are com- plicated. Self-induced abortions are not covered. The schemes are typically financed by premiums collected from the particular community. The premiums start at ¢20,000per adult per year (about US$2.2), with the actual amount depending on the number of members the scheme can enroll and the services provided. Members' ability to pay is taken into account and the poor are subsidized. For example, in the Koforidua Diocese Mutual Health Scheme, which commenced in 2000, many of the families in the scheme are poor and/or unemployed. Even a premium of 500 (US$0.05) per week is too expensive for most ¢ members. To encourage premium collections, the frequency of standard church col- lections has been reduced and 30 percent of the daily church collections is set aside to subsidize MHO premiums. Thus premiums are based on a family's contributions to the church and therefore vary. National Health Insurance Scheme Ghana passed the National Health Insurance Act (NHIA) in 2003, and it became operational in March 2004. The scheme is operated as a decentralized national healthinsurancesystemencompassingdistrictmutualhealthschemesinall110dis- Sreekanth Ramachandra and William C. Hsiao 69 tricts, private mutual health insurance, and private commercial insurance schemes in order to give all Ghanaians the opportunity to join a health insurance scheme of their choice. The central government sets the minimum benefits package, licenses and regulates the health insurance schemes, certifies the providers, and collects a national health insurance levy and uses it to subsidize premiums for the poor. The intent is that insurance will eventually become mandatory, which implies that all Ghanaians would somehow be compulsorily enrolled. The policy document speaks of achieving insurance coverage of 30 to 40 percent of the population by 2010 and 50 to 60 percent by 2015­20. The scheme follows the provisions laid out in the 2003 NHIA and the subse- quent national health insurance regulations. The important provisions of the NHIA are (a) the establishment and functions of the National Health Insurance Council; (b) the types, registration, and licensing of the various health insurance entities included in the overall national insurance scheme; (c) the establishment and opera- tion of district mutual health insurance schemes; (d) the establishment of private health insurance entities comprising private commercial and private mutual health insurance; (e) the general provisions applicable to the operation of all health insur- ance schemes; (f) the establishment of the National Health Insurance Fund; (g) the imposition of a national health insurance levy; and (h) the administration and effective implementation of the NHIA's provisions. NATIONAL HEALTH INSURANCE COUNCIL. The National Health Insurance Coun- cil was set up to govern the insurance scheme and reports to the president through the minister of health. Its objectives are to secure implementation of a national health insurance policy that ensures access to basic health care for everyone. The council has 15 members representing various interest groups of society, including a chair and an executive secretary. The president appoints council members in con- sultation with the Council of State. The council discharges its functions through a health insurance administration and committees as deemed appropriate, but is required to comply with health policy directives issued by the minister of health. Initial appointments to the council were made in May 2004. The council's responsibilities include the following: · to register, license, and regulate health insurance schemes; · to supervise the operations of health insurance schemes; · to accredit health care providers and monitor their performance; · to ensure that health care services rendered to beneficiaries of schemes by accredited health care providers are of good quality; · to determine, in consultation with licensed district mutual health insurance schemes, the contributions that their members should make; · to approve health identity cards for members of schemes; · to provide a mechanism for resolving complaints by schemes, members of schemes, and health care providers; · to recommend health insurance policies to the minister of health; · to undertake sustained public education on health insurance; · to devise a mechanism for ensuring that the basic health care needs of indi- gents are adequately provided for; · to maintain a register of licensed health insurance schemes and accredited health care providers; 70 Ghana: Initiating Social Health Insurance · to manage the National Health Insurance Fund; · to monitor compliance with the NHIA and its regulations and to pursue action to secure compliance. TYPES, REGISTRATION, AND LICENSING OF HEALTH INSURANCE SCHEMES. Three types of health insurance schemes were established under the NHIA, namely: · district mutual health insurance schemes (district based and state sponsored), · private commercial health insurance schemes (private for-profit schemes), · private mutual health insurance schemes (community-based nonprofit schemes). A district mutual health insurance scheme is to be established in every district. It will be responsible for establishing a district administration, enrolling and main- taining membership, collecting contributions from people who can pay, applying a means test to determine who is indigent, and administering subsidies received from the National Health Insurance Fund for the indigent. People have a choice of enrolling with a private commercial health insurance scheme or a private mutual health insurance scheme instead of joining the local district mutual health insur- ance scheme. The private commercial health insurance schemes are established under the 1963 Companies Code and required to comply with relevant provisions of the Insurance Law of 1989. They are not eligible for a subsidy for the indigent under the National Health Insurance Fund. The NHIA specifies the operating principles for registered health insurance schemes: the establishment of a governing body; the management of the scheme, including financial management, reporting requirements, and audits; the staffing; the membership registration, identification, and termination; and the specification of benefits packages. The NHIA also lays out procedures for hearing and settling complaints, accrediting health care providers, and monitoring the quality of ser- vice providers; safeguards to prevent the excessive use or the abuse of benefits; schedules for settling outstanding payments to providers; and inspection proce- dures. Each health insurance scheme is required to comply with directives from the Health Insurance Council and may be required to appoint an actuary if the council has reasonable grounds to think that a particular scheme has contravened provisions of the NHIA or related regulations. NATIONAL HEALTH INSURANCE FUND. The purpose of the National Health Insurance Fund is to provide a direct subsidy to the district mutual health insur- ance schemes that offer the minimum health care benefits stipulated by the NHIA, to reinsure the district funds against random fluctuations in the costs of health care, to subsidize the costs of health care for the indigent, and to support pro- grams that improve access to health services. The health fund is financed through a health insurance levy of 2.5 percent on goods and services produced in Ghana or imported, with some exceptions; 2.5 percent of the social security and pensions scheme funds;3 transfers from the state budget allocated to the fund by parliament; returns on investments made by the National Health Insurance Council; and vol- 3. The funds collected from social security payments also serve as the premiums paid by workers in the formal sector, as they are automatically covered by virtue of their contri- butions into the Social Security and National Insurance Trust. Sreekanth Ramachandra and William C. Hsiao 71 untary contributions to the fund (grants, donations, gifts, and other sources of financing). The council may modify these sources of funding to keep pace with developments in the health insurance industry. Monies for the fund are held in bank accounts approved by the accountant-general. Transfers from the fund to the district mutual health insurance schemes are approved annually by parliament. The fund is managed by the council. This includes managing the fund's liquid- ity, investing temporary surpluses, maintaining appropriate accounts, submitting annual reports, and conducting regular audits of the fund's financial activities. The 2005 NHIF allocation formula was as follows: ¢139.2 billion (US$15.5 mil- lion) for 1.74 million indigents, 48billion (US$5.33 million) for 600,000 people aged ¢ 70andolder,and 60billion(US$6.66million)for2millionchildrenunder18.Other ¢ beneficiaries are 81 billion (US$9 million) for 132 financially distressed schemes ¢ and 60 billion (US$6.66 million) for Social Security and National Insurance Trust ¢ contributors automatically covered under the law. The remainder includes operat- ing expenses of the National Health Insurance Council Secretariat, contingencies, administrative expenses, investments, and reinsurance of the National Health Insurance Scheme. MINIMUMBENEFITSPACKAGE. AccordingtotheNHIA(Article64),everyenrollee in any scheme is entitled to those minimum health care benefits specified by the minister of health on the advice of the council. The minimum benefits package pro- vides a comprehensive list of services, including the following: · certain outpatient services, including general and specialist care, requested investigations (for example, X-rays and ultrasounds), medications listed on the National Health Insurance Scheme drugs list, symptomatic treatment for opportunistic infections of HIV/AIDS, simple surgeries, and physiotherapy; · certain inpatient services similar to those included for outpatients with the addition of room and board in a general ward and cervical and breast cancer treatment; · basic oral health services; · eye care services, including cataract removal and eyelid surgery; · maternity care, including prenatal, postnatal, and delivery services; · all emergencies or crisis situations that demand urgent intervention. Emergencies are covered at any health facility. For all other services, the first point of attendance must be a primary health care provider. The list of services cov- ered can be expanded subject to the payment of an additional premium as agreed upon by the scheme and its members. District mutual health insurance schemes, however, cannot expand upon this list without prior council approval. Public health services are available free to every Ghanaian, irrespective of whether they are scheme members. These services include immunization; fam- ily planning; inpatient and outpatient treatment for mental illness; treatment for tuberculosis, onchocerciasis (river blindness), Buruli ulcer, and trachoma (a type of blindness); and HIV testing. What specific immunizations are offered or to what services family planning specifically refers to is unclear at this time. Design Issues The scheme's design has a number of troubling aspects: the administrative struc- ture needs to be articulated better; the financing mechanism for the NHIF and the 72 Ghana: Initiating Social Health Insurance disbursement of funds need to be streamlined with a focus on the scheme's long- term sustainability; a universal minimum benefits package needs to be spelled out after a proper costing exercise; a clear understanding of how this benefits pack- age will interact with the free public health services provided by the government is needed; the definition of indigents needs to be evaluated to show clearly who the beneficiaries of the subsidy are and whether everyone who needs a subsidy will get it; the purchasing mechanism from providers needs to be spelled out; and the pos- sibility of an uneven playing field needs to be taken into account given the supply- side subsidies available to public providers. Administrative Structure The bureaucratic setup now has three new additions: the National Health Insur- ance Council, the National Health Insurance Fund, and the district administrative entities. However, consensus on which entity is the final authority and on what roles the different entities are responsible for seems to be lacking. Figure 4.2 shows a rough outline of the structure based on available data. The reformers initially argued that health insurance was the sole responsibil- ity of local government, but after a year, there seemed to be an attempt to bring health providers and local governments together to initiate the process of estab- lishing schemes. This leaves the Ghana Health Service and district administrations unsure about who is in the driving seat. In addition, community ownership and participation will likely be affected if care is not taken to clearly spell out the roles and responsibilities of different parties. Also, as in many low-income countries, Ghana has a shortage of skilled man- agers and few who have been trained in managing social insurance systems. The rapid growth of MHOs has far outstripped growth in technology and in skilled personnel. A decentralized insurance system based on local mutual health insur- ance schemes will be resource intensive and require significant management Figure 4.2. Interaction between Administrative Entities Ministry of Health National Health Insurance Fund National Health Insurance Council District Mutual Health Insurance Scheme Health Service Private Commercial Health Insurance Scheme Private Private Mutual Health providers Insurance Scheme Source: Authors. Sreekanth Ramachandra and William C. Hsiao 73 development so that the district MHOs, as well as the central administration, have appropriately skilled staff. Financing and Disbursement Issues The details of National Health Insurance Fund financing were discussed earlier, but no explanation of how the authorities arrived at this breakdown is available. There is no evidence of any estimation of the budget required, although the Minis- try of Health claims that replacing the cash-and-carry system and financing health care in public facilities will cost the government between ¢70 billion and 120 bil- ¢ lion (more than US$13 million) per year.4 Another issue is the allocation and disbursement of funds. The NHIA does not specify any mechanisms for this, but simply delegates this responsibility to the National Health Insurance Council. Section 77.1 of the NHIA indicates how the fund can be used, such as providing subsidies to and reinsuring district MHOs, paying for the health care costs of indigents, and facilitating access to health ser- vices. Parliament has approved an allocation of funds based on a formula, though how these funds are to be disbursed and how they will be allocated to various dis- trict mutual health insurance schemes has yet to be spelled out. Costing of the Minimum Benefits Package Many questions concerning the scope of the mandatory benefits package and the packagefavoredbydistrictMHOsfortheirmembersremainunresolved.Theoptions chosen will be important for the evolution of providers. The Ministry of Health has published a broad package covering almost all care except chronic diseases for feed- back. Such a package is too extensive to be sustainable over the long term, and the package does not appear to have been costed or to have considered cost escalation, especially given the expected overutilization of services. Given traditional relations between the state and the public in most African countries, whereby government- sponsored initiatives are often exploited, this is a legitimate fear. Administrative costs might also tend to be higher in the health sector because of the burden of han- dling insured patients requiring complex and lengthy medical treatments. In the case of MHOs, the packages were designed to suit scheme members and were costed accordingly, whereas the new packages might cost a lot more, which will put pressure on the private schemes. One fear is that schemes will try to compensate for this by cutting corners. In an extreme case, cost pressure could even lead to the crowding out of private health insurance. Another fear is that as the cost structures of private insurers and the district mutual health insurance schemes vary with the providers they use, this could lead to differences in the quality of health care and also different insurance systems for people in different socioeconomic strata. Another issue concerns the free public health services provided by the govern- ment. No effort seems to have been made to reconcile these with the new national health insurance system or even to set out exactly what the free services will consist of. This needs to be done in order to cost the national health insurance system. 4. IRIN News. "GHANA: National Health Insurance Scheme Launched." United Nations Office for the Coordination of Humanitarian Affairs. http://www.IRINNews.org. 74 Ghana: Initiating Social Health Insurance Actuarial Soundness of the District Mutual Health Insurance Schemes NoactuarialcostingseemstohavebeendonebeforetheNHIAwasenacted,although theNHIAdidrecognizetheimportanceofactuarialexpertiseandsuggestedemploy- ing one actuary. Since the NHIA came into effect, the Social Security and National Insurance Trust, with the assistance of the International Labour Organisation, has undertaken an actuarial analysis, but Ghana lacks sufficient reliable data to prepare reasonable cost estimates. The main result of the analysis seems to be that the pub- lic health sector is unlikely to face expenditure problems that cannot be mitigated by adjustments to the health care financing system, such as increasing premiums for members of district mutual health insurance schemes or temporarily reducing investment in health services infrastructure. The key uncertainty in the projections remains the future increase in the coverage rate of those in the informal sector. The analysis also assumes a growth in formal employment that is not reflected in cur- rent trends. In addition, the model depends on the current definition of indigents, which probably vastly underestimates the actual number of indigents. Exemptions for Indigents The NHIA exempts the poorest Ghanaians, referred to as indigents, from having to pay premiums. The NHIA (Article 104) defines an indigent as "a person who has no visible or adequate means of income or who has nobody to support him or her and by the means test qualifies as an indigent." The minister of health, on the advice of the council, is supposed to impose a means test for determining who is indigent. This is a stringent test that will qualify only the poorest of the poor. According to Regulation 58, people cannot be classified as indigent unless they meet each of the following four criteria: (a) they are unemployed and have no visible source of income, (b) they do not have a fixed place of residence, (c) they do not live with a person who is employed and who has a fixed place of residence, and (d) they do not receive any identifiable and consistent support from another person. Furthermore, if the list of indigents submitted by any district mutual health insurance scheme exceeds 0.5 percent of the entire membership of the scheme, the council shall verify the list by whatever means it deems fit. There is a complaints mechanism for those who feel they were misclassified as nonindigent. Clearly, the number of indigents in each district will be limited, which could mean that many of the poor and needy who deserve a subsidy may be excluded. The Ministry of Health estimates the total number of indigents at 1.74 million, or about 9 percent of the population. This seems unreasonable in a country where 40 percent of the population is poor and more than 25 percent is extremely poor (see the section on economic indicators). This gives rise to two issues. First, the state would find col- lecting the premiums it envisages difficult, as many people would be unable to pay. Second, many poor people would be uninsured and would be denied access to care. Thus the scheme could end up hurting the very people it was intended to benefit. At the same time, if the definition of indigents is relaxed to cover more people, then the question of how to muster additional financing crops up. Purchase of Services and Pricing Issues The current system of payment is based on government budget allocations and out- of-pocketpaymentsbypatients.Thesalariespaidtohealthworkerstodeliverservices Sreekanth Ramachandra and William C. Hsiao 75 are not related to how much work individuals do or the volume of services delivered. With insurance, health care services will be paid on a capitation or fee-for-service basis. The previous system was a cash-and-carry method that functioned on a fee- for-service basis, and this still seems to be the Ministry of Health's favored method. A traditional fee-for-service system, based on individual services and detailed invoices, has two major disadvantages for insurers: assessing individual claims to verify whether all services claimed have actually been delivered (and were neces- sary) is difficult, and allowing providers to claim as much as they feel is necessary places the entire financial risk on the insurer. Meanwhile a capitation system needs to be based on a mutual understanding of the probable cost of services for a given population. Without correct and detailed information on expected costs, usually derived from historical data, capitation puts a high financial risk on providers. If the capitation fees do not reflect real costs, providers usually react by refusing ser- vices to members or by asking patients for additional (informal) payments. Also the varying cost structures of different schemes imply a need for different capitation fees. Thus a capitation system seems inappropriate at the start of the new district mutual health organization schemes, especially in the hospital sector. Uneven Playing Field The scheme's design could create an uneven playing field in two ways. First, the Min- istry of Health provides a huge supply-side subsidy to public providers and the costs at public hospitals for either ambulatory care or for inpatient treatment are bound to be much lower than those at private facilities, which receive no subsidies. In addition, the health sector has no pricing policy for services provided, and public, mission, and private facilities provide services based on different price indexes and input costing systems. This has the potential to lead to a two-tiered health care system. The reputa- tions of the private and the mission hospitals are good and their services are osten- sibly more expensive than those of public facilities. Also most private facilities are concentrated in urban areas where people can afford to pay for their services. Second, an inequity arises between private insurers and district MHOs. The district mutual health insurance schemes receive a direct subsidy and subsidies to cater to indigents. Private insurers will need to depend on paying enrollees, as 100 percent of their revenues will come from the premiums collected. Faith-based and community-based schemes might still receive aid from other sources, but private for-profit organizations risk being crowded out. The foregoing discussion might lead to the conclusion that private for-profit insurers would have to compete in relation to the quality aspect of services by con- tracting with private hospitals and in relation to the cost of services by contracting with public providers, but in a more cost-efficient manner than the state sponsored plans. Thus the situation could turn out to be one in which the rich and the healthy enroll into private, commercial plans; the less rich join community-based, private plans; and the indigent are covered by state-sponsored plans. Implementation Issues In terms of implementation, the handling and integration of a number of plans could prove to be a major hurdle. Enforcing the scheme and ensuring premium col- lection is also bound to be problematic. These issues are elaborated in this section. 76 Ghana: Initiating Social Health Insurance Additional issues that are more generic to heath system functionality in general include strengthening the public and private health infrastructure, ensuring the availability of health personnel, and accrediting public and private facilities. Financing and Integration Issues The recent expansion in the number of MHOs is clearly driven by the acute nature of the cash-and-carry burden facing individual families and their awareness of the existence of MHOs as an alternative to user fees. Against this backdrop, the govern- ment has chosen to emphasize district insurance schemes rather than the spontane- ously developing and dynamic mix of all kinds of MHOs, including districtwide schemes and subdistrict, village, and community schemes. Experience in Ghana and elsewhere shows that coverage by district schemes is usually quite poor. Subdis- trict community schemes and those based on professional solidarity tend to reach a much larger percentage of their target populations and also tend to be better able to implement more robust risk management techniques.5 Outside Accra and a few other large cities, Ghana's social structure remains highly decentralized and concen- trated around the household and ethnic, religious, and occupational groups. Those who live in rural areas, who account for 57 percent of the population (Ghana Statisti- cal Service 2000), and the poor distrust central programs sponsored by the central government based on past difficulties in financing and managing public services. The NHIA does recognize other schemes, such as the private and commer- cial MHOs, but does not support them through subsidies. This means that such schemes would either need to be profitable to survive or should be funded by NGOs. These scenarios become tougher to realize if the government imposes a universal minimum benefits package without doing a cost-benefit analysis. Thus existing community schemes could be compromised by the government's focus on district insurance schemes. In relation to integration, assuming that private and commercial schemes are in effect alongside the district mutual health insurance schemes, the outcome will be a large number of plans run by different administrations as they see fit. Integrat- ing these into a national health insurance scheme is bound to create problems, as the current MHOs are independent of each other and have different premium rates and benefits packages. All these issues still have to be resolved. Enforcement and Enrollment Adult Ghanaians are to pay a monthly minimum subscription of 60,000 (US$0.66), ¢ with premiums adjusted according to income. The government will subsidize the cost of health care--meaning no premiums--for the aged, the poor, and the chil- dren of parents who both subscribe to the scheme. While the Ministry of Health has declared that the scheme will be mandatory, it has not indicated how enrollment into the scheme will be enforced. Moreover, as the council has set targets for coverage over the next 5 to 10 years, this raises questions about the compulsory nature of the scheme and the enforceability of membership. 5. As noted earlier, the biggest scheme in Ghana, in terms of both numbers and cover- age, is not a district scheme, but the enterprise-based Ashanti Civil Servants Association, which represents nearly 100 percent of its target group. Sreekanth Ramachandra and William C. Hsiao 77 Table 4.4. Employment by Category, Selected Years (percentage of the total labor force) Category 1987/88 1988/89 1991/92 1998/99 Formal sector 17.3 18.1 15.4 13.2 Government 8.0 7.9 7.8 5.9 State enterprises 1.9 2.3 1.2 0.6 Private enterprises 7.4 7.9 6.4 6.7 Informal sector 82.7 81.8 84.7 86.8 Farmers 58.7 54.6 56.7 55.7 Nonfarm self-employment 19.5 24.2 23.5 27.3 Unpaid family work and unemployed 4.2 3.0 4.5 3.8 Total 100 100 100 100 Rough estimates from table 4.4 show that less than 15 percent of the popula- tion is employed in the formal sector. More than 80 percent are self-employed farm- ers and workers in the informal sector, of whom only a small proportion will be considered indigent. Enrolling and collecting premiums from nonindigent farmers and informal sector workers will be a huge task. As concerns dependents, according to the NHIA, dependents of working people who contribute to the scheme will be covered, even if they are not elderly or chil- dren. The number of dependants per active contributor could end up being relatively large, and the financial implications of this undertaking need to be determined. Accreditation of Providers The council has the authority to limit the services a facility will provide depend- ing on its equipment and its standards of service. The NHIA states as follows: "A scheme shall not use the services of any healthcare provider or any health facility in the operation of the scheme unless the healthcare provider or the health facility has been approved and accredited to the scheme by the Council. Regulations may prescribe the qualifications, requirements and such other matters as the Council considers necessary in respect of healthcare providers and healthcare facilities that operate under the schemes." The current version of the law does not distinguish between accreditation for different types of facilities or between modern health care professionals with high- level formal training versus community-level providers, such as traditional birth attendants, small-scale sellers of drugs, and practitioners of traditional medicine. With the heavy utilization of services provided by the latter category, the govern- ment needs to be careful not to deny access to those who use traditional medicine. While setting up a permanent accreditation scheme will take time, a temporary one needs to be instituted as soon as possible. The primary focus at this time needs to be an accreditation procedure aimed at ensuring consumer safety and enabling financial access to services in previously underserved areas. Successes and Challenges After lengthy deliberations, the National Health Insurance Council was finally set up in 2004. Merely establishing the National Health Insurance Scheme should be 78 Ghana: Initiating Social Health Insurance termed a success. Other than the protection provided to those enrolled in volun- tary MHOs, protection against the financial risks of illness was previously achieved mainly through access to publicly financed health services. The new health insur- ance system will introduce a formal mechanism for pooling revenues and spread- ing risks across population groups. In June 2005, parliament approved an allocation to the National Health Insur- ance Fund whereby the national health insurance levy accounts for 77 percent of the total and the Social Security and National Insurance Trust accounts for 23 per- cent. The collection of funds has started and the National Health Insurance Council has been able to mobilize more than ¢700billion (US$77.74 million). According to President Kufuor, 40.6 billion (nearly US$5 million) from the ¢ Heavily Indebted Poor Countries Initiative has been used to fund the scheme at the district level. A similar amount has been set aside for health workers who agree to work in deprived rural areas. Thus the current scheme is being subsidized by the Heavily Indebted Poor Countries Initiative, and with the Group of Eight's recent announcement in relation to canceling the debts of the world's poorest countries, Ghana is likely to have more such funds for health care. According to the Ministry of Health, more than 90 districts are at various stages of implementation and 22 districts are about to start the implementation process. The sooner the scheme can be made visible and operational in places, the better its chances of catching on. Challenges include the following: · Increasing coverage and building up political support. In the current context with rival political parties accusing the government of indecent haste in institut- ing the scheme,6 the challenge is to demonstrate small successes in the near future that can build up support. To this end, the National Health Insurance Scheme needs to be on track to reach its purported goal of covering 30 to 40 percent of the population within the next 5 years and 50 to 60 percent within the next 5 to 10 years. This would send the right signals to Ghanaians and to external donors and build support for the scheme. · Rationalizing the administrative structure. From an administrative standpoint, defining the roles of the various entities involved is crucial. The Ministry of Health, the National Health Insurance Council, and the Ghana Health Ser- vice will become increasingly interdependent in the course of formulating and executing health policy. The Ministry of Health currently funds and manages public facilities. The challenge is for the ministry to become the pol- icy maker and regulator of health care rather than the provider of services. · Enforcing mandatory insurance. No specific guidelines have been provided on how this is to be achieved. While universal coverage is a long way off and the government seems to have targeted certain milestones, specifying the groups or people who will be covered during the first phase is important. Given the concentration of health care infrastructure in urban areas and the higher per- centage of people employed in the urban formal sector, and from an ability to pay perspective, a logical assumption is that the urban population will be the 6. IRIN News. "GHANA: National Health Insurance Scheme Launched." United Nations Office for the Coordination of Humanitarian Affairs. http://www.IRINNews.org. Sreekanth Ramachandra and William C. Hsiao 79 first to be covered. However, this defeats the entire purpose of making health care accessible to the poor and to rural populations, of risk pooling, and of subsidizing the poor. The challenge lies in targeting needy groups and set- ting out a plan to ensure coverage for them in the next few years. · Ensuring financial sustainability. This is one of the biggest challenges that needs to be confronted. An actuarial analysis was recently undertaken, but it needs to be refined with utilization data. The sooner this is done and a benefits package is decided on following a proper costing exercise, the better for the scheme. A recent statement by parliament implies that the average person employed in the formal sector would be paying ¢350,000 per year in premiums (nearly US$40), which is steep compared with the 72,000 per ¢ year (US$8) paid by community residents who enroll in an MHO. In addi- tion, collecting premiums from those outside the formal sector will not be easy. Also the definition of indigent could end up affecting the coverage process, as enrolling those defined as nonindigent and collecting premiums from them will be a major hurdle. The high level of dependence on foreign aid is another concern and finding other sources of revenues to sustain the scheme in the long run is important. · Integrating schemes into a single national health insurance scheme. This could be an onerous task as has been evidenced by other countries such as Germany and Korea, especially given the variations in terms of scheme membership, benefits, premiums, and types of providers. · Creating health care infrastructure. Increasing coverage calls for building up the health infrastructure and the provider network. This places dual pressures on the system: ensuring the availability of health care professionals through- out the country and undertaking a rapid accreditation process to certify them. The scheme's success will depend on how early enrollees perceive it and whether they are able to access care. In this respect, the preponderance of health care facilities in urban areas is an issue. To ensure access, the gov- ernment needs to ramp up public facilities, facilitate the setting up of private facilities in rural and poorer areas, and recognize the role of traditional heal- ers. A related challenge lies in developing clear accreditation guidelines and setting a target for implementation. · Safeguarding against crowding out and adverse selection. Competition between schemes could lead to a crowding out of disadvantaged schemes or to fail- ures through adverse selection. The supply-side subsidy to the Ghana Health Service needs to be re-examined along with the issue of indigent subsidies to the district mutual health insurance schemes. Private providers would face a higher cost structure and private insurers would accept only those who could afford to pay higher premiums, while the poor would enroll in district mutual health insurance schemes, leading to adverse selection. The challenge is to structure the subsidy in such a way as to level the playing field between the different kinds of insurers and to ensure proper cross-subsidization and pooling of risk between the rich and the poor. References Aiken, M. 2003. Emerging Community Health Insurance Schemes/Mutual Health Organizations in Ghana: Achievements and Challenges. Accra: Danish International Development Agency. 80 Ghana: Initiating Social Health Insurance Apoya, P., E. Maaweh, A. R. Yakubu, E. Ewuntomah, S. G. Anarwat, and D. Subiniman. 2001. Health Seeking Behaviour and Coping Strategies of Households to Health Financing: Constraints in Northern Region of Ghana. Accra: Ghana Development in Partnerships, Ministry of Health, and Danish International Development Agency Health Sector Support Office. Arhin-Tenkorang, D. C. 1995. Rural Health Insurance: A Viable Alternative to User Fees. London, London School of Hygiene and Tropical Medicine. ------. 2001. Health Insurance for the Informal Sector in Africa: Design Features, Risk Protection, and Resource Mobilization. Report submitted to Working Group 3 of the Commission on Macroeconomics and Health. Geneva: World Health Organization. Atim, C. 2000. Training of Trainers Manual for Mutual Health Organizations in Ghana. Bethesda, MD: Abt Associates. Atim, C., S. Grey, and P. Apoya. 2002. A Survey of Mutual Health Organizations in Ghana. Bethesda, MD: Partners for Health Reform, Plus. Baeza, C., F. Montenegro, and N. Nunez. 2002. Extending Social Protection in Health through Community Based Health Organizations: Evidence and Challenges. Geneva: International Labour Organisation. Bennett, S., A. Creese, and R. Monasch. 1998. Health Insurance Schemes for People Outside Formal Sector Employment. Analysis Research and Assessment Paper 16. Geneva: World Health Organization. CIA (Central Intelligence Agency). 2006. The World Factbook: Ghana. Langley, VA: CIA. http:// www.cia.gov/cia/publications/factbook/. Dzikuna, H., and H. Thorup. 2003. "Health Insurance Policy Development in Denmark 1860 to 1974: Its Relevance to the Emerging Health Insurance Movement in Ghana." Unpub- lished Paper. Danida Health Sector Support Office, Accra, Ghana. Ghana Statistical Service. 2000. Statistics. Accra: Ghana Statistic Service. Government of Ghana. 2003. "Ghana Poverty Reduction Strategy: A Framework for Action in the Ghana Health Service." Draft 7. Government of Ghana, Accra. Huber, G., J. Hohmann, and K Reinhard. 2003. Mutual Health Organization (MHO): Five Years Experience in West Africa--Concerns, Controversies, and Proposed Solutions. Bonn: German Agency for Technical Cooperation. Jakab, M., and C. Krishnan. 2001. Community Involvement in Health Care: A Survey of the Litera- ture on the Impact, Strengths, and Weaknesses. Report submitted to Working Group 3 of the Commission on Macroeconomics and Health. Geneva: World Health Organization. Musau, S. 1999. Community-Based Health Insurance: Experiences and Lessons Learned from East and Southern Africa. Partnerships for Health Reform Project, Technical Report 34. Bethesda, MD: Abt Associates. Inc. Transparency International. 2006. Corruption Perception Index. London: Pluto Press. UNAIDS (Joint United Nations Program on HIV/AIDS). 2006. Epidemiological fact sheets on HIV/AIDS and sexually transmitted diseases. http://www.who.int/hiv/pub/ epidemiology/pubfacts/en/. UNDP (United Nations Development Programme). 2003. Millennium Development Report, 2003: Ghana. New York: UNDP. Waelkens, M. P., and B. Criel. 2004. Les mutuelles de santé en Afrique sub-saharienne: État des lieux et réflexions sur un agenda de recherche. Washington, DC: World Bank and Institute of Tropical Medicine, Department of Public Health, Anvers, Belgium. WHO. World Health Organization. 2006. World Health Report 2006. Geneva: WHO. World Bank. 2005. World Development Indicators 2005. Washington, DC: World Bank. ------. 2006. World Development Indicators 2006. Washington, DC: World Bank. 5 The Philippines: Extending Coverage beyond the Formal Sector Matthew Jowett and William C. Hsiao The Philippines has a wealth of experience in implementing health insurance schemes, little of which is appreciated by the many countries currently in the pro- cess of establishing similar schemes around the world. The National Health Insur- ance Program (NHIP), known as Medicare and modeled on the U.S. system, was initiated in the early 1970s and revamped in the mid-1990s. Local health insurance schemes operated by both local governments and NGOs flourished in the 1980s and early 1990s to fill the gaps in coverage left by Medicare, which focused almost exclusively on those in formal employment. The Philippines also has a vibrant pri- vate health insurance market that provides supplementary coverage for the mid- dle class. A further interesting dimension of the situation in the Philippines is that health insurance has developed within a service delivery system characterized by a rich mix of public and private providers and, since 1991, a far-reaching process of decentralization in the management and financing of government health services. Background ThePhilippines,alower-middle-incomecountry,comprisesmorethan7,000islands located between the Pacific Ocean and the South China Sea (figure 5.1). Ninety per- cent of Filipinos are Christian (80 percent of them Roman Catholic) and account for 60 percent of all Christians in Asia. The Philippines has a long history of coloniza- tion: Spain ruled (through Mexico) for more than 350 years (1542­1899), followed immediately by the United States for a further 50 years (1899­1946). Independence came following liberation from the occupying Japanese forces at the end of World War II. Table 5.1 provides basic statistics regarding the country's demography, econ- omy, health status, and health system. Economy The 1990s saw limited economic growth, and in 1997­98 the economy suffered as a result of both the Asian financial crisis and the negative effects of El Niño. GDP grew at 6.1 percent in 2004, the highest rate in 15 years, underpinned by a posi- tive international economic climate. Manufacturing was the biggest contributor, 81 82 The Philippines: Extending Coverage beyond the Formal Sector Figure 5.1. Map of the Philippines Source: World Bank. expanding by 5.0 percent, with the agriculture and services sectors also registering improved performances. Despite improvements in growth, the economy suffers from high levels of debt--78.6 percent of GDP--second only to Argentina. The Philippines exports more labor than any other country in the world except Mexico. Remittances from overseas Filipino workers amounted to 10 percent of Matthew Jowett and William C. Hsiao 83 Table 5.1. Background Statistics, Philippines Category Statistic Source Demographics (2005) Population (millions) 83.1 World Bank 2006 Growth rate (%) 1.7 World Bank 2006 Urban population (%) 62.7 World Bank 2006 Economy GDP per capita, purchasing power parity (2005) 4,920.0 World Bank 2006 GDP annual growth rate, per capita, real purchasing power parity (past decade) 1.7 World Bank 2006 Percentage of the population below the national poverty line (2003) 37.0 CIA 2006 Percentage of the population living on less than $2/day (2000) 45.1 World Bank 2006 Percentage of the population living on less than $1/day (2000) 14.4 World Bank 2006 Agricultural employment (percentage of the population) (2005) 36.0 World Bank 2006 Corruption (rank out of 163 countries, 2005)a Transparency 121 International 2006 Health status Life expectancy (years) (2004) 70.8 World Bank 2006 Infant mortality/1,000 live births (2004) 26.0 World Bank 2006 HIV/AIDS prevalence (percentage of the population age 15­49) (2005) 0.1 World Bank 2006 Health system Health spending (percentage of GDP) (2003) 3.2 World Bank 2006 Health spending (US$ per capita, purchasing power parity) (2003) 157 World Bank 2006 Physicians/1,000 population (2002) 1.16 World Bank 2006 Hospital beds/1,000 population (2001) 1.0 World Bank 2006 Health system financing (percentage of total health expenditure) (2004) SHI 9.3 WHO 2006 General government revenue 33.3 WHO 2006 Out-of-pocket 44.3 WHO 2006 Private insurance 6.2 WHO 2006 Other private 6.8 WHO 2006 GDP in 2003 (Ratha and Timmer 2006). Currently, many Filipinos are responding to the strong demand for qualified nursing staff in the United States and other coun- tries of the Organisation for Economic Co-operation and Development with rapidly aging populations. Poverty rates are high: estimates indicated that in 2003, 37 percent of the popu- lation was living below the official poverty line. Income distribution is also highly unequal, with an estimated Gini coefficient of 0.466 in 2003. 84 The Philippines: Extending Coverage beyond the Formal Sector Health Overall, the health of Filipinos has improved over the past 14 years, with the infant mortality rate declining from 57 per 1,000 live births in 1990 to an estimated 26 in 2004. However, malnutrition among newborns to children aged five has declined only slightly, from 34.5 percent of children being underweight in 1990 to 32.0 per- cent in 1998. Maternal mortality fell substantially from 209 maternal deaths per 100,000 live births in 1990 to 172 in 1998. The country is on target to meet the Millen- nium Development Goals only in relation to child health. The two leading causes of mortality are heart disease and diseases of the vascular system. Unusually, inju- ries are the fourth leading cause of death, 67 percent of them intentional (stabbings, shootings, and so on). As with income, the distribution of health status across the population exhibits significant inequalities. For example, the infant mortality rate is estimated to be 2.3 times higher among households in the poorest quintile than among those in the richest quintile and the mortality rate among children under five is estimated to be 2.7 times higher in the poorest quintile than in the richest. Similarly, in 2000, life expectancy among adults in the Autonomous Region of Muslim Mindanao was estimated to be comparable to that reached at the national level in 1970. Health Financing and Expenditure At 3.2 percent of GDP, health expenditure in the Philippines is among the low- est levels in the region. Overall public expenditures decreased by eight percent- age points during 1997­2002, principally as a result of falling general government expenditures at both the national and local levels. In 2002, overall peso expendi- tures by the government were down by 18.2 percent compared with 2001, reflecting the country's growing fiscal crisis. Between 2000 and 2002, the share of financing from private sources increased by eight percentage points, reflecting a real increase. In 1995, out-of-pocket payments accounted for 50.4 percent of total health expenditures. These fell to 40.5 percent in 2000, but have risen since then (figure 5.2). Overall, the government accounted for less than half of total health expenditure in 2004 (figure 5.2). Less than 10 percent of total health expenditure was financed through SHI. The majority of expenditure, 57.3 percent, was, however private. Out- of-pocket payments account for the bulk of private health expenditure. The Health Sector Reform Agenda set a target for PhilHealth reimbursements to represent 25 percent of all health expenditures by 2004; in 2002 it represented just over 9 per- cent. PhilHealth has, however, become an increasingly important source of public funding, growing from only 8.9 percent of public funding for health in 1998 to 23.4 percent in 2002. Chronology of Health Insurance Approval of the Philippines Medical Care Act in 1969 was the first step on the road to the NHIP (table 5.2). The Philippine Medical Care Commission, established to manage NHIP was largely successful in implementing the act, which mandated the enrollment of workers in regular employment in both the public and private sectors. In terms of reaching out to the poor and to other workers in the informal economy, however, the program was far less successful. Matthew Jowett and William C. Hsiao 85 Figure 5.2. Composition of Health Expenditures, 1998­2004 100 80 46.3 43.3 40.5 43.9 46.8 44.0 44.3 expenditures 60 5.8 health 6.2 4.9 5.6 5.9 6.1 6.2 6.4 5.8 6.3 6.3 6.4 7.0 6.8 6.8 total 40 3.8 5.0 7.9 9.5 9.3 9.0 of 20 39.1 39.2 40.6 36.2 31.0 34.2 33.3 percentage 0 1998 1999 2000 2001 2002 2003 2004 year out-of-pocket other private private prepaid SHI plans general revenues Source: WHO 2005. Note: Totals may not add up to 100 percent because of rounding. Table 5.2. Milestones in the Development of the NHIP Date Event Phase 1 Covering the formally employed sector August 4, 1969 Philippines Medical Care Act approved 1972 Philippine Medicare Commission formed, start of mandatory enrollment of employees in the formal sector 1991 Local Government Code introduced; dealing with service delivery, enrollment of indigents Phase 2 Shifting toward universal coverage 1995 PhilHealth created to implement Republic Act 7875 1999 Department of Health launches the Health Sector Reform Agenda October 1999 Launch of indigent program December 1999 Equalization of benefits across member type resulting in a 32% increase in benefits for government employees July 2000 Launch of outpatient consultation and diagnostic package December 2001 Plan 500 implemented (enrollment of poor households) February 2002 Implementation of the relative value scale 2001 April 2003 Launch of package for directly observed treatment short-course for tuberculosis May 2003 Launch of maternity and SARS packages July 2003 Launch of the Partnership with Organized Groups Initiative to enroll workers in the informal sector; accreditation of ambulatory surgical clinics February 2004 Launch of Plan 5/25: mass enrollment of indigent households 86 The Philippines: Extending Coverage beyond the Formal Sector Difficulties in extending the NHIP to those working in the informal sector led to growing public concern about the lack of coverage for this group. In response, the late 1980s and early 1990s saw growth in the number of small-scale local health insurance schemes (German Agency for Technical Cooperation 2003). Republic Act 7875, which embodied a clear mandate to reach out to all Filipinos to achieve universal coverage, passed into law in 1995 and represented a further legislative landmark for SHI. A new government-owned and operated corporation, PhilHealth, was established to take this new agenda forward in place of the Philip- pine Medical Care Commission. The new legislation also represented a philosophi- cal shift away from ex post reimbursement of hospital costs toward more proac- tive promotion of health, for example, by extending benefits to cover outpatient services. Nevertheless, PhilHealth still has considerable scope for moving further in this direction in order to become a driving force for good health rather than a purchaser of clinical services. Insurance coverage has grown steadily since 1972, but only since the establish- ment of PhilHealth has some progress been made in extending coverage to the poor and to informal sector workers. For PhilHealth, as well as for similar entities in many other countries, achieving universal coverage depends on the success of strategies to enroll these two groups. An equally challenging task is to ensure that any extension in coverage is sustainable. PhilHealth's target for achieving universal coverage is 2010. A World Bank discussion brief identifies several structural problems in the health system, including the following:1 · the excessively high prices of medicines that lead to inadequate and irratio- nal use; · the insufficient effort expended on prevention, especially of noncommuni- cable diseases; · the excessive reliance on high-end hospital services rather than primary care and outpatient specialist care; · the inefficient organization of the hospital system; · the insufficient quality assurance mechanisms for eliminating poor and wasteful medical practices. In 1999, the Department of Health launched a major initiative to reform the health system, known as the Health Sector Reform Agenda, one pillar of which, health financing, stressed the important role of PhilHealth in driving reforms. Delivery of Health Services The Philippines has a vibrant public and private mix in the delivery of health ser- vices. Supply is regulated through licensing by the Department of Health, while PhilHealth operates a separate process of facility accreditation. While there are more beds licensed by the Department of Health in the public than in the private sector, there are considerably more PhilHealth accredited beds in the private sector than in the public sector. 1. World Bank. "Discussion Briefs." Philippines Country Office. http://siteresources .worldbank.org/INTPHILIPPINES/Resources/DB19-Health-June26.pdf. Matthew Jowett and William C. Hsiao 87 Department of Health Facility Licensing Table 5.3 shows the number of hospital facilities licensed by the Department of Health. Almost two-thirds are in the private sector and nearly half are first-level facilities, that is, relatively small primary-level hospitals. However, when the num- ber of licensed beds is analyzed rather than simply the number of facilities, the dominance of the private sector disappears, as more than 53 percent of licensed beds are in government facilities (table 5.4). More than half of beds in government facilities are in third-level referral hospitals. Similarly, of all licensed beds in pri- vate facilities, 55 percent are in third-level referral facilities. More than 33 percent of all licensed beds in government facilities are located in the national capital region, compared with 27 percent of private beds. Following introduction of the Local Government Code in 1991, ownership of rural health units was transferred to local chief executives as part of the decentral- ization process. PhilHealth Facility Accreditation PhilHealth conducts a separate, independent facility accreditation process and catego- rizes hospitals into primary, secondary, or tertiary hospitals (table 5.5). Approximately 39 percent of the hospitals accredited by PhilHealth are owned by the government, compared with 61 percent owned by the private sector. Table 5.3. Department of Health Licensed Hospital Facilities by Ownership and Level, 2003 Number of facilities Percentage of First Second Third all hospital Ownership level level level Total facilities Government 327 250 82 659 39.6 Private 465 377 164 1,006 60.4 Total 792 627 246 1,665 100.0 Percentage of all hospital facilities 47.6 37.7 14.7 Table 5.4. Hospital Bed Distribution in Department of Health Licensed Hospitals by Ownership and Level, 2003 Number of facilities Percentage of First Second Third all hospital Ownership level level level Total beds Government 6,775 14,261 24,242 45,258 53.2 Private 6,428 11,328 21,293 39,049 46.8 Total 13,183 25,589 45,535 84,307 100.0 Percentage of all hospital beds 15.6 30.4 54.0 88 The Philippines: Extending Coverage beyond the Formal Sector Table 5.5. PhilHealth Accredited Hospital Facilities by Ownership and Level, 2005 Number of facilities Percentage of all hospital Ownership Primary Secondary Tertiary Total facilities Government 291 231 80 602 38.6 Private 389 389 178 956 61.4 Total 680 620 258 1,558 100.0 Percentage of all hospital facilities 43.6 39.8 16.6 The system appears to have too many secondary hospitals, mainly because of the large number of accredited private facilities. Note that PhilHealth does not use the accreditation process as a tool to control supply, but accredits any facility that meets its basic criteria.2 The facility can then begin to make claims from PhilHealth for the treatment of members. Table 5.6 shows the number of PhilHealth accredited beds by hospital facil- ity level and ownership type. The divergence between Department of Health and PhilHealth in terms of accredited beds is greater than that between the number of facilities. Whereas 53.2 percent of Department of Health licensed beds are in gov- ernment facilities, only 44.9 percent of total PhilHealth accredited beds are in such facilities. In addition, PhilHealth accredits 18 teaching hospitals and a number of ambulatory surgical clinics that perform outpatient surgery. Since introducing the outpatient consultation and diagnostic package, which is currently restricted to members of its indigent program, PhilHealth now accredits rural health units. For each indigent enrolled in a rural health unit by a local govern- ment, PhilHealth makes a capitation payment of 300 per year to the local govern- ment for financing and providing this package. By the end of 2004, PhilHealth had accredited 749 rural health units across the country to provide this benefits package. PhilHealth has not accredited any primary care private providers in rural areas. Table 5.6. Hospital Bed Distribution in PhilHealth Accredited Hospitals by Ownership and Level, 2005 Number of facilities Percentage of all hospital Ownership Primary Secondary Tertiary Total facilities Government 4,213 10,160 17,782 32,155 44.9 Private 4,055 10,753 21,801 36,609 55.1 Total 8,268 20,913 39,583 68,764 100.0 Percentage of all hospital facilities 12.0 30.4 57.6 2. Health care institutions must have been operating for at least three years prior to their initial application for accreditation and must conform with PhilHealth's standards in relation to human resources, equipment, physical structure, and other requirements. Matthew Jowett and William C. Hsiao 89 Table 5.7. Facility Accreditation for New PhilHealth Benefit Packages, December 2004 Package Applied Approved Deferred Denied Maternity care clinics 76 74 1 1 Directly observed treatment short-course for tuberculosis 31 29 2 0 Freestanding dialysis clinics 19 8 1 0 PhilHealth has introduced additional benefits packages in recent years, in par- ticular, the maternity package for normal, spontaneous delivery and the directly observed treatment short-course (DOTS) package for tuberculosis. Facilities must undergo a separate accreditation process to qualify as providers of these new ser- vice packages to PhilHealth members (table 5.7). This evolution in the design of ben- efits represents a strategic shift in PhilHealth's thinking, away from simply paying retrospectively for inpatient care and toward a more progressive approach aimed at keeping members healthier, that is, treating problems before they require expen- sive inpatient care. On the positive side, this approach has stimulated a supply-side response. On the negative side, however, it has created an overlap with the pub- lic and primary health functions of the Department of Health, whose funding of health services was more than triple total PhilHealth reimbursements in 2002. In 2005, for the first time in 10 years, PhilHealth revoked the accreditation of three hospitals for filing claims for patients whom they had not admitted and padding claims amounts. Typically, the initial analysis of a facility is strict, with a simple pass or fail outcome, but when PhilHealth is reassessing an accredited facility for renewal, if it does not meet PhilHealth's standards, for example, by hav- ing equipment that is not functioning, PhilHealth may issue a provisional accredi- tation that allows the facility to continue making claims. If the reason for failure is an insufficient number of qualified staff, PhilHealth may re-accredit a facility but with a lower accredited bed capacity, thereby reducing the volume of claims it can make and the level of benefits for members. Registered Health Professionals The Department of Health registers medical professionals (table 5.8). Despite the current huge exodus of health human resources, the ratio of health professionals to population is relatively high compared with other countries in the region. For example, for every 100,000 people, Sri Lanka has about one-third the number of physicians and less than one-fifth the number of nurses than in the Philippines. As noted earlier, the primary demand from richer countries is for qualified nursing staff. Despite this, however, the ratio of 442 nurses per 100,000 people in the Philip- pines is higher than for any other country in the region.3 While many existing staff are leaving for nursing jobs overseas, many school-leavers are training as nurses and workers from other sectors of the economy are being retrained as nurses. 3. This issue requires more detailed analysis, as it is not clear whether the figures include nurses who were trained in the Philippines, but have since moved overseas to work. Furthermore, the average hides an unequal distribution within the Philippines. 90 The Philippines: Extending Coverage beyond the Formal Sector Table 5.8. Health Professionals Registered by the Department of Health, 2004 Category Number per 100,000 population Physicians 116 Nurses 442 Midwives 179 Pharmacists 60 Laboratory technicians 51 Design and Implementation Issues Following the initiation of the NHIP in 1972, the Philippine Medical Care Commis- sion acted principally as a policy-making body, as an accreditor of health facilities, and as an arbiter for claims appeals. The main functions of enrolling the public, collecting contributions, and processing claims were retained by the Government Service Insurance System, which managed government employee members, and the Social Security System, which managed private sector employee members. The creation of PhilHealth merged these functions into a single organization. Even though the chair of the PhilHealth board is the secretary of the Depart- ment of Health (equivalent to a minister of health), the Executive Committee of PhilHealth has considerable autonomy over how PhilHealth organizes and man- ages itself on a day-to-day basis. PhilHealth is, however, still subject to execu- tive and administrative orders issued by the Office of the President,4 but not the Department of Health. PhilHealth can set its own salary scales, although natural limits are built into the system; can spend a maximum of 12 percent of its income from premium contributions on total expenses, including salaries; and must have the salaries of the president and the CEO of PhilHealth approved by the Office of the President, thereby setting a natural limit on the salary scale of the entire orga- nization. The official vision and mission statements of PhilHealth are stated below, followed by several sections detailing key aspects of the NHIP (PhilHealth). A premiere government corporation that ensures sustainable, affordable, and progressive social health insurance which endeavours to influence the delivery of accessible quality health care for all Filipinos. (Vision) As a financial intermediary, PhilHealth shall continuously evolve a sustain- able NHIP that shall lead towards universal coverage; ensure better benefits for its members at affordable premiums; establish close coordination with its clients through a strong partnership with all stakeholders; and, provide effective internal information and management systems to influence the delivery of quality health care services. (Mission) Financial Position PhilHealth is currently in a healthy financial position (figure 5.3). While legally PhilHealth is mandated to maintain a maximum of two years of projected annual 4. For example, Administrative Order 103, issued in 2004, imposed limits on expen- ditures on training, travel, and other line items as a measure to stem growing levels of public debt. Matthew Jowett and William C. Hsiao 91 Figure 5.3. PhilHealth's Financial Position, 1997­2004 50 45 40 35 pesos 30 of 25 20 billions 15 10 5 0 1997 1998 1999 2000 2001 2002 2003 2004 year collections benefits payments investments Source: PhilHealth, Finance Department data. benefit payments in reserve, it maintained the equivalent of 4.08 years in 2004.5 To some extent this picture reflects limits or caps that PhilHealth places on levels of benefit payments. Figure 5.3 shows that both contributions and benefit payments have been increasing steadily since the establishment of PhilHealth. A justification PhilHealth frequently offers for exceeding the two-year reserve rule is the growing risk it faces related to payments to members that are heavily subsidized or exempt from paying premiums. Estimates indicate that 25 percent of benefit payments are made to beneficiaries over the age of 60 and a further 25 percent are paid out to those under the age of 20. However, no detailed financial analysis in terms of projected income and expenditures is publicly available. The fee-for-service approach to provider payment currently used for the bulk of Phil- Health payments is likely to be creating problems of supplier-induced demand. Table 5.9 summarizes PhilHealth's assets as of March 31, 2004 and 2005, most of which are invested in long-term treasury bonds. PhilHealth is subject to fairly stringent rules in terms of how it can invest; for instance, it is not allowed to invest in health facilities given its exclusive role as a purchaser of health services, and as a rule it focuses on low-risk investments. Despite many allegations, no PhilHealth official has ever been convicted of misuse of or misinvestment of funds. PhilHealth estimates collection efficiency based on expected versus actual pre- mium contributions by membership category. In the public sector, contributions by government employers, as determined by the Department of Budget Management, 5. This is likely to be a slight overestimate, as no value of projected benefit payments was available, and hence figures for the 2004 were used. 92 The Philippines: Extending Coverage beyond the Formal Sector Table 5.9. PhilHealth Financial Statement, March 31, 2004, and March 31, 2005 (billions of pesos) 2004 2005 Short-term investments 11.8 11.7 Treasury bills 6.6 10.2 Special savings deposits 5.3 1.5 Property, plant and equipment 0.6 1.1 Land 0.5 0.9 Long-term investments 30.9 36.9 Treasury bonds 30.3 36.8 have remained unchanged since 2001 despite changes approved by PhilHealth. Fig- ure 5.4 shows how employee contributions rose, but employer contributions stag- nated, with growing arrears as a result. Employees and employers are supposed to contribute equal amounts. For example, those government employees earning 20,000 or more per month pay 250.00 as their contribution to PhilHealth, while their employers contribute only 62.50. Clearly, this shortage of 187.50 represents a significant loss of revenue to PhilHealth. The situation is less clear among employees of private firms, although critics have frequently suggested that employers find ways of evading making contri- butions on behalf of their employees. Estimates suggest that in the Individual Payment Program (IPP), only one-third of those registered under the scheme are Figure 5.4. Contributions by Government Employees and Government Employers, 2000­2005 500 450 400 187.5 350 125 300 62.5 250 pesos 62.5 62.5 62.5 200 31.25 150 62.5 62.5 62.5 100 62.5 50 62.5 93.75 125 125 187.5 250 0 2000 2001 2002 2003 2004 2005 year amount employer should have funded but did not employer employee Source: PhilHealth, Finance Department data. Matthew Jowett and William C. Hsiao 93 making payments on a regular basis, a problem that is being tackled by strength- ening group enrollment. Enrollment in the indigent program varies from year to year, depending on the availability and prioritization of public budgets, but has been boosted by considerable political support in recent years. Enrollment and Coverage Rates Membership in the NHIP has increased steadily since it began operations in 1972 (figure 5.5). The figure shows that private sector employees have formed the bed- rock of the program. The sudden drop in membership between 1986 and 1987 resulted from a reduction in private sector enrollees from 23.6 million to 16.2 mil- lion. Discussions with officials who were working with the NHIP at the time sug- gest that the popular uprising against the military junta in 1986, which resulted in a change of political leadership, also led to changes in the management of the Social Security System. As a result, the Social Security System conducted a review of the membership database, followed by a cleansing process to deal with what was seen as a bloated list of members. PhilHealth membership is segregated into four categories as discussed in the following subsections. EMPLOYED PROGRAM. Membership in the NHIP is compulsory for all govern- ment and private sector employees, including household help and Filipinos work- ing on ships. Contributions are currently set at 2.5 percent of employees' salaries, divided equally between employers and employees. A salary cap is set at 25,000 per month, that is, contributions do not increase when earnings reach this level or beyond. The salary cap has increased considerably in recent years and will con- tinue to rise in the future. By bringing contributions more in line with ability to pay, financing is becoming more equitable. Figure 5.5. Enrollment in the NHIP, 1972­2005 90 TOTAL POPULATION 80 The area under the total 70 population line and above Indigent/sponsored the four insured groups program (millions) 60 represents the uninsured Individual 50 PhilHealth established payment program under the 1995 National IPP coverage 40 Health Insurance Act 30 (government) 20 Employed (private) population Employed 10 0 1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 year Source: PhilHealth, Corporate Planning Department data. 94 The Philippines: Extending Coverage beyond the Formal Sector Employers automatically deduct and withhold monthly contributions and remit them to PhilHealth. Evasion is reportedly a major issue, particularly among small shops and businesses. While accurate estimates of the extent of the problem are not available, the Office of the Actuary estimates that collection efficiency is as low as 30 percent, that is, 70 percent of those who should be contributing are not doing so. INDIGENT PROGRAM. The enrollment of indigents into PhilHealth is initiated by local governments (cities or municipalities), which are responsible for identi- fying them. The local governments then pay their share of the indigents' annual premiums of 1,200 each. The share varies with the class of LGU, which depends on its total income rather than on the extent of poverty in the LGU, and this share changes over time, eventually rising to 50 percent for all local governments. (The remainder of the premium is paid directly to PhilHealth by the central Department of Budget Management. Indigent enrollment was a major issue during the 2004 national elections. Presi- dent Gloria Macapagal Arroyo, running for re-election, launched Plan 5/25, which aimed to enroll 5 million indigent families, or 25 million beneficiaries, into Phil- Health. To do so, funds were earmarked from the Philippine Charity Sweepstakes Office, which paid these families' premiums in full, that is, without any LGU con- tributions. While controversial, the program was successful in boosting numbers. However, the downside to the Plan 5/25 initiative was that technical difficul- ties resulted from the creation of a second category of indigents. When local gov- ernments pay part of the premium, indigent members are eligible for PhilHealth's additional outpatient consultation and diagnostic package and the local govern- ments receive a capitation payment of 300 per member in return from PhilHealth, but those indigents funded under Plan 5/25 were not eligible for the outpatient consultation and diagnostic package, as there was no local government sharing. Furthermore, a problem has arisen in that some local governments are not mak- ing their contributions in the expectation that the Philippine Charity Sweepstakes Office will continue to pay the full amount. Sustaining this rapid increase in indigent members is now, not surprisingly, a major issue. While some success in maintaining funding has been achieved by earmarking a portion of the recently introduced so-called sin taxes, the numbers covered were expected to fall in 2005.6 To a large extent, the initiative was closely associated with the election campaign of the incumbent president, and as such was subject to accusations of electioneering. As a result, the program runs some risk of not having full political support in the future. An alternative approach proposed by PhilHealth is the possibility of deducting premium contributions from internal revenue allotments to local governments in the same way that contributions are deducted from salaries at source; however, this is not considered politically fea- sible. A more positive development is new legislation that earmarks 4 percent of recently increased value added tax receipts to fund PhilHealth premiums for indi- gents. Specifically, these funds would be used to cover that portion of the indigent premium that local governments currently pay. Estimates indicate that this new source of funding would amount to 2 billion per year, but PhilHealth would only 6. PhilHealth officials estimated that the figure would drop from 5 million families (25 million beneficiaries) to 2 million families (10 million beneficiaries). Matthew Jowett and William C. Hsiao 95 receive these funds from 2008 onward. A further positive development is the ear- marking of a portion of the additional revenues resulting from the recent increase in sin taxes. The 2.5 percent of incremental receipts is estimated to be worth just over 100 million per year to PhilHealth, and will be used to cover the central gov- ernment counterpart for indigent premiums. However, this arrangement is only valid for the period 2005 to 2010. These developments are a further indication that, following the use of charity sweepstakes funds in 2004, the funding of indigent premiums is slowly being cen- tralized. This in turn reflects growing recognition that local governments will not adequately prioritize PhilHealth enrollment for their poor constituents. INDIVIDUAL PAYMENT PROGRAM. Those not eligible for either the employed pro- gram or the indigent program can join the NHIP voluntarily through PhilHealth's IPP. This program targets nonindigent informal sector workers, and their annual premium of 1,200 is currently the same as that for indigents. The International Labour Organisation estimates that 50 percent of the working population in the Philippines is in the informal sector. Regularity of payments among this group is a major problem: most IPP mem- bers pay on a quarterly basis, but around two-thirds of members are not paying the required amount on a regular basis. For example, a member may join and pay an initial quarterly payment, miss the next quarterly payment, but pay the following contribution, and so on. This target group is categorized by uncertain and variable income over the course of the year, leading to instability in finances. The lack of regular payments also presents providers with administrative complications when verifying eligibility. PhilHealth is currently considering a proposal to divide those currently eligible for the IPP into several groups and to vary the premium for each. The aim would be to bring premium contributions more in line with ability to pay, given that the IPP targets a heterogeneous group that ranges from relatively wealthy professionals to relatively poor farmers. NONPAYING PROGRAM. PhilHealth's nonpaying program targets those who have reached the age of retirement, as provided for by law, and have paid at least 120 monthly premium contributions to PhilHealth. No contributions are made for this increasingly high-risk group either by themselves or by the government on their behalf. It thus represents a growing risk for PhilHealth from a financial perspective. COVERAGE ESTIMATES. Figure 5.6 breaks down NHIP membership to show the various membership categories. The figure clearly shows the rapid increase in the indigent program from 16 percent of total membership in 2003 to 48 percent in 2004. At the end of 2005, PhilHealth estimated national coverage to be 81 percent, but by September 2005, the official figure was 63 percent, due almost entirely to a fall in the enrollment rate of indigents. Figure 5.7 presents preliminary estimates of cover- age rates by employment sector based on membership data and calculations of tar- get group size using employment structure data from the National Statistics Coor- dination Board. As the figure demonstrates, the greatest gap in coverage is among informal sector workers, followed by indigents and private sector members. 96 The Philippines: Extending Coverage beyond the Formal Sector Figure 5.6. Changes in Membership Composition, 2003­2005 50 48 48 43 40 36 membership 30 total 22 22 of 20 15 16 17 14 12 centage 10 per 4 2 0 1 0 private government indigent IPP nonpaying members members end of 2003 end of 2004 September 2005 Source: PhilHealth, Corporate Planning Department data. ENROLLMENT IN THE INFORMAL SECTOR. Voluntary enrollment by individual householdsleadstothepredictableproblemofadverseselection,commonlyobserved in insurance markets, of which there is some indication in PhilHealth as evidenced by trends in claims. PhilHealth is currently testing a new initiative designed to limit problems of adverse selection and irregular premium contributions. The essence of the strategy is to offer an incentive to groups such as microfinance and cooperative Figure 5.7. Estimated Coverage Rates by Member Sector, 2004 35 35 30 30 target actual 25 23 members 20 oup 20 gr 15 15 15 of 15 10 centage 7 per 5 0 private formal public formal informal indigent sector sector sector Source: PhilHealth, Corporate Planning Department data. Matthew Jowett and William C. Hsiao 97 organizations if they deliver a minimum of 70 percent of their eligible members into PhilHealth's IPP. The hope is that all those involved will benefit as follows: · PhilHealth.Movingawayfromindividualandtowardgroupenrollmentshould limit adverse selection and reduce dropouts, thereby helping PhilHealth to move toward universal coverage in a more sustainable way. If the minimum 70 percent enrollment threshold is not met, no group premium rate is applied and the partner organization receives no income (see the next bullet). If individuals who joined the scheme start to drop out and enrollment levels fall below the 70 percent threshold, again no discount is applied and the group will receive no income. This approach gives a clear incentive to the group both to enroll a large percentage of its members into the IPP and to prevent dropouts. By only entering into partnerships with groups having at least 1,000 eligible members, PhilHealth will also achieve gains in terms of administrative efficiency. · Organized groups. The partnership will allow groups to offer a new product to their clients, thereby serving their needs more effectively. In most cases, the organization is expected to continue charging the full premium to its members, but to remit a discounted group premium amount to PhilHealth, keeping the difference as income to cover administrative costs or use at its discretion. For example, any surplus might be returned to members as divi- dends, might be used for additional health benefits, or might be invested. PhilHealth does not set any conditions on how an organization uses these funds, although it does make ideas, advice, and technical support available. In addition, research shows that a major proportion of bad loans among microfinance organizations results from their members falling ill and facing high hospital bills and expenditures on medicines. Thus the hope is that this strategy will contribute to the financial stability of partner organizations. · Members. Research has shown that for informal sector workers, flexibility in payment is more important than the premium level of 1,200, therefore the success of the strategy will depend on the extent to which the partner organization is innovative in terms of introducing flexible payment methods for its members and in assisting members when they face problems making payments, for instance, by means of savings schemes. Thus PhilHealth is tar- geting organizations that are well managed and have extensive and effective loan collection systems. If the managers of the partner organization decide to pass on some of the discount, the members may also obtain cheaper access to PhilHealth through this initiative than would otherwise be the case. Member Benefits Benefits under the NHIP are principally, but not exclusively, related to inpatient care. Under PhilHealth's implementing rules and regulations, the scope of benefits is defined as follows: · inpatient hospital care · room and board charges · fees of health care professionals · diagnostic, laboratory, and other medical examination charges · charges for the use of surgical and medical equipment and facilities · prescription drugs 98 The Philippines: Extending Coverage beyond the Formal Sector · outpatient care · fees of health care professionals · diagnostic, laboratory, and other medical examination charges · personal preventive services · prescription drugs · health education packages · emergency and transfer services · other health care services that PhilHealth deems are appropriate and cost-effective. PhilHealth excludes the following services from its benefits package: · fifth and subsequent normal obstetrical deliveries · nonprescription drugs and devices · treatment for alcohol abuse or dependency · cosmetic surgery · optometric services · procedures that PhilHealth deems are not cost-effective Benefits are portable nationwide, that is, they can be accessed at any PhilHealth accredited hospital facility in the country, and are defined in terms of a series of ceilings (table 5.10). In principle, PhilHealth Regional Offices (PRO) can tailor the benefits package to their local area, provided that the overall value to the patient does not change. PhilHealth has introduced several new packages in recent years, including the maternity care package for normal spontaneous delivery in 2003. A case rate of 4,500 is paid to accredited health care providers regardless of the length of hospital stay, of which 2,000 is allocated to the health professional and 2,500 to the health facility, which is expected to cover costs related to room and board, drugs, diagnos- tic procedures, operating room expenses, and all other medically necessary care. Also in 2003, PhilHealth introduced an outpatient package for DOTS for tuber- culosis, under which a flat payment of 4,000 per case is paid to an accredited DOTS facility that is expected to cover diagnostic and consultation services and drugs. A first payment of 2,500 is made after the accredited DOTS facility has completed the intensive phase of DOTS treatment and a final payment is made at the end of the maintenance phase. PhilHealth extended the reimbursement of dial- ysis services to free-standing dialysis centers in 2003 and launched a SARS package in response to the regional crisis between 2003 and 2004, under which a maximum of 50,000 per case is paid. FINANCIAL PROTECTION. The expectation was that PhilHealth would cover approximately 70 percent of patients' total hospital costs,7 compared with the esti- mated 30-45 percent at the time. In practice, the situation is highly variable and uncertain. A review of the Health Sector Reform Agenda by Solon, Panelo, and Gumafelix (2002) expects PhilHealth to periodically increase the benefit expendi- ture ceiling, but notes that there is no clear estimate of the support value of the 7. Commentators suggest that the ideal rate of copayment is between 15 and 25 per- cent of the total cost of care to simultaneously meet objectives of efficiency and financial protection. Matthew Jowett and William C. Hsiao 99 Table 5.10. PhilHealth Member Benefits and Cost Ceilings, 1999 (pesos) Hospital category Benefits Primary Secondary Tertiary Room and board not exceeding 45 days per 200 300 400 year for each member and another 45 days to be shared by the member's dependents Drugs and medicines per single period of confinement · Ordinary 1,500 1,700 3,000 · Intensive 2,500 4,000 9,000 · Catastrophic 0 8,000 16,000 X-rays, laboratory changes, etc. per single period of confinement · Ordinary 350 850 1,700 · Intensive 700 2,000 4,000 · Catastrophic 0 4,000 14,000 Professional fees per single period of confinement shall not exceed · General practitioner (per day) 150 150 150 · Specialist (per day) 250 250 250 · Ordinary General practitioner 600 600 600 Specialist 1,000 1,000 1,000 · Intensive General practitioner 900 900 900 Specialist 1,500 1,500 1,500 · Catastrophic General practitioner 900 900 900 Specialist 1,500 1,500 2,500 Other · Operating room Relative value units of 30 and below 385 670 1,060 Relative value units of 31 to 80 0 1,140 1,350 Relative value units of 81 and above 0 2,160 3,490 · Surgeon Maximum of 16,000 · Anesthesiologist Maximum of 5,000 benefits. The problem of assessing the actual support value provided by PhilHealth results from two aspects of the NHIP's design, namely, the design of PhilHealth benefits, which compensate patients for expenses incurred up to a ceiling (or "first- peso coverage") but not above that ceiling, and an approach whereby providers are allowed to charge what they want (or what the market will bear). 100 The Philippines: Extending Coverage beyond the Formal Sector As a result, patients are subject to a high degree of uncertainty about what the cost of care will be, how much PhilHealth will cover, and how much they must pay out-of-pocket. Generalizing about the support value offered by PhilHealth is dif- ficult. In many cases, for example, public hospitals in rural areas, PhilHealth ben- efits will be enough to cover 100 percent of the cost of treatment, but for treatment received in private hospitals in large cities, PhilHealth may only cover a small per- centage of the hospital bill, leaving the patient to shoulder the balance. A recent internal analysis conducted by PhilHealth calculates that the average support value of PhilHealth benefits nationwide is 62 percent, ranging from an average of almost 88 percent in government hospitals to just over 53 percent in private hospitals. However, further studies are required to validate these findings. Low financial protection diminishes the value of joining PhilHealth and can exacerbate adverse selection, with the danger that lower-risk individuals will elect not to join the program. Understandably, PhilHealth faces pressure to translate its considerable reserves into enhanced benefits, but without changing the design of the benefits package and the way that PhilHealth pays providers, such a move may simply benefit providers rather than members. BENEFITS PACKAGE DESIGN AND PROVIDER PAYMENT. A research study conducted in 1991, which analyzed survey data collected from randomly sampled patients in 132 hospitals across the country by the Philippine Institute of Development Studies on behalf of the Department of Health, demonstrates how unregulated private pro- viders increase their prices in order to capture benefit payments from PhilHealth (as reported in Gertler and Solon 2002). The study estimated that private hospitals mark up prices to insured patients by 23.4 percent, with private patients not having insurance charged a 60 percent markup. Overall, the study calculates that hospi- tals extract 86 percent of PhilHealth benefit payments through price discrimina- tion (profits or rents), with only 14 percent going toward financing patient care. The study illustrates how, when PhilHealth increases the caps for its benefits package, private providers simply increase their prices in response. Without some ex ante certainty about the final price charged for the care delivered, a decision by Phil- Health to increase benefit payments will not necessarily result in great financial protection for members. Under the first-peso approach to benefits, PhilHealth limits the financial risk to itself, pushing it onto the patients, who end up absorbing any charges over and above the benefit ceiling. Designing benefits in this way allows PhilHealth to man- age the risk to its finances, but without exerting its power over providers, and exposes its members to potentially poverty-inducing medical expenses. One option would be to reverse the design of the benefits package, with the patient making an initial deductible payment and PhilHealth covering the "second peso," that is, capping the amount members pay.8 However, without negotiation about the final 8. In Vietnam, for example, social insurance fixes providers' fee schedules and covers 80 percent of the total cost of care, leaving the patient with a 20 percent copayment. How- ever, once a patient has paid the equivalent of half his or her annual salary, the patient is not expected to make any more payments out-of-pocket. Matthew Jowett and William C. Hsiao 101 price, the same problem may continue, with the possibility that additional charges beyond the "second peso" level will fall onto the patient. In terms of provider payment, PhilHealth makes most reimbursements to hos- pitals on a fee-for-service basis. Health facilities are accredited and the cost of treat- ment is reimbursed up to a certain ceiling, or cap, depending on the type of care pro- vided (ordinary, intensive, or catastrophic) and in which type of hospital (primary, secondary, or tertiary). Special agreements exist with nonaccredited providers for emergency care. Recent years have seen some positive developments. PhilHealth now uses capitation payments to fund the outpatient consultation and diagnostic package and fixed case payments for the recently introduced DOTS and maternity packages. CROSS-SUBSIDIZATION AND SUSTAINABILITY. Figure 5.8 summarizes data on rev- enues through contributions and expenditures through reimbursements disaggre- gated by membership category. Despite some missing data, the figure shows that private employees are the major net contributors to the NHIP, while nonpaying members and IPP members are the main net recipients. Clearly indigent households are also major net recipients, given that the government pays the premiums on their behalf. Through claims payments, government employees receive almost exactly what they put in through premium contributions. Overall the scheme does exhibit cross-subsidization, with those in regular employment subsidizing the less well-off. In terms of the long-term sustainability of the NHIP, fraud poses a problem, principally in terms of claims made for treatment never provided and in terms of inflated prices. PhilHealth considers the problem to be serious, and it is one that tends to be exacerbated when providers are reimbursed on a fee-for-service basis. Figure 5.8. Cross-Subsidization across Member Categories, 2004 10 9 contributions 8 benefit payments 7 pesos 6 of 5 4 billions 3 2 1 0 public private IPP indigent nonpaying formal formal members sector sector Source: PhilHealth, Corporate Planning Department data. 102 The Philippines: Extending Coverage beyond the Formal Sector Figure 5.9. Trends in Claims by Member Category, 2002­2005 30 public formal sector a private formal sector 25 IPP member indigent having 20 nonpaying members oved Jan.­March 2005 members appr 15 of claim 10 centage 5 per 0 2002 2003 2004 2005 year Source: PhilHealth, Corporate Planning Department data. The Office of the Actuary estimates that between 10 and 20 percent of claims are fraudulent.9 Figure5.9plotsrecenttrendsintheclaimsand figure5.10plotstheaveragevalue of claims, in both cases disaggregated by member category. A noteworthy aspect of figure 5.9 is the rapid increase in claims by nonpaying members, including those age 60 and over, who are, by definition, likely to have greater medical needs than Figure 5.10. Average Value of Claims by Member Category, 2002­2005 9,000 public formal sector private formal sector 8,000 IPP member (pesos) indigent 7,000 nonpaying members claim paid 6,000 of 5,000 value 4,000 average 3,000 2002 2003 2004 2005 year Source: PhilHealth, Corporate Planning Department data. 9. The chief actuary of an international private health insurance company operating in the Philippines estimates that at least 10 percent of claims are fraudulent. Matthew Jowett and William C. Hsiao 103 the average population. Nevertheless, given the relatively small size of this group, the trend is extremely worrying and potentially financially damaging for the NHIP. The claims rate for other member categories is generally stable, although the rate has more than doubled for IPP members since 2002. In terms of the average value of claims, an upward trend is evident across all member categories since 2002. Most worrying, once again, is the trend among non- paying members, which has shown an increase of almost 50 percent since 2002. The average claim value by indigent members is also increasing rapidly, although it remains substantially lower than claims made by other members. Taken together, these two trends suggest that at the current level of premiums, and given the underpayment of contributions by other government departments, the scheme is not sustainable in the medium to long term. The principle approach taken in other countries is to innovate with provider payment mechanisms, for example, by using global budgets. Impact of SHI Attributing changes in health status or poverty levels to the introduction of a national health insurance program is difficult; however, one can argue that the NHIP has had the following effects: · Raising health as a political priority. Through its indigent program, PhilHealth has advanced the issue of access to health services up the political agenda, both at local and central government levels. This in turn has helped secure more funding for health insurance coverage for the poor. · Providing stable funding for the health sector. With its own earmarked source of funding through premium contributions, PhilHealth is not dependent on annual government budget allocations, which are currently facing severe cutbacks in response to high levels of public debt. While premium contributions rise and fall with the performance of the economy, the estab- lishment of the NHIP has nevertheless led to a more reliable, stable, and sustainable source of funding for the continued reimbursement of claims for health services provided relative to a system funded out of regular gov- ernment budgets. · Improving health information. Although limited to those individuals enrolled in the NHIP, PhilHealth collects detailed information on disease patterns and other health-related problems through its claims processing information system. Despite limitations, the database has the advantage of being up-to- date compared with data generated through facility-based routine data col- lection and sample surveys, which are often several years out-of-date. · Establishing a dynamic public sector organization. The establishment of Phil- Health as a separate, independently managed organization has created a relatively dynamic public sector institution open to new ideas and prepared to test them. While an urgent need for further improvements in many policy areas remains, PhilHealth is well positioned both financially and in terms of its legal framework to drive reforms across the health sector given the right leadership and management team. · Improving access to services and financial protection for the poor. The extension of health insurance to a large number of poor households through both 104 The Philippines: Extending Coverage beyond the Formal Sector the indigent program and the IPP program has led to greater access and financial protection for poorer segments of society. The extent of this effect, however, has been limited, given the persistence of relatively low service utilization rates among those groups. This in turn reflects the fact that for the poor, indirect costs, such as those incurred for transport and out-of- pocket payments required over and above PhilHealth benefits, remain a deterrent to seeking care. · Enhancing equity in access to health care. The policy decision made in 1999 to equalize benefits packages across member categories strengthened equity in access to health care. Again, however, the impact of this decision has been limited, given the lower levels of utilization among poorer households. · Raising the quality of healthy services. While difficult to measure, the establish- ment of a separate accreditation process under PhilHealth has probably led to some improvements in the quality of health services over and above the licensing process conducted by the Department of Health. In the future, however, the potential for PhilHealth to use its influence as a purchaser to further drive up the quality of health services is considerable. References CIA (Central Intelligence Agency). 2006. The World Factbook. McLean, VA: CIA. http://www. cia.gov/cia/publications/factbook/. German Agency for Technical Cooperation. 2003. Resource Book of Community-Based Health Care Organization Social Health Insurance Schemes in the Philippines. Eschborn, Germany: German Agency for Technical Cooperation. Gertler, P., and O. Solon. 2002. "Who Benefits from Social Health Insurance in Developing Countries?" Working paper, University of California Berkeley. PhilHealth (Philippine Health Insurance Corporation). 2004. The PhilHealth Chronicles: The Journey Towards Universal Social Health Insurance. Manila: PhilHealth, Corporate Planning Department. ------. "About Us." PhilHealth. http://www.philhealth.gov.ph/about_us.htm. Date con- sulted: November 16, 2006. Ratha, D., and H. Timmer. 2006. Global Economic Prospects 2006. Washington, DC: World Bank. Solon, O., C. Panelo, and E. Gumafelix. 2002. A Review of the Progress of Health Sector Reform Agenda Implementation. Manila: Department of Health. Transparency International. 2006. Corruption Perception Index. London: Pluto Press. WHO (World Health Organization). 2005. Social Health Insurance: Selected Case Studies from Asia and the Pacific. Geneva: WHO. ------. 2006. World Health Report 2006. Geneva: WHO. World Bank. 2006. World Development Indicators 2006. Washington, DC: World Bank. 6 Colombia: Social Health Insurance with Managed Competition to Improve Health Care Delivery Diana Pinto and William C. Hsiao Colombia's national health insurance scheme came into existence with health reforms implemented in 1993. The reforms paved the way for transformation of the health care delivery system and brought in a managed competition model. As the health insurance scheme has now been in place for a number of years, an analysis of the scheme provides valuable insights about the design and implementation of universal health insurance that may be applicable to other countries. Background Colombia is a lower-middle-income country located in northern South America, bor- dering the Caribbean Sea between Panama and Venezuela and the northern Pacific Ocean between Ecuador and Panama (figure 6.1). It is the fourth largest country in South America. Colombia is divided into 32 administrative states, known as depart- ments, which in turn are divided into 1,092 municipalities. There are also four dis- tricts that correspond to major cities. Seventy percent of municipalities are rural and have fewer than 20,000 inhabitants each, and more than 60 percent of the population lives in the six largest urban municipalities. Table 6.1 provides basic statistics regard- ing Colombia's demography, economy, health status, and health system. Economy Untilthemid-1990s,Colombiaexperiencedeconomicgrowthandhadoneofthemost stable economies in Latin America. However, by 1997, the country entered a period of economic recession that hit a low point in 1999, when economic growth rates were negative. The economy recovered, but the impact of the recession on poverty, employ- ment, and public social expenditures was huge, for example (Lasso 2004): · The percentage of the population living in poverty increased from 53.2 percent in 1995 to 58.2 percent in 1999. By 2005, poverty had dropped to 49.2 percent. · The urban unemployment rate increased from 10.0 percent in 1996 to 20.1 percent in 2001. · The level of public social expenditures as a percentage of GDP increased from 6.8 percent in 1990 to 15.3 percent in 1997, but subsequently decreased to 13.6 percent during 2000­1. 105 106 Colombia: Social Health Insurance with Managed Competition to Improve Health Care Delivery Figure 6.1. Map of Colombia Source: World Bank. Health Colombia's health profile fits what has been described as the double burden of dis- ease, which is characterized by an increasing incidence of the chronic and degen- erative diseases typical of developed countries and the persistence or resurgence of infectious and parasitic diseases such as malaria and tuberculosis. Colombia ranks fourth among Latin American countries with respect to the total number of Diana Pinto and William C. Hsiao 107 Table 6.1. Background Statistics, Colombia Category Statistic Source Demographics (2005) Population (millions) 45.6 World Bank 2006 Growth rate (%) 1.5 World Bank 2006 Urban population (%) 72.7 World Bank 2006 Economy GDP per capita, purchasing power parity (2005) 7,769 World Bank 2006 GDP annual growth rate, per capita, real purchasing power parity (past decade) 0.8 World Bank 2006 Percentage of the population below the national poverty line (2005) 49.2 CIA 2006 Percentage of the population living on less than $2/day (2003) 19.4 World Bank 2006 Percentage of the population living on less than $1/day (2003) 7.0 World Bank 2006 Agricultural employment (percentage of the population) (2005) 22.7 World Bank 2006 Corruption (rank out of 163 countries, 2005)a 59 Transparency International 2006 Health status Life expectancy (years) (2004) 72.6 World Bank 2006 Infant mortality/1,000 live births (2004) 17.5 World Bank 2006 HIV/AIDS prevalence (percentage of the population age 15­49) (2005) 0.6 World Bank 2006 Health system Health spending (percentage of GDP) (2003) 7.6 World Bank 2006 Health spending (US$ per capita, purchasing power parity) (2003) 590 World Bank 2006 Physicians/1,000 population (2002) 1.35 World Bank 2006 Hospital beds/1,000 population (2003) 1.1 World Bank 2006 Health system financing (percentage of total health expenditure) (2003) SHI 55.5 WHO 2006 General government revenue 28.6 WHO 2006 Out-of-pocket 7.5 WHO 2006 Private insurance 8.4 WHO 2006 Other private 0.0 WHO 2006 reported cases of HIV/AIDS. Estimates of the number of people living with AIDS in Colombia range widely from 82,000 to 160,000. The period between 1990 and 1998 saw decreasing rates of AIDS cases and mortality and rising rates of asymptomatic HIV infections (UNAIDS). Health Financing and Expenditure Total health expenditures as a percentage of GDP grew from 6.2 percent in 1993 to 9.6 percent in 1997, but contracted to 7.7 percent during the economic recession. 108 Colombia: Social Health Insurance with Managed Competition to Improve Health Care Delivery Health spending has stabilized at around 8 percent of GDP since that time. Social security expenditures grew throughout this period to become the largest financ- ing source. Private financing has decreased, with out-of-pocket spending dropping from 44% of total health expenditure in 1993 to 7.5% in 2003 (Baron 2005). Public and Private Provision of Health Services In the 1970s, as part of an initiative to extend basic services to all Colombians, the Ministry of Health invested in the creation of a network of public providers, which was set up at three levels based on complexity of services and catchment area. First- level care included health posts, health centers, and hospitals that provided general medicine. Second-level care included hospitals that provided basic specialty medi- cal and some surgical services. Third-level care included institutions that provided specialty and subspecialty care and hospitalization for complex cases. The private market for health care services grew in parallel to this structure, targeting the pop- ulation that had the ability to pay and meeting the demand for services and quality not provided by the public sector. The 1993 reform integrated public and private providers into the national health insurance scheme and organized all institutions by the three levels of care. Provisions in the reform also required public institu- tions to transform their management and budgeting structures so that they could compete with private sector institutions. As of December 2003, 54,778 health care providers were registered with the min- istry's database of certified providers. Of these 41,000 were individual physicians and the remainder were health care institutions at various levels of complexity. Of the registered health care institutions, 32 percent were public providers, of which 80 percent were primary care providers, 17 percent were secondary care providers, and 3 percent were tertiary care providers. Of the total number of public hospital beds, 60 percent were at the secondary and tertiary levels. Following a government decree issued in 2002, all health service providers must meet a set of minimum quality, financial, and administrative standards defined by the Ministry of Social Protection (a new entity created in 2003 that combined the for- mer Ministry of Health and Ministry of Labor) in order to operate. To obtain a three- year certification, providers register with the local health authority, which then car- ries out an inspection visit to verify that the providers meet all required standards. Health Care Reform As noted earlier, prior to 1993, Colombia had a three-tiered health care system. This system consisted of the public provision of health care for about 65 percent of the population, a mandated social insurance plan that financed and operated its own facilities for formal sector workers, and a system of private insurance and health care provision for those able to pay out-of-pocket. This system resulted in limited access to even basic health services for a large proportion of the population, opera- tional inefficiencies at all levels of care, and poor quality of services. All these con- tributed to low use and acceptance of the network of public providers. The 1993 Reform and Strategy The problems inherent in the three-tiered system motivated a major reform in 1993 based on Law 100. The law transformed the organization, financing, and delivery Diana Pinto and William C. Hsiao 109 of the health care system in an attempt to improve access to services and their effi- ciency and quality, along with equity. It mandated the creation of a new system for financing and delivering health care whereby public subsidies were to be allo- cated directly to individuals instead of to institutions. The reform introduced four main elements to reach the poor (Escobar 2005): (a) a proxy means testing index to target the allocation of public subsidies (SISBEN);1 (b) the transformation of the tra- ditional supply-side subsidies, which financed the public health care network, into individual insurance premiums for the poor subsidized by the system; (c) an equity fund with financial flows allowing for payroll contributions and treasury resources to cross-subsidize the insurance premium for the poor; and (d) the contracting of health service delivery from both the public and private sectors. The strategy used to improve access and equity was based on the concepts of managed competition and decentralization. Managed Competition To improve efficiency and quality, national health insurance was organized as a dual market managed competition model, as illustrated in figure 6.2. The first mar- ket is an insurance market, where consumers are free to enroll in their choice of public or private health insurance plans. These plans offer services included in a benefits package in return for a risk-adjusted premium that is fixed for all plans at the national level. The second market is the provider market, whereby health plans act as group purchasers for their enrollees by selecting a network of provid- ers based on price and quality. As the premium is fixed, health plans compete for enrollees on the basis of the service and quality features of their benefits packages. In this model, the government has the role of facilitating competition between plans Figure 6.2. Managed Competition Model for National Health Insurance Government consumers Insurance market: consumer choice, Regulations health plan Benefits quality Premium competition Minimum quality standards Enrollment rules Information Price health plans Provider market: Quality provider price and quality Oversight competition providers Source: Authors. 1. The original SISBEN instrument consisted of 62 variables related to aspects of hous- ing, public services, family structure, labor participation, education, insurance, recreational activities, and so on. Each household was assigned a score and classified into one of six lev- els. The instrument was modified in 2003 to improve its sensitivity and specificity. 110 Colombia: Social Health Insurance with Managed Competition to Improve Health Care Delivery by creating a level playing field. It does this by providing information about price and quality and by formulating, monitoring, and enforcing regulations concerning benefits, premiums, quality, enrollment, and standards. Decentralization Implementation of the reform required new roles for local governments and the existing public provider network. The stage for this was partly set up at the time Law 100 was enacted, because in the late 1980s, Colombia's health sector began to participate in a nationwide process of fiscal, political, and institutional decentral- ization that sought to reassign government functions and responsibilities across the national, departmental, and municipal levels of government. Under the decen- tralization framework, the central government concentrated on policy design, regulation, and public finance; departmental governments assumed responsibili- ties for regional planning, management and finance, articulation of policy at the local level, and provision of health care services at the regional level; and munici- pal governments took on policy implementation and provision of health care ser- vices at the primary level (Bossert and others 2000). From 1990 to 1993, legislative mandates defined territorial functions and responsibilities along with new sources of financing for health service provision and their respective allocation formulas. Administrative procedures to certify local governments as decentralized were established, and on the basis of these, authority, responsibility, and budgetary con- trol of resources were to be shifted to the departmental and municipal levels of government. Law 100 also established the legal basis for institutional decentralization of pub- lic hospital facilities through provisions mandating that hospitals were no longer to be administratively governed by local governments. The law facilitated conver- sion of these facilities to semipublic entities referred to as social enterprises of the state, thereby granting them the financial and managerial autonomy necessary to prepare for competition with the private sector under the new health insurance scheme. Design Important design features included administrative structure, target population, benefit design, and financing modalities. Administrative Structure The reform called for reorganization of the administrative structure of the health care system and the addition of new units (figure 6.3). The Ministry of Social Pro- tection took on the role of formulating national policies and regulations and moni- toring their progress and results. The monitoring, inspection, and sanctioning functions of health plans were entrusted to a regulatory body called the National Health Superintendency. A national health insurance fund, referred to as the Soli- darity and Guarantees Fund (SGF) was created to pool the system's revenues and distribute them among regimes (explained later). The National Council for Social Security in Health was created as a policy-making body with regulatory and pol- icy making authority over various aspects of the health care system as shown in Diana Pinto and William C. Hsiao 111 Figure 6.3. Administrative Structure of the Health Care System Ministry of Social Protection Formulation of national Solidarity and National Council of Social policies, operating rules, Guarantees Fund Security in Health and quality standards Flow of funds Benefits, premiums, tariffs, copayments, enrollment rules, subsidized regime consumers cement enfor regulation oversight, insurance plans providers monitoring, regimes local health National Council authorities for Social Security Source: Authors. figure 6.4. The council was chaired by the minister of social protection. Members included the minister of finance and representatives of various stakeholders, such as employers and workers, public and private health plans, and public and private sector providers. Target Population With the objective of achieving universal insurance coverage, but given resource con- straints, the reformers realized that they needed to design two separate insurance schemes that targeted different populations. The first scheme, in which members con- tributed to a national health fund according to their ability to pay, was known as the contributory regime. The contributory regime included all formal sector employees or independent workers able to pay who were already enrolled in some form of either private or public insurance and extended coverage to their families. The second scheme, referred to as the subsidized regime, targeted the poor by subsidizing their insurance premiums using dedicated public resources and resources from the contributory regime. As more resources became available, insurance coverage for the population eligible for subsidies was expanded by mak- 112 Colombia: Social Health Insurance with Managed Competition to Improve Health Care Delivery ing the benefits package more generous. Both regimes were expected to become equal in terms of benefits in 2000, leading to a universal health insurance system. Benefit Design The two regimes required the design of a benefits package within the resource constraints of each scheme. Contributory regime enrollees were entitled to a com- prehensive, standard benefits package; however, this package did not include such amenities as private or semiprivate rooms in hospitals and some services such as expensive imaging. People were allowed to purchase private insurance to supplement the basic package should they chose to do so. The benefits package for enrollees in the subsidized regime was limited in coverage compared with the contributory regime package. The subsidized regime package focused on primary care interventions, some basic surgeries and hospitalizations, and cata- strophic events. According to Law 100, adjustments to the benefits packages were to be based on changes in the population's epidemiological and demographic profile. In the case of the subsidized regime package, the addition of services was limited by the availability of resources. The premium for each package was fixed at the national level each year and was based on estimates of the costs of providing the respec- tive benefits package. If deemed necessary, the premium could also reflect risk adjustment. Financing The 1993 legislation introduced a number of new sources of funding for the health sector and a new entity, the SGF, to administer the flow of funds. CONTRIBUTORY REGIME. For those in the contributory regime, payroll taxes were increased from 8 to 12 percent of income. Of this 12 percent contribution by formally employed individuals, 4 percent was to be paid by the employee and 8 percent by the employer. Self-employed workers would be required to pay the full 12 percent on all earnings exceeding twice the level of the minimum wage. SUBSIDIZED REGIME. Of the 12 percent payroll contributions by those in the contributory regime, one percentage point (or roughly 1/8 of the total) was to be channeled to the subsidized regime. Tax revenues from several sources and social investment transfers to municipalities were to be earmarked for health. The new resources included income from oil revenues and funds from the central govern- ment equivalent to the value of the one percentage point channeled to the subsi- dized regime, as noted above. Resources used for supply-side subsidies were to be transformed to demand-side subsidies. These new sources were intended to finance the expansion of insurance cov- erage that would eventually embrace the entire populace (figure 6.4). The left- hand panel represents the uninsured population divided by ability to pay. The right-hand panel depicts the amount of resources available for insurance by the main sources of financing. Under the macroeconomic and employment growth assumptions for 1993, payroll tax contributions and the number of enrollees in the contributory regime would increase, and therefore so would funds for the Diana Pinto and William C. Hsiao 113 Figure 6.4. Toward Universal Health Coverage, 1993­2001 national and local tax revenues, solidarity matching funds transformation of supply-side subsidies unable to pay solidarity fund payroll contributions able to pay 1993 2001 1993 2001 uninsured population available funds for insurance Source: Authors. subsidized regime. In addition, increases in national and local tax revenues ear- marked for health, coupled with the gradual transformation of the resources allo- cated to finance the supply of health services into demand-side subsidies, would ensure a flow of funds to provide insurance subsidies to cover the poor. ADMINISTRATION AND FLOW OF FUNDS. The SGF was the institution that pooled the system's revenues and distributed these among regimes. The flow of funds took place as follows: · The SGF received the 12 percent contributions from employees in the formal sector from health plans. · The SGF transferred one percentage point of the total from employees in the formal sector to the solidarity fund, which subsidized the contributions of those in the informal sector. · Of the total revenues, the SGF · transferred 0.41 percentage point to a fund that financed health promo- tion and prevention activities by health plans, · set aside 0.25 percentage point for sick leave payments, · set aside 0.25 percentage point for maternity leave payments. The remaining revenues were used to pay premiums to contributory regime health plans and claims for payments filed by health plans for medications and procedures not included in the benefits package. Surpluses were invested in gov- ernment bonds and reserved for future contingencies. Implementation Key implementation issues included expansion of coverage, benefits packages, health plans, contracting and paying providers, and financial sustainability. Expansion of Coverage 114 Colombia: Social Health Insurance with Managed Competition to Improve Health Care Delivery Figure 6.5. Expansion of Insurance Coverage by Regime, 1992 and 1996­2003 40 subsidized 35 contributory population 30 total of 25 20 percentage 15 a as 10 5 enrollment 0 1992 1996 1997 1998 1999 2000 2001 2002 2003 year Source: Ministry of Social Protection data; national quality of life surveys. Colombia's health care reform was successful in expanding insurance coverage from 27 percent of the population in 1992 to more than 63 percent by 2003. Figure 6.5 shows the growth in enrollment for each regime. The growth in enrollment in the subsidized regime is notable. In particular, the subsidized regime played a key role in increasing coverage for the poorest population (figure 6.6), especially in rural areas, where insurance cov- Figure 6.6. Insurance Coverage by Urban and Rural Residence and Regime, 2003 100 80 ea total ar of in 60 40 centage per population 20 0 urban rural uninsured subsidized exceptions contributive Source: 2003 national quality of life survey. Diana Pinto and William C. Hsiao 115 Figure 6.7. Insurance Coverage by Income Quintiles, 1992­2003 90 80 70 insured 60 top 50 people fourth of 40 30 third 20 second percentage 10 bottom 0 1991 1993 1995 1997 1999 2001 2003 year Source: Ministry of Social Protection data. erage increased from 7 percent of the population in 1993 to 52 percent in 2003, with the most rapid increase occurring between 1993 and 1997. Insurance coverage reached the most vulnerable. Estimates of the proportion of insured individuals during 1992­2003 by income quintile show an increase of 37 percentage points for the bottom quintile (from 9 percent in 1992 to 48 percent in 2003), whereas the proportion insured in the top quintile increased by 21 percent- age points (from 60 percent in 1992 to 81 percent in 2003) (figure 6.7). Estimates from a national household survey in 2003 indicated that the total population eligible for subsidies could be around 22 million, as classified by the SISBEN targeting instrument. Of these, 12 million were regarded as relatively high priority for targeting (SISBEN levels 1 and 2), followed by the remaining 10 million classified as SISBEN level 3 candidates for targeting.. Assuming that no additional funds for the subsidized regime would become available, table 6.2 presents a num- ber of scenarios for expanding coverage as of December 2004. Under scenario 1, all those eligible for subsidies (SISBEN levels 1­3) are offered insurance with available demand-side resources collected from employer and employee premiums. The subsidized regime premium would have to be reduced and so would the benefits plan. The main drawback of this scenario would be that reducing benefits would be a politically undesirable measure. Under scenario 2, the SISBEN 1 and 2 populations are offered insurance at the current subsidized regime premium and benefits, using available demand-side resources, and the SISBEN 3 population would have to contribute some of the premium. This con- tribution would be equivalent to about 18 percent of the annual family income of this population. Not only would few families be willing or able to afford such a contribution, but it would also be difficult to collect. Scenarios 3 and 4 are the same as 1 and 2, but assume an aggressive realization of 60 percent of proposed supply-side subsidies. 116 Colombia: Social Health Insurance with Managed Competition to Improve Health Care Delivery Table 6.2. Scenarios for Expanding Coverage by the Subsidized Regime as of December 2004 Assumption Scenario 1 Scenario 2 Scenario 3 Scenario 4 · Demand-side Cover Cover only Transform Transform resources: everyone SISBEN 60% of 60% of Col$2,773 billion eligible for levels 1 and 2 supply-side supply-side · Supply-side resources: subsidy with populations resources resources, Col$1,800 billion available with and cover cover only · Total resources: resources available all those SISBEN Col$4,573 billion resources eligible for levels 1 · Population eligible and require and 2, and for subsidy contribution require SISBEN 1­2: 12,184,143 from SISBEN contribution SISBEN 3: 10,228,873 level 3 from SISBEN Total: 22,413,016 level 3 Premium (Col$) 123,723 193,000 159,908 193,000 Individual annual premium (Col$) 0 151,797 0 72,508 Individual contribution as a percentage of income 0 18 0 8 Supply-side subsidy per capita (Col$) 80,310 80,310 44,125 44,125 Total expenditure per capita (Col$) 204,033 204,033 204,033 204,033 Given the political infeasibility of reducing the subsidized regime benefits package or of accelerating the realization of subsidies, as an alternative, the govern- ment created a third insurance scheme that started in 2004, a partially subsidized regime. The rationale was to expand coverage of the urban SISBEN 2 and 3 popula- tions who were not enrolled in the subsidized regime because of the lack of funds to provide full subsidies. The program was financed through a mix of government subsidies and departmental and municipal resources.2 Health plans already oper- ating in the subsidized regime could participate in the program, providing a pack- age that offered less than the subsidized regime package, and receive a premium that was about 40 percent of the current subsidized regime premium. Benefits Packages Table 6.3 summarizes the main features of the contributory and subsidized regimes packages. As the table indicates, the contributory regime package covered almost all health interventions, whereas the subsidized regime package covered only essential clinical services, a few surgeries, and the treatment of catastrophic dis- eases. Thus the major difference between the two packages lay in the provision of services in the secondary and tertiary levels of care. The partial subsidy pack- 2. Local governments must guarantee resource availability to cofinance the program until 2008. Table 6.3. Comparison of the Contributory and Subsidized Regime Benefits Packages as of December 2004 Characteristic Contributory regime Subsidized regime Partially subsidized regime Scheme nature Compulsory Social welfare Social welfare Model Public/private managed competition model Public/private managed competition model Public/private managed competition model Population coverage Target population Individuals and families with the ability to Individuals and families eligible for subsidies Individuals and families eligible pay: formal workers, informal workers who (SISBEN 1 and 2) for partial subsidies (urban, earn more than twice the minimum wage uninsured SISBEN 2 and 3) Number of enrollees as 14,857,250 13,765,405 1,788,069 of December 2004 Percentage of the total 33 30 4 117 population covered Contents of benefits package Family coverage Yes Yes No Public health education Yes Yes No and outreach Preventive care, Full range of preventive services Same range of preventive services as included No individual and family under the contributory regime Outpatient services All All obstetric services All obstetric services (consultations, All pediatric services for children under one All pediatric services for treatment, diagnostic year of age children under one year of age tests, rehabilitation) Low complexity outpatient services (consultations, diagnostic tests), minor trauma, glasses for children and the aged, family planning Dental care Basic dental care Basic dental care No Table 6.3. Comparison of the Contributory and Subsidized Regime Benefits Packages as of December 2004 (continued) Characteristic Contributory regime Subsidized regime Partially subsidized regime (table continues on following page) Inpatient services All Hospitalization for low complexity care All obstetric services All obstetric services All pediatric services for All pediatric services for children children under one year of age under one year of age All orthopedic care General surgery Cataract surgery, correction of strabismus All orthopedic care Medications All medications in national listing All medications in national listing All medications in national listing required for treatment 118 of covered conditions and primary-level care Catastrophic care Treatment with radiotherapy and Same Same chemotherapy for cancer; dialysis and organ transplant for renal failure; surgical treatment of heart, cerebrovascular, neurological, and congenital conditions; treatment of major trauma; intensive care unit; hip and knee replacement; major burns, treatment for HIV/AIDS Transportation For referrals, catastrophic care For referrals, catastrophic care No Table 6.3. Comparison of the Contributory and Subsidized Regime Benefits Packages as of December 2004 (continued) Characteristic Contributory regime Subsidized regime Partially subsidized regime Excluded conditions Cosmeticsurgery,infertilitytreatment, All conditions not listed above All conditions not listed above treatmentforsleepdisorders,organtransplants (exceptrenal,heart,cornea,andbonemarrow), psychotherapyandpsychoanalysis,treatments forend-stagedisease (table continues on following page) Enrollment rules 119 Provisions against Minimum weeks of enrollment required Minimum period of two years of enrollment None adverse and risk before full coverage of higher complexity care before switching plans selection takes place; minimum period of two years of enrollment before switching plans No denial of coverage or pre-exclusion of No denial of coverage if eligibility criteria are No denial of coverage if existing conditions met or pre-exclusion of existing conditions eligibility criteria are met or pre-exclusion of existing conditions Risk adjustment of premium by age and sex No No Cost sharing Sliding scale copayments for ambulatory care No copayments No copayments and hospitalization; preventive services and public health initiatives are excluded (treatments for hypertension, diabetes care) 120 Colombia: Social Health Insurance with Managed Competition to Improve Health Care Delivery age was a subset of the subsidized regime package that covered catastrophic care, orthopedic services, maternal and child care, and medications. Health Plans Health plans serving the contributory regime were known as health promot- ing organizations. These plans could be private (for-profit or nonprofit) or public. Before the reform, health insurance for employees was provided almost exclusively by public insurers, of which the largest was the National Social Security Institute. Between 1993 and 2000, competition from private health plans reduced the public sector's market shares. Between 1995 and 1998, the market share of private health plans increased from 11 percent of the population in the contributory regime to 30 percent, a net increase of 261.1 percent. As of December 2004, 40 health plans par- ticipated in the contributory regime, of which private health plans had a 72 percent share of the market. In the contributory regime, health plans were free to decide on their network of providers. First levels of care were contracted mainly with private providers, and recently a trend toward vertical integration (ownership of first-level providers by insurance plans along the lines of health maintenance organizations in the United States) has become apparent. Health plans serving the subsidized regime were known as subsidized regime administrators (SRAs). These could be public, private (for-profit or non- profit), or community-based nonprofit organizations. The insurance market for the subsidized regime developed quickly: SRAs were authorized to operate in 1995, and as of December 1999, Colombia had 239 SRAs with a total of 9 million enrollees or 22 percent of the total population (Cardona 1999). The large number of SRAs did not permit adequate risk pooling and generated large inefficiencies, caused primarily by large transaction costs. This motivated the government to issue Decree 1804, which required SRAs to have a minimum of 200,000 enroll- ees and led to a wave of mergers that reduced the total number of SRAs to 43 by December 2004. Of this total, 45 percent were private, 42 percent were com- munity based, 6 percent were public, and the remainder were health plans for indigenous populations. Analyses in 2001 of the market structure at the municipal level indicated that even though most municipalities have several health plans, the market was charac- terized by a dominant health plan with more than 73 percent of enrollees. Only in major cities did the market appear to be competitive (Restrepo, Arango, and Casas 2001). Also, while efforts were made to unify the insurance schemes existing in 1993, exceptions were made for the military and police forces, the education sec- tor, and those working for the Colombian oil company, allowing these groups to be autonomous in the organization and provision of health benefits. About 4 percent of Colombia's population was affiliated with these independent schemes. The Ministry of Social Protection defined minimum quality and financial and administrative standards for health plan operations and health plans are required to demonstrate fulfillment of these requirements through a certification process. However, implementation of this process has been slow, in part because of political pressure by public health plans that might not meet the standards. Contracting and Paying Providers Diana Pinto and William C. Hsiao 121 Table 6.4. Payment Method Used by Health Plans in the Contributory Regime (proportion of sampled firms that use payment method) Capitation and Level of care Capitation only Fee-for-service only fee-for-service Total Preventive care 55 27 18 100 Primary care 46 9 45 100 Secondary care 27 37 36 100 Tertiary care 9 46 45 100 High complexity specialized care 0 73 27 100 For more complex levels of hospital care, health plans contract with both public and private institutions. Health plans are increasingly using capitation payments for preventive services and primary­level care, but most specialist care and hospital care is paid for on a fee-for-service basis as shown in table 6.4. Although SRAs are required by law to contract with public providers for at least 40 percent of their network, they are free to structure the rest of their network as they wish. Otherwise, contracting and health care management strategies are essentially the same as those described for health promoting organizations in the contributory regime. In the contributory regime, health plans can control demand by charging copay- ments for both ambulatory care and hospital services. The central government sets the maximum value of copayments per service on a sliding scale by income level. Copayments are not permitted for preventive services and services for certain conditions of public health interest, such as hypertension and diabetes. The main administrative strategies health plans use to control demand are the use of gate- keepers and utilization management for specialists, hospitals, and diagnostic care. Financial Situation At the time of the reform, the government assumed that the sustainability of the contributoryregimewouldbeguaranteedthroughwagecontributionsandthrough additional resources generated by enrollee copayments and the sale of supplemen- tal insurance. The assumption also implied that premiums would be sufficient to cover health plans' costs. Until 1999, the compensation fund maintained a positive balance, which yielded important financial returns, but by 2000, premium pay- ments to health plans increased at a faster rate than revenues from wage contribu- tions, and by 2001 the compensation fund had a negative balance, leading to the need to use surpluses from previous years. Table 6.5 shows the SGF's balances from 1998 through 2002. The total number of enrollees in the contributory regime has grown mainly because of an expansion in family coverage, as seen by the faster growth rate of beneficiaries than of contributors. In addition, average salaries expressed in multiples of the minimum wage have been decreasing and family size has been increasing. The composition of enrollees and the labor market conditions have together translated into fewer revenues and greater expenditures. The experience of the subsidized regime was similar to that of the contributory regime. Between 1995 and 2002, resources for the subsidized regime increased from Table 6.5. Solidarity and Guarantees Fund (SGF) Balances, 1998­2002 Change from Category 1998 1999 2000 2001 2002 1998 to 2002 (%) Characteristics of enrollees Contributors (number of people) 5,089,574 5,293,294 5,235,106 5,204,702 5,282,773 0.9 Beneficiaries (number of people) 6,481,049 7,734,498 7,692,801 7,732,990 7,976,987 5.3 Total number of enrollees 11,570,623 13,027,792 12,927,907 12,937,692 13,259,760 3.5 Family density 2.27 2.46 2.47 2.49 2.51 Average salary (multiple of the minimum wage) 2.18 2.09 2.07 2.1 2.0 12 Revenues (Col$ millions) 2 Contributions 3,256,567 3,766,973 4,058,810 4,501,359 4,701,245 1.4 Financial returns 57,998 74,641 38,099 37,923 40,198 Total 3,314,565 3,841,614 4,096,909 4,539,282 4,741,443 1.1 Expenditures Premium (Col$) 207,362 241,577 265,734 289,119 300,684 1.5 Total premium payments (Col$ millions) 2,399,308 3,147,215 3,435,384 3,740,533 3,986,998 5.0 Sick and maternity leave payments (Col$ millions) 135,690 156,957 169,117 187,557 176,297 ­1.3 Transfers to solidarity fund (Col$ millions) 310,108 384,052 386,317 382,092 368,346 ­3.5 Transfers to prevention fund (Col$ millions) 135,690 156,957 169,117 187,557 62,683 Total (Col$ millions) 2,980,795 3,845,181 4,159,935 4,497,739 4,594,324 3.0 Balance (Col$ millions) 333,770 (3,567) (63,026) 41,543 147,119 Table 6.6. Total Resources and Sources of Funding for the Subsidized Regime, 1995­2002 Category 1995 1996 1997 1998 1999 2000 2001 2002 Total resources (Col$ billions) 0.63 0.71 0.80 1.12 1.25 1.24 1.71 1.84 Total resources (percentage of GDP) 0.75 0.70 0.66 0.79 0.84 0.73 0.91 0.91 12 Sources of funds (percentage of total resources) 3 Solidarity fund 85.7 63.1 46.7 51.3 38.0 34.6 36.0 31.0 National budget 14.3 36.9 44.0 37.2 50.0 51.9 51.0 66.2 Own resources (municipal taxes) -- -- 3.8 4.5 3.9 5.1 3.0 0.8 Earmarked municipal taxes -- -- 1.8 2.2 3.4 2.1 9.0 -- Transfers from employment funds -- -- 3.7 4.8 4.8 6.2 2.0 2.1 124 Colombia: Social Health Insurance with Managed Competition to Improve Health Care Delivery Col$0.63 billion to Col$1.84 billion, or 0.75 to 0.91 percent of GDP (table 6.6). Initially, the main source of financing was the solidarity fund contribution, but this has gradually been replaced by resources from the national budget, which accounted for more than 66 percent of funding in 2002. Other sources of funding grew moder- ately. The amount of funds available for the subsidized regime is insufficient to pro- vide full coverage for the eligible uninsured given the current subsidized regime premium. Achievements As figure 6.8 illustrates, evidence suggests a decrease in financial barriers to health care. According to a 1992 household survey, the main reason interviewees in the lowest quintiles gave for not using health care was the cost of services. Responses to this same question in a 1997 household survey show a reduction in the percent- age of people reporting cost as a reason for not using health care services for all income levels. Being insured seems to be positively related to having access to a doctor (figure 6.9). Between 2000 and 2003 in both urban and rural areas, not only were treatment rates higher for the insured compared with the noninsured, but they also increased more. The insurance schemes also seem to provide financial protection, especially for the poor. For example, a risk-pooling study compared the percentage of the popula- tion that fell below the poverty line because of a health incident requiring ambu- latory care or a hospitalization by insurance status (Bitran, Giedion, and Muñoz Figure 6.8. Reasons for Not Using Health Care by Income Quintile, 1992 and 1997 1992 1997 100 100 90 90 80 80 70 70 60 60 respondents respondents of 50 of 50 40 40 30 30 percentage 20 percentage 20 10 10 0 0 top top third bottom second fourth third bottom second fourth income quintile income quintile cost denied care no provider other Sources: Ministry of Social Protection calculations based on the 1992 national household survey and the 1997 national quality of life survey. Diana Pinto and William C. Hsiao 125 Figure 6.9. Percentage of Respondents Who Had a Health Problem Who Were Seen by a Doctor by Enrollment Status and Area, 1997 and 2003 1997 2003 80 80 70 70 60 60 50 50 espondentsr espondentsr of 40 of 40 30 30 centage centage 20 20 per per 10 10 0 0 urban rural urban rural insured uninsured Sources: Ministry of Social Protection calculations based on the 1997 and 2003 national quality of life surveys. 2004). The study finds that 5 percent of the noninsured fall below the poverty line because of ambulatory care and 14 percent do so as a result of hospitalization, com- pared with 4 percent of the subsidized regime population that fall below the pov- erty line as a result of either type of event. Challenges While social health insurance has made impressive gains in Colombia, important challenges remain. Expanding Coverage Several factors have negatively influenced the growth and availability of resources for health insurance expansion, of which the most important are economic reces- sion and labor market characteristics. The availability of resources for health insurance declined because of a reduc- tion in wage contributions and fiscal resources. This was largely an outcome of unemployment and of decreases in family incomes, which have fallen by about 25 percent since 1997. In addition, fiscal pressures coupled with a lack of political commitment led the central government to fail in its obligation of providing soli- 3. International Labour Organisation. "Key Indicators of the Labour Market 2004." International Labour Organisation. http://www.esds.ac.uk/international/support/user_ guides/ilo/kilm.asp. Last accessed July 2005. 126 Colombia: Social Health Insurance with Managed Competition to Improve Health Care Delivery darity matching funds for the subsidized regime and directing oil revenues to the health sector. Colombia's labor market characteristics pose challenges to insurance expan- sion and to increases in revenues from wage contributions. In 2000, 55.6 percent of the labor force was in the informal sector.3 Bitran and Associates, Econometrita, and Superior School of Public Health (2002) found that about 31 percent of infor- mal workers fell into SISBEN categories 1­2, and thus were eligible for the subsi- dized regime. Of the remaining informal workers, about 84 percent (10 million people or about 25 percent of the population) could not afford to pay the full con- tributory regime premium, and half of them were uninsured. Evidence indicates that this segment of the population has been growing. The main difficulties in expanding insurance coverage in this segment of the population in the contribu- tory regime are · the limited, unpredictable, and unreliable ability to pay; · the high potential for adverse selection; · the high administrative costs involved in enrolling, monitoring, and collect- ing contributions; · the lack of incentives to enroll, as people can still use services in the public sector, often at minimal cost because of institutions' lack of capacity to col- lect user fees. Preventing Evasion Empirical evidence has shown that evasion of wage contributions and underreport- ing of income are factors that have affected the SGF's balances (Bitran and Associ- ates, Econometrita, and Superior School of Public Health 2002). Estimates for 2000 indicated that only 65 percent of potential contributors were actually paying their obligations, and furthermore, those who did contribute paid much less than they should. This could be decreasing the contributory regimes' revenues by 35 percent. In the absence of this problem, positive balances would have been achieved, pro- viding sufficient resources to enroll an additional 1.5 million people in the subsi- dized regime. Among the factors related to evasion are characteristics of Colombia's labor market, incentives for health plans to underreport income, and lack of an infor- mation system that allows evaders to be detected. As an initial strategy to reduce evasion, pension system data and health insurance contribution data have been cross-referenced. Fostering Competition Managed competition is based on the premise that given mechanisms to guarantee equitable access to health care services and regulations to address failures inherent to health care markets, competition can be used to improve efficiency and quality. The design of the Colombian health care system should lead to a focus in quality; however, after years of reform, the necessary conditions for competition in relation to quality have not emerged. A variety of barriers may be preventing consumer choice from being the driv- ing force behind competition. To generate more sensitive demand for quality, find- Diana Pinto and William C. Hsiao 127 ing effective ways to provide consumers with information about how the system functions, their rights and obligations, and aspects of quality of care relevant to consumer choice is a priority. To this end, the government is currently working on designing appropriate consumer information systems. An alternative is to iden- tify sponsors of managed competition, as originally envisioned by the Enthoven model. In this model, sponsors would pool a large number of enrollees in their organizations, gain greater purchasing clout in the marketplace as a result, and require providers to compete for contracts to serve their members. This greater purchasing clout would pressure providers to compete not only on price, but also on quality of services provided. On an informal basis, employers in the contribu- tory regime and local governments in the subsidized regime seem to be exercising this role, therefore the government may want to consider improving the capacity of these institutions to direct enrollees toward high-quality health plans by pro- viding information about quality. Surveys and focus groups involving consumers, employers, and local health authorities that seek to identify who is making health care choices and how these choices are affected by quality of services could pro- vide useful information to the quasi-managed competition approach now in play. Other problems in reform implementation that presented obstacles to quality competition were the absence of a quality assurance system and weak institutional capacity to conduct quality oversight. In 2003, regulations pertaining to qual- ity were issued, such as minimum standards and accreditation procedures, and whether capacity to monitor and enforce these provisions will be sufficient remains to be seen. At the same time, in some areas of the country the size of risk pools makes it infeasible for more than one health plan to operate or to have a choice of provid- ers. Competitive bidding and performance-based contracts might be alternatives to induce quality improvement under these circumstances. Figure 6.10. Public Hospitals' Incomes and Expenditures, 1993­99 4,000 3,500 income 3,000 expenditure Col$ 2,500 1999 2,000 of 1,500 billions 1,000 500 0 1993 1994 1995 1996 1997 1998 1999 year Source: Ministry of Social Protection calculations based on public hospital budget data 1993­99. 128 Colombia: Social Health Insurance with Managed Competition to Improve Health Care Delivery Transforming Supply-Side Subsidies into Demand-Side Subsidies The implementation of rules governing the transformation of supply-side subsidies into demand-side subsidies was slow. The transformation of subsidies only began in 1997, and by 2000, only 50 percent of what had initially been estimated had been achieved. Most public hospitals are still financed by national resources allocated through historic budgets and spending by public hospitals has increased more than their incomes (figure 6.10) An evaluation of the public network found that between 1994 and 2000, the total income accruing to public hospitals increased more than 100 percent (Gen- eral Comptroller of the Republic 2002). This was more pronounced for first-level hospitals, even though their occupancy rate had fallen. Personnel costs increased 40 percent in real terms since 1995 and total hospital expenditures increased almost 1 percentage point of GDP between 1996 and 2000. During 2003, the main source of revenues for first-level hospitals was the provi- sion of services for the subsidized regime, while the main source of revenues for second- and third-level hospitals was regular government budgeting (table 6.7). For all levels of care, the sale of services for the contributory regime represents a small share of total revenues from the sale of services. First-level hospitals generate 40 percent of their total sales revenues, second-level hospitals generate 36 percent, and third-level hospitals generate 24 percent. Among the most important obstacles to hospital transformation have been the application of bailout policies; the political pressure from interest groups, in par- ticular, strong unions; the scant development of managerial capacity; and the legal restrictions against autonomous management of resources, especially personnel. Also the persistence of a large number of uninsured and the design of the subsi- dized benefits package, which envisages the public network satisfying the demand for a large portion of second- and third-level care, has been used as a justification to continue financing via supply-side subsidies (Giedion, Lopez, and Marulanda 2000; Giedion, Morales, and Acosta 2000; Sáenz 2001). Dealing with Lack of Clarity and Flaws in Rationing Policies Table 6.7. Sources of Public Hospitals' Sales Revenues by Level of Care, 2003 (percentage of total sales revenues) Source Level I Level II Level III Subsidized regime 42 20 20 Contributory regime 5 7 8 Uninsured care 30 47 44 Compulsory insurance for traffic accidents and catastrophic risk and accident fund 0 2 6 Public health package 4 1 0 Other 19 23 22 Total 100 100 100 Percentage share of total sales revenue generated by hospitals 40 36 24 Diana Pinto and William C. Hsiao 129 If Colombian citizens demand a health service or medication that is not covered by the benefits package, they have the option of filing a lawsuit against the health plan or the public provider, basing the claim on the constitutional right to life. If the court upholds the claim, the SGF has to pay for the service or medication. In addi- tion, as concerns medications, the SGF has a mechanism that reimburses health plans a fraction of the cost of drugs not covered in the package if they are able to demonstrate medical need. Use of these options has been increasing in recent years, and the costs to the sys- tem are not negligible. If the total value of these claims were converted to premium equivalents, it would have been possible to pay for insurance for approximately 199,000 contributory regime enrollees in 2002 and 327,000 thousand in 2003. Simi- larly, it would have been possible to pay for roughly 27,000 more subsidized regime enrollees in 2002 and 43,000 in 2003 (Pinto and Castellanos 2004). The costliest claims have been related to a brand medication for Gaucher´s dis- ease, a rare metabolic, lifelong disorder. In one year, the amount the SGF paid for treating three people with this disease could be equivalent to enrolling about 1,000 people in the contributory regime and twice as many in the subsidized regime. Solutions to this problem will require developing and promoting policies around evidence-based medicine and technology assessment and launching an explicit debate around rationing that incorporates both stakeholders of the health care sys- tem and the courts. Addressing Administrative Inefficiency Because of the system's complexity, the administrative costs of health plans can be large and highly variable, ranging from 4 to 60 percent of the value of the premium (Cendex 2000; national health accounts data provided by the National Department of Planning). More than 50 percent of administrative costs are spent on supporting dailyoperations(financial,personnel,andinformationmanagement)andenrollment processes. Little is spent on risk management and quality assurance (Cendex 2000). Delays in the flow of funds from the government to health plans and from health plans to providers are a serious problem. Calculations have showed that 230 days can pass before resources from the SGF reach a provider. All actors contribute to the problem: providers are slow in charging health plans, health plans spend much time auditing claims, and local governments delay payment to health plans to obtain interest on insurance funds (Jaramillo, Poveda, and Bernal 2002). This constrains providers' capacity to invest in technology and infrastructure and, in general, increases the financial risk of operating their services. Measures have been taken to simplify administrative processes in relation to the flow of funds. Closing Information Gaps Colombia's health system has not developed mechanisms for adjusting and mon- itoring premiums and the contents of benefits packages. The parameter that has guided decisions about premiums has been annual increments in the minimum wage, along with considerations about the SGF's balance at the end of the fiscal year. No actuarial studies of the real cost of providing the benefits packages have 130 Colombia: Social Health Insurance with Managed Competition to Improve Health Care Delivery Table 6.8. Ratio of Total Health Expenditures to Total Income from Premiums 2003 and 2004 Regime 2003 2004 Contributory 78­83 80­86 Subsidized 87­95 74­94 Source: Ministry of Social Protection data. been undertaken, mainly because of the lack of representative and reliable infor- mation needed for these calculations. As table 6.8 indicates, 2003 and 2004 estimates of the ratio of total health expen- ditures to total income from premiums suggest that the value of the premium is enough to cover health plans' expenditures and could leave a margin for admin- istrative expenses and utilities. However, the reliability and validity of these esti- mates is questionable, as they are based on financial reports from accounting data that are based on standardized reporting systems and are difficult to verify. If the premiums are actually set above costs, the system would be wasting resources. If the premiums are set below actual costs, consumers will suffer, as health plans will have incentives to cut quality and access. Regarding the contributory regime package and the subsidized regime package, the Ministry of Social Protection does not have systematic mechanisms to revise and adjust their contents. Few modifications have been made to the packages, and decisions to include medical interventions have been a response to requests by interest groups. Epidemiological information has not been updated since 1995. Thus little is known about the appropriateness of the benefits packages in address- ing the country's changing health priorities. Dealing with Risk Selection Despite the existing regulations to prevent adverse and risk selection,4 evidence indicates that the national health insurance system does suffer from selection problems. For example, a study of the distribution of cases related to catastrophic care in the contributory regime showed important deviations from the national average for conditions such as chronic renal disease and HIV/AIDS for several health plans (Ministry of Social Protection data). While the national average num- ber of cases in the contributory regime for chronic renal disease and HIV/AIDS was 41 and 19 per 100,000, respectively, the average for one of the health plans was 137 and 44. The situation motivated the government to adopt two measures that were expected to be one-time solutions to the problem. In one measure, the outlier health plans were reimbursed retrospectively for the value above the expected costs for these two conditions for one year by transferring this amount from the 4. Compulsory enrollment, standard benefits, age and sex risk-adjusted payment, non- denial of enrollment, and prohibition of exclusions and minimum enrollment periods. Diana Pinto and William C. Hsiao 131 premiums of health plans that had less than the national average of such cases. In a second measure, a redistribution of cases was mandated by legally requiring a randomly selected number of patients in excess of the national average to switch to other health plans. Enforcement of this second measure was extremely cumber- some: many of the selected patients could not be found because of information inconsistencies in the databases that were collected for this purpose, a large num- ber of individuals refused to switch, and lawsuits against the measure were filed. In consequence, the measure was dropped in 2004. As an alternative, the Minis- try of Social Protection is studying the alternative of including risk adjustment for these conditions in the premium. However, the challenge for the success of this strategy will be developing reliable diagnostic and cost information. Assuring Equity and Convergence of the Two Regimes The fact that both regimes have not converged and will not do so in the near future raises serious equity and efficiency concerns for Colombia's national health insur- ance. Because of the gaps in service coverage, the subsidized regime population is more likely to experience discontinuities in care that have the potential of reduc- ing the effectiveness of treatments.5 Also, as this population would be accessing services both from health plans (covered services) as well as from local govern- ments and public hospitals, monitoring of access, quality, and efficiency becomes difficult. The introduction of partial subsidies further increases inequities generated by the differences in coverage between the contributory and subsidized regimes. The policy ignores the potential for the subsidized segment of the population to pay even a small percentage of the premium and perpetuates the need for supply-side subsidies. It also concentrates on curative care rather than on preventive care. At this point, no clear answers are available in relation to alternatives for equal- izing benefits for the contributory regime and the subsidized regime. A first step is to evaluate the contents of these packages in the light of updated information on the burden of disease, the utilization of services, the cost-effectiveness of interven- tions, the available resources, and the preferences of patients. This could initiate a process leading to the redesign of both packages starting from a common basis. References Acosta, O. L., M. Ramírez, and C. I. Cañon. 2004. "Principales estudios sobre sostenibilidad financiera del SGSSS." Working paper, Corona Foundation, Bogota. Baron, G. 2005. "Cuentas de salud de Colombia: Aspectos conceptuales y análisis de resul- tados 1993­2002." Working paper, Health System Reform Support Program (Programa de Apoyo a la Reforma de Salud), Ministry of Social Protection and Social Development Division of the National Department of Planning, Bogota. 5. For example, a woman with a positive PAP smear for cervical cancer, which is cov- ered by the subsidized regime, has to pay out-of-pocket to have the diagnosis confirmed by a colposcopy, which is not covered. 132 Colombia: Social Health Insurance with Managed Competition to Improve Health Care Delivery Bitran and Associates, Econometrita, and Superior School of Public Health (Bitrán y Asociados-Econometría and Escuela Superior de Administracion Publication). 2002. Propuestas de reestructuración de los procesos, estrategias y organismos públicos y privados encargados de la afiliación, pago y recaudo de aportes al SGSSS. Health System Reform Sup- port Program of the Ministry of Social Protection (Programa de Apoyo a la Reforma. Proyecto 03-99). Bogota: Ministry of Health. Bitran, R., U. Giedion, and R. Muñoz. 2004. "Risk pooling, ahorro y prevención: Estudio regional de Políticas para la protección de los más pobres de los efectos de los shocks de salud." Washington, DC: World Bank. Bossert, T., M. Chawla, D. Bowser, J. Beauvais, U. Giedion, J. J. Arbelaez, and A. L. Villan. 2000. Applied Research on Decentralization of Health Systems in Latin America: Colombia Case Study. Data for Decision Making Project. Cambridge, MA: Harvard School of Public Health. Cardona, A. 1999. "Participación de las entidades promotoras de salud (EPS) en el mercado de aseguramiento de salud." Revista Facultad Nacional de Salud Pública 17 (1): 52­62. Cendex. 2000. Estudio de costos administrativos. Program to Strengthen the Subsidized Regime (Programa para el fortalecimiento del Régimen Subsidiado). Bogota: Interamerican Development Bank, Corona Foundation, and Social Foundation. Center for Economic Research (Centro de Investigaciones Económicas). 2001. Encuesta a EPS y ARS sobre configuración y participación en el mercado. Medellin, Colombia: Antioquia Uni- versity, Faculty of Economic Sciences. CIA (Central Intelligence Agency). 2006. The World Factbook. McLean, VA: CIA. http://www .cia.gov/cia/publications/factbook/. Escobar, M. L. "Health Sector Reform in Colombia." World Bank Institute. Electronic maga- zine. http://www1.worldbank.org/devoutreach/may05/document.asp?id=308. Date consulted: August 2005. General Comptroller of the Republic (Contraloría General de la República). 2002. Situación de la Red Hospitalaria Nacional. Fiscal Management Financial Report 1. Bogota: General Comptroller of the Republic. Giedion, U., H. R. Lopez, and J. A. Marulanda. 2000. "Desarrollo institucional del sector salud en Bogotá." Coyuntura Social 23: 57­81. Giedion, U., L. Morales, and O. L. Acosta. 2000. "Efectos de la reforma en salud sobre las conductas irregulares en los hospitales públicos." Coyuntura Social 23: 98­126. Jaramillo, L, R. Poveda, and C. Bernal. 2002. "Porqué no fluyen los recursos del régimen subsidiado?" In Ha mejorado el acceso en salud? Evaluación de los procesos del régimen subsidi- ado, ed. D. Arevalo and F. Martinez, 185­268. Bogota: National University of Colombia, Center for Development Research. Lasso, F. 2004. "Incidencia del gasto público social sobre la distribución del ingreso y la reducción de la pobreza." Working paper, Mission for the Design of a Strategy for Pov- erty and Inequality Reduction (Misión para el diseño de una estrategia para la reducción de la pobreza y la desigualdad), National Planning Department, Bogota. Pinto, D., and M. Castellanos. 2004. "En el sistema general de seguridad social colombiano caracterización de los recobros por tutelas y medicamentos no incluidos en los planes obligatorios de salud." Revista Gerencia y Políticas de Salud 3 (7): 40­61. Restrepo, J. H., M. Arango, and L. Casas. 2001. "Estructura y conducta de la oferta del seguro de salud en Colombia." Observatorio de la Seguridad Social 1: 1­10. Sáenz, L. 2001. "Modernización de la Gestión Hospitalaria Colombiana: Lecciones aprendi- das de la transformación de los hospitales en Empresas Sociales del Estado." Document 46. Regional Initiative of Health Sector Reform in Latin America and the Caribbean. Transparency International. 2006. Corruption Perception Index. London: Pluto Press. 7 Thailand: Achieving Universal Coverage with Social Health Insurance Piya Hanvoravongchai and William C. Hsiao Thailand is a middle-income developing country with an interesting history of health financing development. Since its first unsuccessful effort to introduce an SHI law in 1950s, Thailand has learned from its experiments with various types of financing mechanisms. It recently became one of the few developing countries that have achieved universal health insurance coverage. The mobilization of political and technical support led to this achievement, although Thailand's health financ- ing reform remains unfinished. Background Thailand is a lower-middle-income country in tropical Southeast Asia that covers an area of about 514,000 square kilometers (about the size of France) (figure 7.1). Thailand is divided into 76 provinces that are further subdivided into roughly 800 districts. One-third of the population lives in urban areas. Almost 95 percent of Thai are Buddhist and somewhat less than 5 percent are Muslim. Table 7.1 provides basic statistics regarding Thailand's demography, economy, health status, and health sys- tem. It has a population of 63 million people and a relatively young demographic structure: 25 percent of the population are younger than 15 and only 6 percent are 65 years old or older. Thailand is one of the few countries in Pacific Asia that have never been colo- nized. Its political system is a constitutional monarchy, with the king as head of state and the prime minister, elected by parliament, as head of the administration. A political reform initiated in 1995 after the fall of the military government led to a new constitution that was promulgated in 1997. This constitution is frequently referred to as the people's constitution, because it set new standards for judicial, executive, and legislative systems and strengthened political party systems and civil society. In 2001, the first election under this new constitution resulted in the Thai Rak Thai Party gaining a majority in parliament. Its leader, Thaksin Shinawa- tra, became prime minister, and implementation of the universal health care cover- age program, one of the key features of the party's platform, began. 133 134 Thailand: Achieving Universal Coverage with Social Health Insurance Figure 7.1. Map of Thailand Source: World Bank. Economy Over the past 30 years, the economy has developed from primarily agrarian to largely industrial. Thailand was one of the world's fastest-growing economies during 1985­95, with real average annual growth of 8.4 percent. Until the Southeast Asian economic crisis in 1997, the steady growth had allowed the country to achieve Piya Hanvoravongchai and William C. Hsiao 135 Table 7.1. Background Statistics, Thailand Category Statistic Source Demographics (2005) Population (millions) 64.2 World Bank 2006 Growth rate (%) 0.8 World Bank 2006 Urban population (%) 32.3 World Bank 2006 Economy GDP per capita, purchasing power parity (2005) 8,551 World Bank 2006 GDP annual growth rate, per capita, real purchasing power parity (past decade) 1.7 World Bank 2006 Percentage of the population below the national poverty line (2003) 11.9 CIA 2006 Percentage of the population living on less than $2/day (2002) 25.9 World Bank 2006 Percentage of the population living on less than $1/day (2002) 2.0 World Bank 2006 Agricultural employment (percentage of the population) 49.0 World Bank 2006 Corruption (rank out of 163 countries, 2005)a 63 Transparency International 2006 Health status Life expectancy (years) (2004) 70.5 World Bank 2006 Infant mortality/1,000 live births (2004) 18.2 World Bank 2006 HIV/AIDS prevalence (percentage of the population age 15­49) (2005) 1.4 World Bank 2006 Health system Health spending (percentage of GDP) (2003) 3.3 World Bank 2006 Health spending (US$ per capita, purchasing power parity) (2003) 282 World Bank 2006 Physicians/1,000 population (2000) 0.37 World Bank 2006 Hospital beds/1,000 population (1999) 2.2 World Bank 2006 Health system financing (percentage of total health expenditure) (2003) SHI 19.7 WHO 2006 General government revenue 41.9 WHO 2006 Out-of-pocket 28.7 WHO 2006 Private insurance 5.6 WHO 2006 Other private 4.1 WHO 2006 consistent upward growth in its per capita income and a significant drop in the num- ber of poor from 45 percent of the population in 1988 to 17 percent in 1996 (figure 7.2). The 1997 economic crisis resulted in an abrupt economic slowdown, with the currency devaluing by more than 50 percent and a GDP growth rate of negative 10 percent. This was accompanied by a fall in the demand for labor, a reduction in wages, an increase in the prices of key commodities, a reduction in government spending on social services, the erosion of social capital, and a rise in the number of poor. However, the economy started to get back on track in 2001, and by 2003 and 2004, the growth rate had reached more than 6 percent per year. 136 Thailand: Achieving Universal Coverage with Social Health Insurance Figure 7.2. Poverty Trend, 1988­2004 45 poor 40 35 considered 30 25 20 population the 15 of 10 5 percentage 0 1988 1990 1992 1994 1996 1998 2000 2002 2004 year Source: National Economic and Social Development Board data (http://www.nesdb.go.th). Health The population's health status has improved considerably during the past decade and a half as evidenced by a significant reduction in the infant mortality rate from 31 per thousand live births in 1990 to 18.2 in 2004, and an increase in life expec- tancy from 68 to 70.5 years over the same period (World Bank 2006). Rates of child malnutrition and anemia in pregnant women have also declined steadily. Despite successful HIV/AIDS prevention programs and declining prevalence (currently less than 2 percent), HIV/AIDS is still the leading cause of death, especially among the working-age population. This is followed by traffic deaths. Other leading causes of death include heart disease, cancer, and lower respiratory tract infections. In addition, Thailand is experiencing an epidemiological transition, with health problems from degenerative and behavior-related diseases on the rise. The increase in life expectancy and the drop in the fertility rate are resulting in increasing numbers of elderly, which are expected to account for 10 percent of the population by 2020, and a further increase in the disease burden from noncommunicable diseases. The prevalence of chronic illnesses, such as diabetes and hypertension, is also expected to increase more rapidly than in the past. During 1987­97, the cancer prevalence rate rose from 53.8 to 60.4 per 100,000 population. Health Financing and Expenditure Thailand's health expenditure, 3.3 percent of GDP, is low considering its level of economic development. Health expenditure increased steadily from 1980 until peaking in 1996 at 4 percent of GDP. Figure 7.3 shows how health spending shrank during the recession and remained at 3.3 percent in 2003. The amount privately funded, mostly out-of-pocket payments, continued to decline throughout this period as more people enrolled in various insurance Piya Hanvoravongchai and William C. Hsiao 137 Figure 7.3. Health Expenditure as a Percentage of GDP, 1994­2004 4.0 3.8 3.6 percent 3.4 3.2 3.0 1992 1994 1996 1998 2000 2002 2004 year Source: National health accounts. schemes. More than 60 percent of health expenditure is now publicly financed. Fig- ure 7.4 shows the changes in financing and figure 7.5 shows what the Thai buy with their health funds. Public and Private Provision of Health Services The health service system involves both public and private institutions. Table 7.2 shows the share of public and private health facilities and utilization rates in 2003. The Ministry of Public Health is the major health service provider with an exten- sive network of hospitals and health centers. Of the total 136,201 hospital beds in 2003, 64 percent were Ministry of Public Health facilities and an additional 14 per- cent were also public sector facilities. Every province has at least one Ministry of Public Health general hospital and every district has one ministry district hospital. General and district hospitals pro- vide outpatient and inpatient medical services, including diagnostic and surgical services and medical treatment. More than 10,000 community health centers are available at the subdistrict level that are equipped with trained personnel (two years of training after high school) who provide primary health care, including health promotion and prevention programs. In addition, each village has one or two village health volunteers who participate in health promotion and prevention activities. The private sector participates in health care delivery, but private facilities are concentrated in Bangkok and other large cities. The economic boom in the late 1980s and early 1990s promoted rapid private investment in and growth of the health sector, but this has slowed significantly since the 1997 economic crisis, as shrinking consumer demand and an oversupply of hospital beds forced 80 private hospitals out of business, while the remaining private hospitals had to reduce their bed capacity and scope of services. 138 Thailand: Achieving Universal Coverage with Social Health Insurance Figure 7.4. Share of National Spending on Health by Source, 1994­2001 100 90 80 B 70,770 B 78,430 B 93,581 B 87,206 B 78,155 B 73,183 B 73,373 B 74,424 70 60 cent B 8,850 B 11,274 50 B 6,459 B 8,601 B 7,907 per B 4,940 B 5,985 B 18,028 B 3,702 B 18,874 B 18,358 B 19,519 B 21,708 40 B 12,949 B 15,615 B 12,432 30 20 B 40,751 B 51,518 B 61,922 B 77,450 B 67,181 B 62,676 B 65,405 B 62,797 10 0 1994 1995 1996 1997 1998 1999 2000 2001 year private social security scheme civil servant medical benefits scheme government Source: National health accounts. Figure 7.5. Share of National Spending on Health by Spending Category, 2001 health administration and health insurance 8% prevention and public health 8% capital 5% inpatient services 33% medical goods 6% outpatient services 40% Source: National health accounts. Piya Hanvoravongchai and William C. Hsiao 139 Table 7.2. Numbers and Use of Public and Private Health Facilities, 2003 Ministry of Public Other Health facilities public facilities Private facilities Utilization Utilization Utilization Total Category Number (%) Number (%) Number (%) (number) Hospitals 868 67 94 7 331 26 1,293 Beds 87,752 64 19,088 14 29,361 22 136,201 Outpatient cases 22,872,801 78 2,442,229 8 4,144,852 14 29,459,882 Inpatient cases 5,804,571 73 634,574 8 1,487,351 19 7,926,496 Inpatient days 26,473,752 75 4,671,715 13 4,338,363 12 35,483,830 Bed occupancy rate (%) 82.7 67.1 40.5 71.4 Source: Wibulpolprasert 2005; Ministry of Public Health data (http://www.moph.go.th/). In 2000, Thailand had a total of 18,025 doctors, with one doctor for every 3,400 people. About two-thirds of the doctors work in the public sector, more than half of them for the Ministry of Public Health. Public sector physicians are paid primarily on a salary basis and receive additional pay when they are on call outside normal working hours. A significant proportion of public sector physicians also works in the private sector outside normal office hours for extra income. Physicians in pri- vate hospitals receive consultation fees that are charged directly to patients. Doc- tors in private clinics earn most of their income from user fees and profits from dispensing drugs. The number of private clinics has doubled from around 7,000 clinics nationwide in 1984 to around 14,000 today. In 2001, Thailand had 13,000 pharmacies, almost all of them privately owned and operated. Half of the pharmacies are classified as modern pharmacies with full-time pharmacists. The remainder are often referred to as drugstores and sell prepacked drugs that do not require a prescription. Traditional medicine drug- stores also exist, but by 2001, their share had declined to less than 15 percent of all drugstores and pharmacies. In addition, local grocery stores are major suppliers of drugs for common and minor ailments for villagers in several areas. In some places, prescription drugs such as steroids, antibiotics, and certain injections may be sold illegally by village grocery stores. Before the implementation of universal coverage, patients had to pay out-of- pocket for both services and drugs at public hospitals and health centers unless they had health insurance, were covered by public welfare systems, or were eligi- ble for free care. Nevertheless, public medical services were generally priced lower than services at private facilities, because staff salaries and capital investments were subsidized by the government. Public hospitals were allowed to keep reve- nues from user fees, which were frequently spent on drugs, medical supplies, and labor costs not already supported by the government. 140 Thailand: Achieving Universal Coverage with Social Health Insurance Health Insurance Development Before Universal Coverage Before implementation of the universal coverage policy in 2001, health care financ- ing systems were complex and involved multiple financing schemes. Each financing scheme had its own rules, regulations, and benefits packages for specific beneficiary groups. Approximately 70 percent of the population was covered by four public health insurance schemes (table 7.3), while private health insurance played almost no role. The remaining 30 percent, more than 15 million people, had no health insur- ance and paid out-of-pocket for health services and medicines. The development of Thailand's health insurance system has been a long process (figure 7.6). The first social security bill was proposed in 1954, but was never taken up in parliament. The first medical welfare program started in 1975, when the gov- ernment decided to provide medical services in public hospitals and health facili- ties to the poor free of charge. This program subsequently expanded to cover other underprivileged groups, the elderly, and children. Health insurance schemes for formal sector employees followed. The medical benefits scheme for civil servants, public employees, and their families (the civil servants medical benefits scheme or CSMBS) was established in 1980. The social security scheme (SSS) for private employees was first introduced in 1990. Efforts to expand coverage to informal sector workers were attempted with community financing schemes in 1983 and the voluntary health card scheme (VHCS) in 1991, but both programs failed. Universal coverage by health insurance was not accom- plished until 2002. Medical Welfare Scheme The government's medical welfare scheme started in 1975, when free medical care was offered to low-income individuals who passed a means test based on a cash income of less than B 1,000 per month. In 1981, the program was transformed into the low-income card scheme, whereby a card for free care was offered to eli- gible individuals (23 percent of the population at that time). The program further expanded to cover the elderly (those older than 60) and all children younger than 12, the disabled, monks and religious leaders, and war veterans. Beneficiaries were entitled to free medical care at public facilities: outpatient care, inpatient care, diagnostics, and medicines. The government paid public hos- pitals and health centers based on capitation and service utilization rates, but at a price much lower than cost recovery because of budgetary limitations. Hospi- tals therefore had to shoulder the resulting costs and had no incentive to provide quality services to this group of patients, resulting in long waiting times and low responsiveness. Health care providers cross-subsidized the scheme by charging other patients who belonged to other insurance schemes more. Despite its rapid expansion to cover 30 percent of population by 1999, the effectiveness of the medical welfare scheme in relation to means testing was lim- ited, affecting its ability to define eligible beneficiaries. Several studies showed that a significant proportion of medical welfare scheme members should not have been entitled to the benefits, as their incomes were higher than they should have been for eligibility. Moreover, some of those who really were poor and should have been included were not. The weakness of the means testing system resulted Table 7.3. The Four Main SHI Schemes before Universal Coverage, 1999 Civil servants medical Characteristics Medical welfare benefits scheme Social security scheme Health card Type of scheme Social welfare Fringe benefit Compulsory Voluntary Model Public integrated Public reimbursement Public contracted Voluntary integrated Coverage Groups covered Poor, elderly, children Government employees, Private formal sector Nonpoor households ineligible under 12, secondary retirees, and their employees working in for the medical welfare school students, disabled, dependents (parents, establishments with more scheme, community leaders, veterans, monks spouses, children) than 10 employees families of health volunteers Number of people, health 19.8 5.5 4.4 11.5 welfare scheme (millions) 14 1 Percentage of the 32.10 8.90 7.10 18.60 population covered Benefits Ambulatory services Designated public facilities Public facilities Public and private facilities Ministry of Public Health facilities Inpatient services Public facilities Public and private Public and private facilities Ministry of Public Health (emergencies only) facilities facilities Choice of provider Referral Free choice Contracted hospital or its Referral network, registration required Conditions included Comprehensive package Comprehensive package Comprehensive package Comprehensive package Maternity benefits Yes Yes Yes Yes Annual physical checkup No Yes No Yes Prevention, health promotion Limited No Health education, Yes immunization (table continues on following page) Table 7.3. The Four Main SHI Schemes before Universal Coverage, 1999 (continued) Civil servants medical Characteristics Medical welfare benefits scheme Social security scheme Health card Financing Source of funds General taxes General taxes Contributions by employees, Household: B 500 + per year, employers, and the government subsidy B 1,000 government of 1.5% of per year payroll each (reduced to 1% 14 2 in 1999) Management authority Ministry of Public Health Ministry of Finance Ministry of Labor Ministry of Public Health Payment mechanism Global budget Fee-for-service Capitation Proportional reimbursement Copayments No No except for inpatient care Maternity and emergency No at private hospitals services if beyond ceiling Expenditure per capita (B) More than 363 + additional 2,106 1,558 534 + additional cross-subsidy cross-subsidy by public by public hospitals hospitals Per capita tax subsidy 363 + additional subsidy 2,106 519 250 Piya Hanvoravongchai and William C. Hsiao 143 Figure 7.6. Development of Health Insurance Schemes, 1975­2002 1975 free medical 2001­2 care for the poor 1981 low-income card scheme 1991 voluntary health universal 1983 mother and child card scheme coverage health fund 1992 elderly 1984 health card 1994 children 1975 1980 1985 1990 1995 2000 1994 social security scheme for firms with 10 or more employees 1980 civil servants medical 1990 social security scheme for firms benefits scheme with 20 or more employees Source: WHO 2004. from difficulties in estimating the incomes of those working in the informal sector. Involvement by community leaders in proposing who should be included was help- ful in some communities, but it also created nepotism in some places. The medical welfare scheme was eventually merged into the universal coverage scheme in 2001. Civil Servant Medical Benefit Scheme The CSMBS is the government's medical benefits program for government employ- ees and retirees, state enterprise employees and retirees, and family members (spouses, children, and parents). The CSMBS is considered the country's most gen- erous insurance scheme, because it provides comprehensive health benefits, includ- ing inpatient and outpatient services in public hospitals and emergency services in public and private hospitals. Beneficiaries are not required to pay any premiums, as the government covers all costs. Health providers also like the scheme because of its generous fee-for-service reimbursement system. Until 1998, the CSMBS fully reimbursed all costs of treatment and prescription drugs on a fee-for-service basis without any cost sharing by patients. For outpa- tient services, patients paid up-front and were reimbursed by the scheme. For inpa- tient services, hospitals submitted their charges directly to the scheme and were reimbursed in full. The flexibility of the CSMBS made it a major source of income for many public hospitals, which cross-subsidized their losses from serving medi- cal welfare scheme patients using profits from the CSMBS. Statistics show that the admission rate per CSMBS beneficiary was higher than for the medical welfare scheme and that the length of stay was much longer. Some health providers also overcharged the CSMBS by overprescribing diagnostic services to generate the maximum amount of income. The CSMBS's expenditure per capita was therefore the highest among all the health insurance programs and increased by 20 percent per year in nominal terms during 1988­97 despite a low medical inflation rate. With the pressure to control costs, especially after the onset of the economic cri- sis, the Ministry of Finance, which is in charge of the CSMBS, has slowly imple- mented several cost-cutting measures since 1998, including reducing benefits and reforming the payment system. The private inpatient care benefit was discontin- ued and drug benefits were reduced to those listed in the national essential drugs list unless deemed necessary by a doctor. In 2003, the scheme changed its provider payment system for inpatient care from full reimbursement to diagnostic-related group reimbursement. 144 Thailand: Achieving Universal Coverage with Social Health Insurance Social Security Scheme The Social Security Act was promulgated in September 1990. All private enterprises with at least 20 employees had to participate in the SSS. In 1994, the SSS expanded to cover all enterprises with at least 10 employees, and in 2002 expanded again to cover all enterprises. The SSS covers 6 million people (10 percent of the population). It is regulated by the Ministry of Labor and managed by the Social Security Office. The SSS provides a comprehensive package of benefits, including coverage for sickness and disability and maternity and death benefits. Contributions to the Social Security Fund come from employers, employees, and the government at the rate of 1.5 percent of payroll each. This rate was reduced to 1 percent each during the economic crisis to reduce pressure on enterprises. Public and private hospitals are contracted to provide inpatient and outpatient services to the beneficiaries. These hospitals are frequently referred to as main con- tractors. Beneficiaries can select a contractor of their choice and can then obtain health services only from the network of the main contractor they selected with no copayments. The Social Security Office pays each hospital on a capitation basis, that is, payments are based on the number of beneficiaries registered with that hospital. The capitation payment is B 1,505 per person (in 2002) and is intended to cover the costs of fees, outpatient and inpatient care, and drugs. Under this kind of system, providers have a big incentive to keep costs to a minimum. They have no financial reason to deliver unnecessary services or to overprescribe, so the cost to the scheme is easily controlled. At the same time, the capitation payment system also created several problems, such as an undersupply of procedures that were more costly for providers; pref- erence for younger, healthier clients; and poor responsiveness to patients. There was an initial concern about the quality of care, as the quality control system was initially weak and the information system was incomplete; however, several mechanisms have since been introduced to improve the scheme, including quality reviews, hospital accreditation, and adjustments to provider payments to combat the problems noted above. The Social Security Office also recently introduced extra payments for high-cost services such as emergency care and renal dialysis and pro- vides supplementary payments based on utilization rankings and the number of beneficiaries with chronic cases. Voluntary Health Card Scheme The VHCS evolved from a community financing pilot initiative in 1983 that aimed to cover informal sector workers. At that time, health cards were sold in each com- munity to finance the village mother and child health development funds. The card entitled its purchaser to minimal benefits, including maternal and child health services, vaccinations, and simple treatments. The funds functioned as revolving funds, to be lent to members for toilet construction and other public health activi- ties. Because of limited participation, unclear government policy, and weak sup- port, the initiative was subsequently phased out. In 1991, the program was transformed into a voluntary health insurance scheme managed by the Ministry of Public Health and was implemented nation- wide in 1994. Its financing came from premiums collected from selling the cards and matching funds from the government. A health card for a family of up to five Piya Hanvoravongchai and William C. Hsiao 145 members was priced at B 500 and the government contributed an additional B 500 later (B 1,000). Health card benefits included outpatient care for illness and inju- ries, inpatient care, and mother and child health services at local public provid- ers. Drug benefits included medicines on the national essential drugs list with no copayments. With extensive mass media advertising and government support, the VHCS was quite popular in several areas, with up to 12 percent of Thai as beneficiaries at one point. The funds from card sales and government matching contributions for each province were collectively managed by the local provincial health office. Most health care providers received a global budget allocated from the local health card fund based on the number of cards they sold, but this amount was inadequate given the actual costs of the medical care the hospitals provided and low cost recovery. For example, in 2000 the average expense per card was B 2,700, whereas the income per card was only B 1,000 to B 1,500. One main reason for the low cost recovery was the voluntary nature of the VHCS, which subjected the scheme to adverse selection and system abuse. Utiliza- tion statistics reveal that VHCS beneficiaries used outpatient and inpatient services more often than the general population. Self-selection was evident, as pregnant women and patients with chronic diseases frequently purchased health cards fol- lowing diagnosis. Low levels of cost recovery and questionable financial sustain- ability led to the government's decision to end the program in 2001. Summing Up In summary, before the introduction of universal coverage, Thailand had several health financing schemes under various management organizations with dif- ferent payment mechanisms and funding methods. Budget subsidies, benefits packages, quality of service provision, and rates of utilization differed from one scheme to another (table 7.4). Inefficiencies were prominent in almost all schemes for various reasons, including adverse selection, moral hazard, and allocative inef- ficiency. Despite the effort to increase coverage by expanding the SSS to formal sec- tor employees, extending the public welfare system to cover more underprivileged groups, and introducing voluntary health insurance to capture informal sector workers, 30 percent of the population was still uninsured (figure 7.7). With a large proportion of the population working in the informal sector, the consensus was Table 7.4. Insurance Schemes and Utilization of Services, 1996 Insurance Payment Outpatient Admission Average length program mechanism visits/capita rate/capita of stay (days) CSMBS Fee-for-service 5.5 13.6 11.9 SSS Capitation 1.4 2.6 5.6 VHCS Capitation 1.7 5.8 4.3 Medical welfare scheme Global budget 0.7 3.0 5.1 Out-of-pocket Fee-for-service -- -- -- Source: Donaldson, Pannarunothai, and Tangchaorensathein 1999. Note: -- = not available. 146 Thailand: Achieving Universal Coverage with Social Health Insurance Figure 7.7. Trends in Coverage by Health Insurance Schemes, Selected Years 100 ed 80 cover 60 population the 40 of 20 centage per 0 1991 1996 1999 2004 year uninsured universal coverage private and other insurance VHCS medical welfare scheme SSS CSMBS Sources: NHSO 2005; Pramualratana and Wibulpolprasert 2002. that relying on voluntary insurance would not be a realistic approach for achieving universal coverage. Introduction of the universal coverage scheme in 2001 was a major political decision accompanied by significant health financing reform to improve the acces- sibility, equity, and efficiency of the health financing system. The universal cover- age scheme merged two existing publicly financed programs, the medical welfare scheme and the VHCS, and introduced major changes to funding and payment mechanisms. Coverage also expanded to include all those who were not already covered by the CSMBS or the SSS. Achieving Universal Coverage Before the implementation of universal coverage in 2001, the issue of how to achieve universal coverage had been regularly debated and discussed since the early 1990s. Policy makers and technocrats recognized that relying on existing schemes to expand coverage to the uninsured population, especially those in the informal sec- tor, was unlikely to be successful. Several meetings and workshops had been organized to discuss possible approaches to universal coverage. For example, in 1993, the National Economic and Social Development Board, the Ministry of Public Health, and the World Bank organized a workshop on expanding health insurance, but the participants could Piya Hanvoravongchai and William C. Hsiao 147 not reach consensus on how to achieve universal coverage. In 1996, the Ministry of Public Health and the Standing Committee on Health of the House of Represen- tatives prepared a draft of the National Health Insurance and Standard Medical Care Bill, which would set up a compulsory health insurance scheme for those not already covered by other schemes. The new scheme would be funded from pre- mium collections and the government would subsidize premiums for the poor; however, parliament was dissolved before the bill was considered. The 1997 constitution resulted in an environment conducive to developing uni- versal coverage. The constitution provides a strong foundation for the concept of people's right to health. Article 52 notes that every Thai has "an equal right to receive standard public health services and the indigent shall have the right to receive free medical treatment from public health facilities of the state." Article 82 adds that the state "shall thoroughly provide and promote standard and efficient public health services." Universal coverage by health insurance was seen as an important element of people's right to health. In the spirit of the 1997 constitution, the Working Committee on Universal Cover- age was formed in 2000 with the support of the Health Systems Research Institute to study potential alternatives for universal coverage. The committee studied and con- firmed the financial feasibility of achieving universal coverage, noting that it would be affordable assuming an efficient health insurance system. A total of B 100 billion per year was the expected level of funding required for universal coverage, which was not much higher than the B 76 billion per year spent on various health insur- ance programs and public health budgets. However, this implied a major financing reform that would reroute funds from many different sources to a single payer sys- tem. In addition, the committee proposed a focus on strengthening cost-effective pri- mary medical care and a split between health care purchasers and providers. Three alternative approaches to achieving universal coverage were proposed: expanding the existing financing system, introducing a single national health insurance system, and introducing a dual system of public and private health insurance. In parallel with the work of the committee, in 2000, a coalition of civil society groups drafted a version of the National Health Security Bill based on the Ministry of Public Health's bill prepared in 1996. The primary interest of the civil society groups was to increase consumer protection and consumer participation in health system management. The proposed bill would establish a single national fund to implement universal coverage based on a tax-financed mechanism instead of com- pulsory insurance with contributions by beneficiaries as in the Ministry of Public Health's bill. The media also played an influential role. They were active in keeping the pub- lic informed about various universal coverage debates and developments, resulting in heightened awareness of the issue. Universal coverage became a major national agenda item when the Thai Rak Thai Party--a new political party established by Thailand's richest business tycoon, Thaksin Shinawatra--embraced the plan to achieve universal coverage as a key component of its platform for the 2001 elections. The party's leaders adopted a policy on universal coverage, as they had envisioned its popularity and had learned about its financial feasibility and technical achievability from available studies and evidence. Following the party's landslide victory in the elections, the universal coverage policy was incorporated into the government's policy. A work- shop on universal coverage implementation was organized soon after with the 148 Thailand: Achieving Universal Coverage with Social Health Insurance prime minister as the chair. A task force to implement the universal coverage pro- gram was established and the Ministry of Public Health was tasked with program implementation. With several members of the Working Committee on Universal Coverage par- ticipating in the workshop and the task force, the implementation plan was laid out similarly to the ways the committee had previously recommended. This entailed a major reform of the Ministry of Public Health's payment system for health facilities (from a historical budget system to capitation) and universal coverage beneficiary cards were distributed first to previous medical welfare scheme and VHCS mem- bers and then to the uninsured in April 2001. The program started in six pilot prov- inces and expanded to cover the entire nation in April 2002. Along with the CSMBS and the SSS, Thailand has achieved universal coverage with three health financing schemes (table 7.5). Universal coverage cardholders are eligible to receive comprehensive health care at the provider with which they registered with minimal copayments per out- patient visit of B 30 (US$0.75). Government budgets for the medical welfare scheme and the VHCS were merged with the Ministry of Public Health budget for health care delivery to create a common budget line for universal coverage. Health care providers receive capitation payments based on the number of cardholders regis- tered with them. A contingency fund was set up to provide temporary alleviation to health care providers whose capitation budgets were significantly lower than their previous historical budgets based on the number of staff and hospital capacity. Concurrent with implementation of the universal coverage policy, the Ministry of Public Health drafted the National Health Security Bill in April 2001 to set up the NHSO and the National Health Security Fund as formal mechanisms to manage and finance universal coverage. The fund would be financed by general tax rev- enues to cover health services for beneficiaries not enrolled in the CSMBS or the SSS. The bill also envisions that this fund will be a single health care purchaser in the future when the CSMBS and the SSS transfer health services purchasing func- tions to the fund. Enactment of the National Health Security Bill was not without opposition. It easily passed the Council of Ministers in July 2001, but faced strong lobbying pres- sure from various interest groups in parliament. The medical profession supported the bill in relation to universal coverage, but opposed the medical liability clause provided in the bill and voiced their concern about the power of a single health care purchaser. The labor unions were worried about the diversion of funds from the Social Security Fund to the universal coverage scheme and perceived the benefits under universal coverage to be inferior to what they already had. The unions also organized several street protests to voice their concerns against the merging of the SSS with the universal coverage scheme. Technocrats in the Ministry of Finance expressed worries about potential long-run budget implications of the tax-financed program, noting that it could increase public debt, which was already high. In addi- tion, the private health care sector exerted pressure to ensure that the bill would allow private sector participation in the new scheme so that monies from the uni- versal coverage fund would also be available to them. With minor revisions of the text to accommodate requests and strong control of parliament by the Thai Rak Thai Party, the National Health Security Bill was passed and the National Health Security Act was promulgated in August 2002. The NHSO Table 7.5. The Three SHI Schemes after Universal Coverage, 2004 Characteristics Universal coverage CSMBS SSS Type of scheme Compulsory Fringe benefit Compulsory Model Public contracted Public reimbursement Public contracted Coverage Groups covered All those not covered by the CSMBS Government employees, retirees, and Private formal sector employees and SSS their dependents (parents, spouses, children) Number of people, health welfare scheme (millions) 46.6 4.4 8.2 14 9 Percentage of the population covered 75.24 7.07 13.16 Benefits Ambulatory services Public and private facilities Public facilities only Public and private facilities Inpatient services Public and private facilities Public and private (emergencies only) Public and private facilities facilities Choice of provider Contracted hospital or its network, Free choice Contracted hospital or its network, registration required registration required Cash benefit No No Yes Conditions included Comprehensive package Comprehensive package Comprehensive package Maternity benefits Yes Yes Yes Annual physical checkup No Yes No Prevention, health promotion Health education, immunization No Health education, immunization Table 7.5. The Three SHI Schemes after Universal Coverage, 2004 (continued) Characteristics Universal coverage CSMBS SSS (table continues on following page) Financing 15 Source of funds General taxes General taxes Contributions by employees, 0 employers, and the government of 1.5% of payroll each (reduce to 1% in 1999) Management authority NHSO Ministry of Finance Social Security office Payment mechanism Global budget + capitation Fee-for-service (experimenting with Capitation (experimenting with diagnostic-related group payments diagnostic-related group payments for inpatient care) for inpatient care) Piya Hanvoravongchai and William C. Hsiao 151 was then set up as a quasi-public organization with the minister of public health as chair of the NHSO's executive board. Fund operations and management for the universal coverage program were transferred from the Ministry of Public Health to the NHSO, which became the health care purchaser for universal coverage ben- eficiaries. Several committees have been set up to improve the universal coverage system and to prepare for the future merger of health purchasing activities. Lessons Learned from Universal Coverage Thailand has achieved universal coverage through a long process of health financ- ing development and by means of technical knowledge, powerful public support, and strong political commitment. The country has learned from the successes and failures of several financing schemes. Implementation of the universal coverage program, albeit rapid, experienced several technical hurdles. The first problem was how to identify those who were uninsured (not already covered by the SSS and the CSMBS) given that no database of CSMBS beneficiaries existed. A comprehensive information system was therefore rapidly developed using a government registration database to avoid the duplica- tion of health insurance benefits. This was not an easy task considering that more than 50 million beneficiary records had to be created. The second problem concerned the reform of provider payment that accompa- nied the implementation of universal coverage. Despite the clear decision to use capitation as the payment mechanism, several detailed policy options required careful planning and implementation. First, an appropriate level for the capitation payment was the subject of major debates at the time of implementation. Based on prior costing studies, setting a suitable capitation rate would ensure cost recovery and would provide providers with incentives to deliver good quality health ser- vices. At the same time, the capitation payments could not burden the government with an overly high budget. However, the costing exercises resulted in a range of capitation rates depending on the assumptions used. The government initially went with the lowest choice, but this created a public outcry from health providers, who were not satisfied with the level. The reform of the provider payment system also included salary costs in the cap- itation rate, which meant that payments from the central government to public hos- pitals did not include fixed salary payments. Public health facilities therefore faced higher risks of operating at a loss, as wages and salaries are generally a major share of their operating budgets. The inclusion of salaries in the capitation rate gave large hospitals with many staff in areas of low population density a much smaller total budget than they had received prior to universal coverage. With limited authority to fire or relocate staff, the fixed costs of wages became a major burden and there were protests from large public providers. With strong political pressure, the uni- versal coverage program finally allowed the Ministry of Public Health to separate salary payments from the total Ministry of Public Health capitation budget at the central level and allocate the remainder to providers on a capitation basis. How to contract providers efficiently and effectively is another issue of major concern and technical difficulties. The universal coverage program learned from the experience of the SSS, which has general and provincial hospitals as the main 152 Thailand: Achieving Universal Coverage with Social Health Insurance contractors. Smaller hospitals and clinics can participate as subcontractors in the network of these main contractors to ensure adequate risk pooling among benefi- ciaries, especially for inpatient medical care costs, which could be high. A recent proposal suggested using clinics as the main contractors for outpatient services to promote primary care and system efficiency. In 2004, outpatient service contracting with clinics started in urban settings, where private clinics are allowed to be main contractors and receive capitation payments for outpatient care and for prevention and promotion services. In 2004, three-quarters of all contracted facilities were Ministry of Public Health hospitals. More than 90 percent of universal coverage beneficiaries registered with Ministry of Public Health facilities, while only 5 percent registered with private pro- viders. The payment system has also evolved from all-inclusive capitation of main contractors to a more complex system of payments. Of the B 1,309 per capita bud- geted for the universal coverage scheme (table 7.6), 85 percent was spent on medical care such as outpatient care, inpatient care, and prevention and promotion activities by means of capitation payments to main contractors (B 1,112.74 per head in 2004). The remaining 15 percent (B 195.80 per head) was retained by the NHSO, which spent the funds on high-cost care, accidentand emergency care, and no-faultliability payments using diagnostic-related group payments, point systems, or claim based payments. Several changes have been introduced to improve efficiency and equity. The decision to keep budgets for high-cost care and emergency care at the central level is to help with risk pooling across contractors. The NHSO also introduced an adjustment for age structure in the calculation of capitation rates to compensate for any differences in the risks contractors face. In addition, the NHSO pays bonuses for priority programs or treatments, such as an extra per case payment of B 1,000 for a cataract operation and B 100,000 per open heart surgery case. It also allocated B 115 million for the Department of Health to strengthen the control and preven- tion of cervical cancer. Table 7.6. Breakdown of the Universal Coverage Scheme Capitation Rate and Payment Methods, 2004 Percentage share Capitation of total average Reimbursement item payment rate payment per capita Payment method Outpatient care 488.18 37 Capitation payments Inpatient care 418.36 32 Capitation payments Prevention and promotion 206.00 16 Capitation payments High-cost care 66.30 5 Diagnostic-related group payments Accident and Diagnostic-related emergency care 19.70 2 group payments Emergency medical services 10.00 1 Point systems Remote areas 10.00 1 Add to capitation Capital replacement 85.00 6 Add to capitation No-fault liability 5.00 0 Claim based payments Total 1,308.54 100 Piya Hanvoravongchai and William C. Hsiao 153 Note that the payment system for Ministry of Public Health contractors is more complex than for private contractors. With political interests in retaining some financing power at the central Ministry of Public Health level, the ministry success- fully negotiated to have the total capitation budget go through the central Ministry of Public Health instead of allowing the NHSO to make direct capitation payments to the Ministry of Public Health's main contractors. One argument in support of this arrangement was to smooth the transition phase and to allow the Ministry of Public Health to deduct a proportion of the funds for salary payments before allo- cating the remainder to main contractors. This arrangement clearly reduced the pressure on large public hospitals with many staff, as their salaries are guaranteed; however, it is detrimental to hospitals that currently have low staff levels and need to expand to cover a large beneficiary base. Despite the problems encountered during implementation, the universal coverage scheme is considered to be the most popular government policy and received high public satisfaction ratings in all public polls. The scheme brought a new perspective to the population's "right to health," which contrasts with the previous view of those receiving free health care services as being welfare recipients. Capitation payments preservealinkbetweenconsumersandservicepurchasers,asmoneyfollowspatients. Utilization statistics show an increase in outpatient visits at health centers and district hospitals and a decrease in outpatient services provided by general hospitals, reflect- ing an effective gatekeeping system that emphasizes primary medical care. Challenges Despite its success, the universal coverage scheme still faces several challenges. The first concerns long-term funding. The NHSO currently receives funds from general tax revenues, which requires annual negotiations on the capitation budget between the NHSO and the Ministry of Finance. Recent years have seen a clear tendency toward approving a lower budget per capita than what is adequate for cost recov- ery. In addition, funding available to the NHSO to pay for universal coverage does not have any endowments, and thus is not prepared for long-term expenses. With changes in health care technology and an increase in population aging, the pro- gram's costs are expected to continue to increase and the scheme's financial sus- tainability will be at risk. Alternative long-term funding mechanisms are therefore necessary. Several studies are currently exploring the possibilities of getting funds from other sources, such as a sin tax (on tobacco or alcohol) or premium payments. The second major challenge concerns the merger of the universal coverage scheme, the CSMBS, and the SSS. As noted earlier, the three schemes receive dif- ferent levels of funding per person, have different payment mechanisms, and cover different benefits (table 7.5). Article 9 of the National Health Security Act indicates that the universal coverage scheme will eventually be the only health insurance scheme, as the CSMBS and the SSS will purchase health services for their benefi- ciaries from the NHSO. However, the process of merging is not easy given the large differences between the schemes. To facilitate the future merger, the Coordination Committee on Health Insur- ance Development was set up in January 2004 and tasked with coordinating the development of and monitoring progress on benefits packages, duplication of eli- gibility, payment methods, standards of care, health facilities, and portability of eligibility and sharing information on claims, financial reports, claims audits, and 154 Thailand: Achieving Universal Coverage with Social Health Insurance medical audits for the three schemes. A successful merger requires (a) developing information infrastructure and a management information system, (b) setting up a common benefits package and quality and standards of care, and (c) developing payment methods and claims processing systems. As of October 2005, the three schemes had started to synchronize their provider payment administrations by merging the claims processing activities to be carried out through the NHSO. Conclusion The development of SHI and a health financing system in Thailand has been a lengthy process. The plan to establish a health insurance scheme started more than 50 years ago, when the SSS for private employees was devised. The first public health financing scheme started in 1975 as a medical welfare scheme for the poor and was expanded to cover other underprivileged groups. Formal sector employ- ees, both public and private, have health insurance programs that provide medical care benefits, but the expansion of health insurance coverage to the informal sector had been slow and ineffective, and the reliance on a voluntary health insurance program to achieve universal coverage proved to be impractical. In 2002, Thailand achieved universal coverage through the government's strong political commitment and with support from civil society and a strong knowledge base. The new financing system relies on tax financing to cover the informal sector. The CSMBS and the SSS still exist to cover government employees and dependents and the private formal sector, respectively. The major challenge facing Thailand's health financing system is how to merge the three existing schemes together into a single fund. This requires a good infor- mation system and a convergence of benefits packages and provider payment meth- ods. There is also concern about the financial sustainability of universal coverage as the population structure ages and technology changes. The system is still develop- ing, and changes are needed to improve the system's efficiency, equity, quality, and sustainability. References CIA (Central Intelligence Agency). 2006. The World Factbook. McLean, VA: Central Intelli- gence Agency. http://www.cia.gov/cia/publications/factbook/. Donaldson, D., S. Pannarunothai, and V. Tangchaorensathein. 1999. Health Financing in Thai- land. Thailand: Health Financing and Management Study Project, Asian Development Bank Report 2997 THA. Nonthaburi, Thailand: Health Systems Research Institute. NHSO (National Health Security Office). 2005. Annual Report 2004: Progress and Achievement. Nonthaburi, Thailand: NHSO. Pramualratana P., and S. Wibulpolprasert, eds. 2002. Health Insurance Systems in Thailand. Nonthaburi, Thailand: Health Systems Research Institute. Transparency International. 2006. Corruption Perception Index. London: Pluto Press. Wibulpolprasert, S., ed. 2005. Thailand Health Profile 2001­2004. Nonthaburi, Thailand: Min- istry of Public Health. WHO (World Health Organization). 2004. Regional Overview of Social Health Insurance in Southeast Asia. New Delhi: Regional Office for Southeast Asia. ------. 2006. World Health Report 2006. Geneva: WHO. 8 Lessons Learned and Policy Implications William C. Hsiao and R. Paul Shaw SHI has been touted as an instrument for mobilizing additional revenues for health and pooling risks between the rich and the poor, the sick and the healthy in low- and middle-income countries. SHI also appeals to policy makers as an approach to improving the performance of national health systems along four dimensions: rev- enue mobilization for health, equity of access, efficient use of inputs, and financial risk protection. Revenue mobilization tends to be weak and disorganized in low-income coun- tries because the government's capacity to tax households and production is lim- ited. On average, tax revenues amount to only 14.5 percent of GDP and social secu- rity contributions amount to only 0.7 percent of GDP. These revenues are completely inadequate to finance public provision of quality health care in rural, peri-urban, and urban areas. Meanwhile, out-of-pocket spending on health typically accounts for 50 to 80 percent of total health expenditures. Access is inequitable because relatively poor households, which account for 50 to 60 percent of the population in low-income countries, tend to live in rural and more isolated areas where public, NGO, and private for-profit health services are sparse. The inequities tend to be aggravated by the concentration of public expendi- tures and subsidies for health on urban areas, where more households are relatively rich and better informed and have the means to access care, and on tertiary-level care. Figure 8.1 illustrates this situation for Ghana, where the poorest 20 percent of households benefit from only 11 to 12 percent of public expenditures on health, whereas the richest 20 percent capture more than 50 percent of such expenditures. The use of inputs tends to be inefficient because public finance and service pro- vision seldom face competitive pressures to demonstrate value for money, and private providers have few incentives to provide cost-effective health care. Ineffi- ciencies in the public provision of health services have been such that some stud- ies have suggested that publicly provided health services have no statistically significant impact on health outcomes in poor countries (see, for example, Filmer, Hammer, and Pritchett 1998). NGO providers, such as religious missions, have a much better track record in this regard, but are thinly distributed in poor areas and underfinanced. Financial risk protection is weak because so many households are vulnerable to catastrophic financial loss at times of serious illness or injury. Without access to well-functioning public hospitals or insurance by SHI, private health insurance, 155 156 Lessons Learned and Policy Implications Figure 8.1. Percentage of Total Public Expenditures on Health by Income Quintile, Ghana, 1992 and 1998 60 1992 1998 50 40 30 percent 20 10 0 bottom second third top two Source: World Bank living standards measurement surveys, 1992 and 1998. or community health funds, poor households must resort to out-of-pocket spend- ing at times of crisis. Data from China, India, and many African countries suggest that 30 to 40 percent of poor households incur indebtedness and/or sell important assets, such as farm animals or land, to pay for health care at times of crisis. This feeds a vicious cycle whereby the reduction of household assets and debt servicing undercuts spending on food, housing, and other inputs required for good health. Combined, these four performance problems shackle progress in countries where huge improvements in health system performance are required to respond to national and donor-driven challenges of improving the health of the poorest households, attaining the Millennium Development Goals, and achieving other national priorities. The problems place a premium on finding workable alternatives to more traditional responses (box 8.1). In developing nations, this sets the stage for intense interest in SHI, not merely as a way of mobilizing earmarked funds for health, but as an instrument for reforming the financing, organization, payment arrangements, and regulation of national health systems and the behavior of both providers and clients. SHI is further viewed as an organizational lever that can help shift financing and provision from predominantly supply-side strategies to demand-side strategies. As conveyed by the experience of three of the countries presented in this volume-- Colombia, the Philippines, and Thailand--demand-side strategies can be pro- moted in two ways. On the one hand, member contributions into an SHI fund can be used to purchase services that members want, from providers they chose, in close proximity to where they live. On the other hand, the SHI fund can contract with both public and private providers and hold them accountable for quality and client satisfaction. The fund can also use provider payments to incentivize pro- viders. For example, when capitation is used to remunerate providers and when William C. Hsiao and R. Paul Shaw 157 Box 8.1. Traditional Responses to Poor Health Sector Performance Traditional responses to poor health sector performance are guided by an epidemio- logical focus on reducing disease through cost-effective interventions that are largely financed and provided by the government as supply-side strategies. In support of such efforts, WHO has urged governments to increase revenue mobilization to the extent that national health spending exceeds 5 percent of GDP in low- and middle-income countries. International forums on development assistance have also urged multilat- eral and bilateral donors to increase grants and loans for health to help scale up public efforts at reducing disease. Multilateral and bilateral donors are also promoting debt relief to enable poor country governments to allocate savings from debt service pay- ments to health and education. The good news is that debt relief involves large sums, for example, US$30 billion up to 2005, with US$52 billion more in debt reduction promised by the Group of Eight during a recent meeting, much of it applicable to Sub-Saharan Africa. As part of these efforts, international NGOs are making laudable efforts to fill critical gaps in immu- nization and the fight against HIV/AIDS. The bad news is that traditional, supply-side responses often result in long-standing performance problems. One problem involves reliance on top-down government health systems in which the public sector dominates provision, allocates public expenditures largely to urban areas and tertiary care, has limited capacity to regulate quality, and lags in its ability to create an enabling environ- ment for private providers. While new infusions of funds come to the rescue tempo- rarily, the fact remains that government capacity to sustain higher spending through general revenue taxation will remain limited in poor countries and unable to capture individuals' out-of-pocket expenditures on health. Moreover, the introduction of user fees in public facilities raises little more than the equivalent of 10 to 15 percent of recur- rent costs, while often deterring the poorest households from seeking care. A second problem involves lack of confidence in the government as a provider of health services, as well as a lack of meaningful representation by stakeholders in health outcomes, including private and NGO providers, communities, and households. This lack of confidence is fully justified in many poor countries, where drugs are available for only one or two weeks per month in public hospitals and staffing is far below quan- titative and qualitative norms. A third problem involves the continued fragmentation of poor countries' health systems by a myriad externally funded projects, separate project implementation units, and independent monitoring and evaluation studies. Harmonization of such efforts has always been a problem, renewing criticisms that donors and international NGOs frequently create parallel delivery systems that invariably undercut the capacities of national systems. Moreover, external aid tends to be short-term and volatile, and often dries up when donors pull out or fail to deliver on their commitments. International Monetary Fund analysts suggest that aid is 20 times as volatile as government revenues as a percentage of GDP and 40 times as volatile as government revenues in constant U.S. dollars per capita. On average, only 60 percent of promised aid actually makes it to programs (Levine 2005). clients have a choice among several providers, providers must satisfy the clients or risk losing them, and thus their capitation payment, to a competing provider. In other words, the system must be responsive to the client rather than simply supply tax-financed services for free while expecting clients to come forward whether or not they are happy with the services provided. This is the essence of giving more power to the demand side and promoting greater accountability in relation to the financing and provision of health care through SHI. 158 Lessons Learned and Policy Implications In short, done properly, SHI can be expected to improve a country's risk protec- tion and health status outcomes by · mobilizing more revenue for health; · improving equity by risk pooling contributions by the rich and the poor, the sick and the healthy; · reducing catastrophic financial loss at times of serious illness or injury, and thus the vicious cycle of indebtedness, debt servicing, and reduced house- hold expenditures on necessities; · expanding access to quality services by the insured; · facilitating more efficient purchasing of health care for SHI members in the quest to achieve value for money; · facilitating the separation of financing and provision, whereby the SHI fund manages the financing and contracts out to public and private providers to deliver services; · involving more stakeholders--industrial groups, cooperatives, religious groups--in the determination of their own health outcomes; · being more responsive to clients' preferences and complaints; · freeing up scarce public revenues (from general taxation) for more effective targeting to the poor. Using SHI as an instrument of reform in low- or middle-income countries is not to exclude the public sector or deny that it will continue to play essential functions. Rather, SHI envisages a redefinition of and increased emphasis on various public roles. Typically, this implies making better use of the public sector's comparative advantage in being a steward of good health system practices, financing public goods and health services with positive externalities, and subsidizing the poor. The Development Context Matters While high hopes may accompany efforts to launch SHI in developing nations, the reality is that a country's level of economic development has a major impact on how many people can be covered and how rapidly SHI can expand toward universal coverage. Among richer nations, for example, several structural features of the economy contribute positively to the enabling environment for SHI. As discussed in chapter 1, these pertain to the share of the population employed in the formal sec- tor; the extent of administrative capacity; and the levels of per capita income, pov- erty, and dependency ratios; and so on. This context can be sharply distinguished from conditions in low-income, developing countries, as illustrated in table 8.1. A rule of thumb is that a country is in a good position to achieve universal coverage through SHI when its per capita gross national product is above US$6,000 per year. Another reality is that some countries may not be politically ready or willing to support the introduction of SHI. Some of these countries may have a strong tradition of health care financing and provision modeled after the Beveridge approach, which relies on general revenue taxation. In such contexts, switching to SHI or complementing general revenue financing with earmarked funding for health from SHI may confront widespread resistance. In other contexts, SHI might be resisted as benefiting relatively rich employers and employees at the expense of the poor and creating a two-tier health system in the process. In other words, William C. Hsiao and R. Paul Shaw 159 Table 8.1. Selected Characteristics of Country Groupings by Income Level, 2000 Total per capita Percentage Percentage expenditure of the of the Per capita on health population population Dependency Income level GNP (US$) (US$) that is urban that is poor ratio Low 390 29 10­30 40­60 0.9 Upper-middle 3,730 242 30­70 20­40 0.7 High 26,160 2,977 70­100 Less than 15 0.5 underlying political economy dimensions tend to be critical. Resistance is likely to be aggravated if governance and corruption issues have been problematic. With- out widespread political support, based on extensive stakeholder consultation, favorable economic conditions alone will be unable to sustain the introduction of SHI and its expansion toward universal coverage. The Five Country Case Studies The five case studies in this volume shed light on the trials and tribulations of implementing SHI in contexts that are far less hospitable than those in relatively rich countries. They were selected to reflect on a continuum and time line of oper- ational stages, beginning with the initial design of SHI, the first phase of imple- mentation, the expansion to cover larger segments of the population, and on up to completion, whereby SHI becomes the predominant form of health insurance in a country. Accordingly, Kenya has been selected for illustration of the design stage, Ghana for initiation, the Philippines for extension of population coverage, Colom- bia for fuller implementation of SHI and managed care, and Thailand for univer- sal coverage. These countries also differ in relation to various criteria known to be important to the enabling environment for SHI. A low per capita income, a small formal sector, a high prevalence of poverty, and a high dependency ratio can be expected to be particularly demanding. As explained in chapters 1 and 2, SHI is a complex instrument of reform. Done well, it can yield positive outcomes over time. Done hastily, it can be backward, dis- ruptive, and possibly hazardous. Against this cautionary background, lessons and policy implications from the five case studies can be summarized in terms of the following three dimensions: · positive changes that can be attributed to SHI, · major problems that challenge implementation, · implications for policy makers. Positive Changes Attributable to SHI The experiences of the five countries presented in this book suggest that at various stages of implementation, SHI can be credited with at least 10 positive changes as follows: 160 Lessons Learned and Policy Implications 1. The initiation of SHI facilitates national debate and consensus on financ- ing health care and allocating resources. SHI requires legislation to provide a legal framework for authorizing mandatory, earmarked contributions for health, as well as creating a new, autonomous organizational entity to man- age the program. The legislative process prompts national debate by par- liamentarians, raising public awareness of the problems facing the health sector, especially sustainable financing and possible options to improve it. 2. The implementation of SHI has succeeded in raising more revenues for health in addition to existing revenues raised by general taxation. In the Philippines, for example, the share of SHI in total public health expenditures grew from 8.9 percent in 1998 to 23.4 percent by 2002 and from 3.8 percent of total health expenditures in 1998 to 9.1 percent by 2002. In Colombia, health expenditures increased from 6.2 percent of gross national product in 1993 to 8.1 percent by 2002, with the share of SHI rising from 1.7 percent of GDP in 1993 to 4 percent by 2002. This is no small accomplishment in countries struggling to meet the general target set by WHO that countries spend at least 5 percent of their gross national product on health. 3. The implementation of SHI constituted a formal mechanism for pooling rev- enues and spreading risks across population groups, from rich to poor and across the life cycle. In some countries such as the Philippines, this was vis- ible in a reduction in out-of-pocket expenditures for health from 50.4 percent in 1995 to 40.5 percent by 2000, though out-of-pocket expenditures climbed to 46.8 percent by 2002. In Colombia, it is visible in the transfer of approxi- mately 10 percent of SHI revenues to subsidize the poor, which reduced financial barriers to health care. Moreover, only 4 percent of the popula- tion covered by SHI fell below the poverty line in Colombia as a result of an ambulatory or hospital shock, compared with up to 14 percent of non- SHI members. While acknowledging that progress on improving financial fairness and risk pooling has been extremely slow in many countries, (WHO 2000) SHI clearly represents a promising response to the challenge. 4. The implementation of SHI has forced more careful and rational planning regarding the imperative of equating SHI revenues with SHI expenditures. In the context of SHI, vague, overly generous, and unaffordable benefit enti- tlements are more readily viewed as a pathway to bankruptcy. In Kenya, this influenced the president to refuse to sign into law the legislation to imple- ment the program in 2004. Moreover, SHI management is increasingly using unit costing to determine the actual costs of different benefits packages and to ascertain their affordability. Tough choices have been required regarding services to be included versus those to be excluded. Copayments and deduct- ibles have had to be considered as a way of shoring up lagging revenues. 5. The social contract between contributing members and the SHI system pro- vides for grievance procedures if benefit entitlements have not been honored. Such provisions that aim to make SHI more responsive to clients are among the criteria WHO (2000) uses to assess the performance of national health systems. In other words, SHI has given new meaning to the term "account- ability" in several developing countries in ways not previously seen in pub- licly financed and provided health care. William C. Hsiao and R. Paul Shaw 161 6. The increasingly popular policy of separating public finance from public provision of health care is incorporated in SHI. This policy stems from the impression that governments often do a good job of raising funds for health, but perform less well in directly providing health care through publicly owned and operated hospitals and clinics. The proposed alternative is for government and other entities that collect revenues for health to contract with various types of accredited providers: public, NGO, private-for-profit. This means that the SHI fund performs more as a purchaser of quality health care than as a direct provider, striving to get the best value for money from public, private, or NGO providers under contractual agreements. Moreover, economists view the increased use of contracting with both public and pri- vate providers as a way of stimulating the market to be more competitive and of making providers more accountable for their performance. In Kenya, for example, SHI legislation proposes that national hospitals become financially autonomous from the Ministry of Health within 2 years of the establishment of SHI, provincial hospitals within 4 years, district hospitals gradually over 6 to 10 years, and health centers and dispensaries within 10 years. In Colom- bia, the encouragement to set up provider health plans expanded coverage by private health plans from 11 to 30 percent of the population between 1995 and 1998. At the same time, expansion of the market did not exclude public providers previously on the government payroll, because a minimum of 40 percent of health plan contracts go to public providers by decree. 7. The debate on subsidizing and expanding coverage for the poor and the indi- gent that has accompanied SHI has inspired more realistic consideration of equity. As explained in chapter 2, the publicly financed and operated health care system often evolves into a three-tiered health care structure. While the reality is widely known, critical review and public debate about the unequal access to reasonable health care between the poor and rich are limited. Under SHI, the lack of coverage of the poor and near-poor is visible and quantifiable and creates political pressure for remedial action. Thus SHI in Ghana aims to expand membership incrementally so that 30 to 40 percent of the population is covered in the first 5 years and 50 to 60 percent in next 5 to 10 years. SHI in the Philippines recently stepped up subsidies to the poor to the extent that indigents as a percentage of the membership increased from 16 percent of the total in 2003 to 30 percent by 2005 (and discounting the 40 percent in 2004, a temporary increase during an election year as explained in the case study). SHI in Colombia doubled subsidies to poor households to the extent that 28 percent of poor households were covered in 2003, compared with 14 percent in 1991. By 2003, 50 percent of the poorest income quintile were included in SHI, compared with fewer than 10 percent in 1991. These trends and commit- ments demonstrate that SHI can enhance equity. 8. The onus on SHI to achieve value for money has encouraged new thinking and experimentation with different forms of provider payments. Several countries, such as Ghana and Kenya, continue to rely on retrospective fees- for-service, but others are shifting to prospective capitation for primary and basic curative care, perhaps combined with a fee-for-service arrangement for more expensive surgical procedures as in Colombia and Thailand. The shift 162 Lessons Learned and Policy Implications away from fee-for-service has helped SHI funds reduce the administrative complexity of processing thousands of individual bills, as well as provid- ers' tendency to ramp up the volume of patients in the quest to manipulate earnings. 9. The creation of a new organization responsible for raising earmarked reve- nues for health and contracting with providers has resulted in a clarification and redefinition of the roles of ministries of health. In Ghana, for example, the Ministry of Health aims to become a policy maker and regulator of health care rather than a provider of services. Accreditation of providers is an especially important role that needs to be filled through a government regulatory body. Also critical is the need to better target public funds from general taxation to the poor, either in the form of directly financed provision of care or though subsidized membership in SHI. Moreover, with more earmarked SHI funds for health, government should be better equipped to channel public revenues to public health goods and services. This can be done in conjunction with SHI, as in the Philippines, where providers that offer services such as DOTS and maternity coverage are reimbursed by SHI. In Colombia, approximately 4 percent of SHI revenues are placed in a promotion and prevention fund to finance health promotion and prevention activities by contracted health plans. Another 2 percent of SHI contributions is set aside for maternity leave benefits. 10. A commitment to SHI over the long term has seen success in expanding membership rather than simply stalling or leveling off. Thus in the Philip- pines, the formalization of SHI in 1995 resulted in a steady climb in the pro- portion of the population covered by insurance from about 50 percent to 78 percent by 2004; in Colombia, coverage grew from less than 30 percent in 1992 to more than 60 percent by 2003; and in Thailand, where 68 percent of the population had been covered under various risk-pooling schemes, the expansion and consolidation of SHI is what set the stage for almost complete universal coverage in 2004. Major Challenges Inherent in Implementation The experiences of the five countries also indicate that at various stages of develop- ment, SHI can expect to encounter at least nine major implementation problems, namely: 1. Mandatory SHI needs to be enforced. Passing a law and creating an orga- nization to collect premiums is relatively easy, but actually collecting those premiums is another matter. Collection will be easiest for civil servants using regular payroll lists and monthly deductions. Collection from large formal sector employers and employees will also be relatively easy. Collec- tion will be much harder among smaller enterprises in the formal sector, with evasion by both employers and employees being a major problem. In the Philippines, for example, the Office of the Actuary estimates that for small employers in the formal sector, only 30 percent of those who should be contributing actually do. In Colombia, estimates indicate that only 65 William C. Hsiao and R. Paul Shaw 163 percent of potential contributors are actually paying, with evasion decreas- ing revenues by up to 35 percent. 2. Dependents of contributing members present another challenge. The poorer the country, the higher the dependency ratio, largely because of high levels of fertility and large family size. If dependents are excluded from SHI bene- fits, as was initially the case in Costa Rica's and Thailand's SHI systems, then revenue from workers' contributions only pays for the workers' health care costs. To cover workers' dependents, the contributions must be more than doubled. A favorable financial condition for including dependents is that the premium rate for a child is typically only about one-quarter of that for an adult. In the Philippines, approximately 25 percent of benefit payments go to beneficiaries less than 20 years of age, 50 percent to those aged 20 to 60, and 25 percent to those older than 60. Thailand eventually extended its coverage to dependents, but only after SHI accumulated a hefty surplus. 3. Accurate estimates of the benefits package and of costs determine the financial sustainability and survival of SHI. Actuarial costing of the ben- efits package is essential, with such techniques increasingly being applied. Actuarial methods require such information as household utilization data broken down by age, sex, employment, and income. Furthermore, utiliza- tion and unit costs will have to be projected to reflect moral hazard, induced demand, and price increases associated with the introduction of SHI. For example, health services of higher quality than in the past, without financial barriers to their use, are likely to lead to a large increase in primary care consultations. The contribution can only be determined when accurate actu- arial costs have been completed. Is a contribution of 6 percent of payroll, split evenly between employer and employee, enough? Is 9 percent enough or is a higher percentage called for? What level is politically acceptable and collect- ible? If required contributions are too high, as usually happens, what has to be cut back in the benefits package? All this means that the benefits package is determined by means of a cycle: designing the benefits package, estimat- ing its costs, undertaking political consultation, then adjusting the benefits package and estimating its costs, and so on. This process requires technical skills and data. In countries with a per capita annual income of US$2,000 or less, these problems will be a major challenge; in countries with a per capita annual income of US$500 or less, they will be monumental. In all countries, sound actuarial analysis combined with a political process will be essential. The Kenya case clearly illustrates what happens when this process is not fol- lowed. Many districts in Ghana encountered difficulty in implementing SHI because the premium was too low. 4. Enrollment of those in the informal sector or the self-employed will always be a major challenge, because mandatory enrollment is not easily enforce- able. If their enrollment is voluntary, adverse selection--meaning the poor and the sick with the highest medical bills will be the most likely to join-- will be a major threat to the financial sustainability of SHI. In addition, the administrative costs to enroll, monitor, and collect contributions from this population can also be high. In the Philippines, two-thirds of voluntary enrollees did not pay their premiums on a regular basis, motivating the SHI fund to give religious and cooperative organizations group discounts as a 164 Lessons Learned and Policy Implications way of enrolling their entire membership, which helped somewhat, but did not solve the problem. In Thailand, coverage and collection problems were such that the government decided to use general revenues to pay for all informal sector and self-employed workers. The more confidence workers have in government and SHI, the more recourse government and SHI have to effectively punish those who evade. 5. All stages of SHI face major problems in relation to defining, certifying, and subsidizing the poor. As implied by the Ghana and Kenya case studies, the poorer the country, the worse the problem, because the number of poor will be large and the capacity to monitor and evaluate them will be limited. Moreover, actually getting public subsidies to the poor and the indigent will be a major hurdle, requiring proper cross-subsidization and pooling of risks between rich and poor regions if SHI contributions are held at the regional or district level. 6. Supply will have to be built up progressively if clients in peri-urban and rural areas are to have access to adequate health care. Moreover, improving performance through contracting (on the supply side) and through choice of providers (on the demand side) will be compromised without sufficient pro- viders to allow some form of competition. 7. Provider payment mechanisms that aim to shift the financial risk of provi- sion to the provider, such as capitation, will have to be continuously moni- tored and evaluated. As Thailand's experience shows, capitation without special provisions for the indigent or for expensive cases can lead to cream skimming and risk selection, because providers have an incentive to keep their costs to a minimum. 8. Improvement of the administrative efficiency and effectiveness of SHI requires attention on several fronts. For example, how should a country con- solidate existing social insurance and other risk-pooling schemes? In Colom- bia and Ghana, several different insurance schemes already existed prior to the introduction of universal SHI, resulting in fragmented risk pools, inef- ficiencies, and large transaction costs. In Ghana, SHI faces the challenge of integrating various schemes that vary in terms of membership, benefits, premiums, and types of providers. In Colombia, the administrative costs of health plans were so variable, ranging from 4 to 60 percent of the value of the premium, that the government issued a decree that required plans to have a minimum of 200,000 enrollees. This led to a wave of mergers among plans serving the subsidized population, with a reduction in their numbers from 239 in 1999 to 43 by 2004, with 45 percent being private, 42 percent being community based, 6 percent being public, and the remainder being health plans for indigenous people. Moreover, administrative budgets have to be realistically determined and adhered to. In Kenya, these are set at 5 percent of SHI revenues, but are as high as 12 percent in the Philippines. Within those budgets, SHI administrations have to be more effective in discharging their responsibilities because they have little recourse to soft budgets and govern- ment bailouts should cost overruns occur. This places a premium on estab- lishing efficient administrative operations and recruiting good managers. 9. Leakage of SHI funds because of corruption will be a perpetual threat. In Kenya, compulsory SHI for hospital services suffered greatly from poor man- William C. Hsiao and R. Paul Shaw 165 agement and corruption, with only 22 percent of the fund actually used to pay for benefits and a large portion of the accumulated reserve lost through corruption. As a result, Kenya's plans to launch a new national SHI system include rules to constrain the board's behavior, including a 5 percent limit on administrative costs, a 3 percent limit on reserves, and a requirement for board members to abstain from voting on investments or contracts if they have any financial links to them. Fraudulent claims will hound SHI, as in the Philippines, where the Office of the Actuary estimates that somewhere between 10 to 20 percent of claims are fraudulent. Policy Implications In dozens of countries, special technical groups are advising ministers of health and ministers of finance on better policy making and capacity building for SHI. Many of these technical groups have participated in brainstorming meetings or dedicated learning events on SHI sponsored by the World Bank Institute and/or WHO. More informed assessment of the preconditions for SHI, as discussed in chapter 1, has motivated some groups to put their plans for SHI on hold. Others have decided to forge ahead, aware that huge challenges await. The policy implica- tions summarized here are cautionary statements, intended to minimize miscon- ceptions and mistakes. SHI Is Complicated: Effective and Efficient Implementation Takes Many Years SHI is a sophisticated financing method and effective implementation requires reorganizing the ministry of health and establishing a new SHI agency. The min- istry of health has to be transformed from being a funder, manager, and operator of public health services to being a policy maker, a regulator, and an overseer. The new SHI agency has to have a sound organizational structure, effective leadership and management, capable and dedicated professional staff, and sophisticated man- agement information and information technology systems. For example, the SHI agency has to recruit or contract capable doctors, actuaries, accountants, financial managers, information technology specialists, policy analysts, and planners and its executives and managers have to be educated in the nature and functions of SHI. In managing health services, the SHI agency needs to select and contract health care providers and monitor their services, which requires hospitals to develop modern accounting, financial, and medical record-keeping systems. Most low-income coun- tries simply do not have the required human resources and knowledge and these will have to be developed over time. For SHI to Achieve Universality Takes Decades Passing a law to introduce the principle of universal coverage through SHI is only the first step. In relatively rich countries, the number of years between the first law related to health insurance and the final law that effectively implemented universal coverage through SHI ranged from 70 to 100 or more years in several Western Euro- pean countries to 30 to 50 years for Costa Rica, Japan, and Korea. Moreover, cover- age tended to grow slowly, usually taking decades to approach universality. 166 Lessons Learned and Policy Implications The most severe constraints to achieving universality in low- and middle- income countries are tax revenues and the portion of workers employed in the formal sector, which depend on a country's stage of economic development and employment structure. Poor countries not only have small tax bases, but the tax bases are also narrow because of the countries' large informal sectors. Enrolling civil servants and employees of large firms is easy, but enrolling the self-employed and informal sector workers is much more difficult. The challenge for enrolling the poor is funding them. Economic growth can have four positive effects on SHI enrollment: (a) it lifts people out of poverty, meaning that more people can afford to pay; (b) it brings more workers into the formal sector, which increases the number of people in the contributory regime; (c) it raises general revenues for the government, meaning that more of the poor can be subsidized; and (d) it tends to increase the government's administrative capacity to collect taxes and insurance premiums. Thailand took more than 25 years to reach universality, at which point annual per capita income had reached US$2,400. Thus expanding SHI coverage and attaining universality depend on a nation's rate of economic development. The Government Should Recognize That Having the Same Benefits Package for All Groups May Not Be Possible Initially All countries would like to offer a comprehensive benefits package to all citizens. Unfortunately low- and middle-income countries may not be able to afford such a package. Formal sector employees demand comprehensive benefits and might be able to pay for it. However, a difficult trade-off has to be made for the poor and near-poor. The choice involves covering fewer poor and near-poor with a compre- hensive package or covering more of them with a less comprehensive package. For example, Colombia had to limit the benefits package of the poor to make it only half as expensive as that of formal sector employees so it can be affordable. Now Colombia is expanding its coverage to the near-poor with a benefits package less comprehensive than that of the poor. The Benefits Package Must Be Designed and Costed Costing out the benefits package is a major technical and political hurdle in imple- menting SHI. Usually, politicians like to promise the most and pay the least. A sound SHI program requires adequate and sustainable financing. The initial "wish- ful" benefits package has to be costed to ascertain its affordability and acceptability by the funders. In practice, it requires several rounds of designing a benefits pack- age, estimating its actuarial costs, modifying the benefits package, and then reesti- mating its actuarial cost to reach a balance between "wishes" and economic reality. Kenya did not conduct such a process, which resulted in the president's refusal to sign the SHI bill into law. Ghana has yet to cost its benefits package. The Philip- pines has not undertaken actuarial studies of the real costs of providing benefits packages, even after 10 years of SHI. Absence of this kind of information disem- powers managers. User Fees Must Be in Place to Motivate People to Join William C. Hsiao and R. Paul Shaw 167 People have to have incentives to pay SHI premiums, especially if employers cannot garnish their wages. This applies especially to self-employed workers, who would not be motivated to join SHI if government services were available at little or no cost. This is a significant issue in countries with high levels of poverty and strong political pressures to eliminate user chargers at public facilities. In other words, people will be motivated to pay for SHI if user fees are relatively high, if patients have to purchase drugs and supplies, or if public services are so poor that many patients pay out-of-pocket for private providers and are susceptible to catastrophic financial loss at times of serious illness or injury. SHI Must Create Adequate Incentives for Workers to Enroll The mainstay of SHI membership will be workers in the formal sector, who can be identified as employees of organizations and mandated to pay monthly premi- ums. However, even when contributions are mandatory, adverse selection can be a critical problem, with evasion most likely among smaller employers. To overcome this, SHI has to create incentives for workers to enroll by requiring employers to pay a share of the SHI premium. Thus Colombia, Ghana, Kenya, the Philippines, and Thailand all designed their SHI so that employers pay at least half of the pre- mium. This parallels the experience of SHI in many countries that are more devel- oped than these five, where the contribution rate is also split between employers and employees, for example, 4.5 percent of payroll each in Bulgaria and 6.1 percent each in Romania. Large General Revenues Are Needed to Cover the Poor All low- and middle-income countries have large poor populations who are unable to pay the premium, and nations use different criteria to establish the poverty level. The indigent population usually accounts for 40 to 50 percent of the population of low- income nations. Even in the United States, 12 percent of the population falls below the poverty line. The government has to have the budget to subsidize the poor, and to do this, it could reallocate funds from other programs to health and/or raise new tax revenues. For example, Ghana imposed a new 2.5 percent value added tax to help finance the subsidized regime, but whether the funds will be sufficient to pay for all the poor as Ghana's SHI coverage expands is not clear. The Philippines used revenues from a national sweepstakes lottery to help finance premiums for the poor. Stakeholders Must Be Convinced of the Actuarial Soundness of SHI SHI is required to maintain its own solvency, meaning expenditures "out" must not exceed revenue "in," and thus must be transparent and accountable. Solvency cannot be adequately assessed without actuarial calculations that consider near and longer-term characteristics of the workforce and the level of workers' earn- ings. These, in turn, will depend upon many economic and demographic factors, including future birth rates, death rates, labor force participation rates, economic development rates, and wage increases. In the United States, for example, actuarial calculations must show predicted revenues and expenditures for 25 years into the future for Medicare, a program financed primarily by a payroll tax that covers 38 million elderly and disabled Americans. The absence of actuarial studies will leave 168 Lessons Learned and Policy Implications SHI policies and implementation plans vulnerable to intense public scrutiny and criticism in relation to solvency. Supply-Side Subsidies Must Be Reduced In most low-income countries, the government subsidizes public health facilities with a large annual budget. Under SHI, public health facilities will receive their revenues from SHI payments. The supply-side subsidy should be reduced in syn- chronization with the implementation of SHI, with the savings used to expand the subsidy for the poor or the like. Otherwise, the public health facilities will be overpaid, a poor use of scarce resources. Colombia had planned to reduce this supply-side subsidy as SHI expanded, but it has been unable to do so, because of the political power of public health workers' unions. As a result, the planned expansion of SHI to cover the near-poor has been retarded and public health facil- ities used the additional revenues from SHI payments to increase staff compensa- tion and to undertake new capital projects. The SHI Agency Should Be Insulated from Political Interference To represent the interests of the insured and prevent corruption, SHI needs to be independent from the government. The new independent agency must be transpar- ent in relation to its finances, which requires independent audits. Many countries establish their SHI agency under the ministry of health without adequate represen- tation by the insured and by premium payers. Typically, the ministry of health is dominated by medical professionals who tend to protect supply-side interests, as was the case in Colombia and the Philippines. Under such governance, much of the new revenue went to increasing the salaries and profits of providers. The SHI Agency Should Be a Prudent Purchaser of Medical Services and Goods For the ministry of health to manage SHI is hard, because doing so requires a transfor- mation of its corporate culture from that of a funder and operator of public services to thatofanactive,prudentpurchaserofservicesfortheinsured.Thecontrastingexperi- ences ofthe Philippines and Thailand providesa good example. The SHI agencyinthe Philippines has acted like a traditional, passive, private insurance company, that is, as just a financial intermediary. It enrolls members, collects premiums, and pays claims. By contrast, the NHSO in Thailand selectively contracts with provider networks and pays them a capitation rate, a payment system designed to discourage overuse. As a result, Thailand has not had the same problems the Philippines has of new funds being used to benefit suppliers instead of the insured. Qualified Providers Must Be Certified Before or Concurrently with Implementation of SHI As a purchaser of services, SHI will need to contract with public, NGO, and pri- vate-for-profit providers. SHI members will expect these providers to provide more or less uniform quality of services and have more or less uniform capacity to deliver them. To a large degree, public providers will be trying to conform to estab- William C. Hsiao and R. Paul Shaw 169 lished, published, provider guidelines, with the onus on public providers to live up to those guidelines if contracted by SHI. At the same time, the quality of private sector health services is highly variable, and detecting fraud and price gouging is difficult when SHI pays claims. These kinds of deficiencies have to be remedied before or concurrently with the implementation of SHI if it is to gain enduring pub- lic support, perform its role of assuring reasonable quality, and sustain its opera- tions financially. ASingle Fund Is Preferable to Many Funds Three main arguments support the establishment of a single insurance fund instead of many funds. First, low-income countries lack the human resources, experience, and information technology systems to start even one fund properly, so having multiple funds would further dilute the human resource pool. Second, hav- ing multiple funds would increase administrative costs at both the insurer and pro- vider levels, meaning that fewer of the scarce resources would be spent for health services. Third, a system with multiple funds develops political and bureaucratic barriers to universal SHI with equal access. Thailand's major hurdle to equalizing access is to merge its various funds. Colombia, Ghana, Kenya, and the Philippines learned from the struggles of more advanced economies such as Germany, Korea, and Taiwan (China) that had multiple SHI funds, and established a single SHI fund. Nonetheless, several economists argue that many funds would give people a choice and that competition among funds would promote greater efficiency; however, no credible evidence to support this theory is available. Donors Could Play a Valuable Role in Supporting the Implementation of SHI Donors are keen to target international or national public subsidies to improve the health of the poor. Their traditional approach has been to channel donor assistance to publicly financed and provided health goods and services on the presumption that these funds actually benefit the poor. However, as studies of public funding typically show, households in the richest quintiles are the ones that typically ben- efit the most by "capturing" the public subsidy through having public hospitals built in urban areas where the rich can easily access them. In addition to multilat- eral and bilateral donors, international NGOs channel large infusions of funds to supply-side provision of public health goods and services, such as much needed vaccinations supported by the Global Alliance for Vaccines and Immunizations, which directly affect health status outcomes. Donor support of SHI is needed to complement traditional development assistance, because SHI shifts the emphasis from improving health status alone to improving both health status and financial risk protection. Avenues of support include the following: · using funds from debt forgiveness in developing countries that are under- taking SHI as a way to subsidize membership by the poor as Ghana is plan- ning to do; · providing direct grants to SHI that are earmarked for enrolling the poor and the indigent and/or expanding special benefit entitlements to targeted groups, as the Philippines did with DOTS and maternity benefits; 170 Lessons Learned and Policy Implications · increasing technical assistance to manage and operate SHI in the form of resident, seconded staff who have experience in planning, budgeting, under- taking actuarial analysis, contracting, and monitoring performance. SHI Should be Linked to a National Health Insurance Policy Finally, an important role of government is to formulate a policy that clearly links the initiation of SHI with other risk-pooling mechanisms and other dimensions of health systems development. First, SHI should not be seen merely as a way of mobilizing and earmarking more revenue for health and for providing formal sec- tor workers with better health services. It should be seen as a building block in national efforts to provide universal risk pooling and coverage and, where possible, it should be used to connect and integrate risk-pooling initiatives by other groups in society. The clearer government intentions in this area are, the better can differ- ent stakeholders assess current and future benefits of SHI to themselves, and the more solidarity can be facilitated in the process. Second, it is important to be clear that legislation in support of SHI, as well as laws to implement it, will have nationwide implications for all sources of revenue mobilized for health. Key issues requiring clarity are the extent to which the gov- ernment will continue to mobilize financing for health from general revenues and other taxes and how it plans to target these revenues more effectively to provide public health goods and services and subsidize the poor. In other words, as SHI evolves, how will it coexist with core government funding in a way that maximizes financial risk protection and assures access to quality care nationwide? Third, the potential or intended role of SHI as an instrument of health sector reform should be specified. SHI can be a powerful instrument for reforming the health care delivery system to improve the efficiency and quality of health care by being an active and prudent purchaser of health goods and services. SHI can also be an instrument for helping to horizontally integrate the many vertical externally funded programs typically found in low- and middle-income countries. Conclusions Theory on SHI is at its best when it speculates on what to do when designing SHI and at its worst when it advises on how to do it. The reason is simply that imple- menting SHI in economically, socially, and culturally diverse environments pro- duces unexpected challenges, thereby opening the door to new kinds of learning. The five case studies presented here are all about how to do it. Often the chal- lenges of implementation involve pitfalls, midcourse corrections, and accomplish- ments that fall short of expectations. Nevertheless, trying to answer the question posed at the outset of this monograph seems reasonable, namely, where might a low- or middle-income country expect to get to in designing and implementing SHI in, say, 10 years time? An informed answer must begin by "It depends . . ." If the country has a relatively low annual per capita income, say less than US$1,000, and a relatively small formal employment sector, it could succeed in enrolling a relatively small share of its population, roughly 10 to 20 percent, in a sus- tainable SHI initiative assuming (a) adequate planning and stakeholder consultation William C. Hsiao and R. Paul Shaw 171 during the design phase; (b) realistic costing of the benefits package and assurance that revenues from premiums will cover costs; (c) ability to identify, enroll, and col- lect premiums from employers and employees; (d) ability to contract with providers in sufficient proximity to SHI members; (e) capacity to regulate quality and handle grievances; (f) adequate human resources and administrative skills to manage the SHI fund; and (g) political commitment to initiate and expand SHI over time. In this case, SHI will obviously exist in parallel with other major methods of health care financing and provision that are probably dominated by the government. Important transition issues will arise, such as dealing with a two-tier system of health care in which SHI typically provides better quality, assuring that public funds for health that are freed up by SHI are better targeted to public health goods and services, helping the poor, and so on. Ghana is a good case to watch in this respect. If the country is lower middle-income, for example, annual per capita income of US$1,000 to US3,000 with 30 to 60 percent of its workers in the formal and/or urban sector, it could succeed in enrolling 20 to 40 percent of its population in an SHI initiative, assuming that the same criteria noted for the relatively low per capita income country are present. Additional prerequisites would be adequate provider markets and capacity for competition between public, private, and NGO providers. The Philippines is a good case to watch in this regard. If the country is middle-income, say annual per capita income of US$2,500 to US$6,000 with upward of 50 to 70 percent of its workers in the formal and/or urban sector, it could succeed in covering a majority of its population in a sustainable SHI program, assuming the foregoing criteria, strong political commitment to social solidarity, and fiscal capacity to cross-subsidize membership by informal sector workers. Colombia and Thailand are good cases to watch in this regard. If the country is higher middle-income, for instance, annual per capita income of more than US$6,000 with a track record of successful risk pooling among formal sector workers, risk-pooling initiatives for informal sector workers, and an unusual degree of political commitment, then achieving universal coverage through SHI would appear to be feasible. Note, however, that the foregoing generalizations and the findings summarized in this volume are part of an unfinished research agenda. Much remains to be eval- uated regarding design, implementation, impact on financial risk protection, and the costs and quality of health care. In this regard, we join others undertaking par- allel strands of operations research, including the World Bank (Preker and Carrin 2004; Schieber and Gottret 2006), WHO (Carrin and James 2005), the International Labour Organisation (Cichon and others 2000), and the German Agency for Techni- cal Cooperation (2004). To conclude, no magic bullet is available that will solve the woes of health care financing and provision in developing countries. SHI clearly has the potential to make a significant positive contribution, but success comes slowly, because of the drawbacks and risks involved. Theory on SHI provides only partial insights into the risks and bottlenecks involved. The real story comes from the trials and tribu- lations of implementation. Our hope is that by contributing to awareness of these issues, as experienced by several low- and middle-income countries, this mono- graph will help policy makers in developing nations chose the right approaches, and then, once chosen, to implement them correctly. 172 Lessons Learned and Policy Implications References Carrin, Guy, and Chris James. 2005. "Social Health Insurance: Key Factors Affecting the Transition Towards Universal Coverage." International Social Security Review 58 (1): 45­64. Cichon, Michael, H. Yamabana, D. Dror, and C. Normand. 2000. Modelling in Health Care Finance. Geneva: International Labor Office. Filmer, Deon, Jeffrey Hammer, and Lant Pritchett. 1998. "Health Policy in Poor Countries: Weak Links in the Chain." Policy Research Working Paper 1874, World Bank, Policy Research Department, Washington, DC. German Agency for Technical Cooperation. 2004. Social Health Insurance: Systems of Solidar- ity. Eschborn, Germany: Federal Ministry for Economic Cooperation and Development. Levine, Ruth. 2005. "From Aid Dependency to Dependable Aid," Global Future 2nd quarter: 7­8. Preker, Alex, and Guy Carrin, eds. 2004. Health Financing for Poor People. Washington, DC: World Bank. Schieber, George, and Pablo Gottret. 2006. Health Financing Revisited: A Practitioner's Guide. Washington, DC: World Bank. WHO (World Health Organization). 2000. World Health Report 1999: Health Systems-- Improving Performance. Geneva: WHO. "Social health insurance is clearly one -- albeit not the only -- mechanism to increase the overall resources envelop for health in developing nations on the basis of equity and solidarity. is book is a unique opportunity for governments and health planners to benefit from the experience of a fairly representative set of developing countries that have introduced and are operating social health insurance. e authors, both outstanding experts with the unique combination of an excellent academic background and solid hands-on experience in health insurance, also offer a solid, well balanced analysis of the necessary prerequisites, the positive potential but also the policy challenges that nations will experience when introducing SHI. Required reading for health system planners not only in developing countries." Michael Cichon, Director, Social Protection Sector International Labour Office, Geneva "Building equitable health financingsystems in developing nations is imperative for improving the health and welfare of the socially disadvantaged. By focusing on one of theprimarymethodsof healthcarefinancinganddrawingonasetof in-depthcountry case studies, this edited volume provides pearls of practical wisdom on the promise and pitfalls of social health insurance." Timothy Evans, M.D., Ph.D., Assistant Director-General for Information, Evidence, and Research World Health Organization, Geneva TMxHSKIMBy369494zv":<:*:/:* THE WORLD BANK ISBN 0-8213-6949-0