A GUIDE TO COMPETITIVE VOUCHERS INHEALTH A GUIDE TO COMPETITIVE VOUCHERS IN HEALTH A GUIDE TO COMPETITIVE VOUCHERS IN HEALTH © 2005 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 07 06 05 04 The findings, interpretations, and conclusions expressed herein are those of the au- thor(s) and do not necessarily reflect the views of the Board of 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 judgment 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 work is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The World Bank encourages dissemination of its work and will normally grant permission 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, www.copyright.com. All other queries on rights and licenses, including subsidiary rights, should be ad- dressed to the Office of the Publisher, World Bank, 1818 H Street NW, Washington, DC 20433, USA, fax 202-522-2422, e-mail pubrights@worldbank.org. Library of Congress Cataloging-in-Publications Data A guide to competitive vouchers in health. p. cm. Includes bibliographical references and index. ISBN 0-8213-5855-3 1. Medical care, Cost of--United States. 2. Medical care--Finance--United States. I. World Bank. RA410.5.G855 2004 338.4'33521'0973--dc22 TABLE OF CONTENTS Foreword ix Preface xi PART I WHAT ARE HEALTH CARE VOUCHERS AND 1 HOW DO THEY WORK? Chapter 1 Delivering Public Subsidies in Health Care 3 Why Do Government Subsidies Do So Little to Help the Poor? 4 When Are Subsidies for Health Services Justified? 5 Demand-Side versus Supply-Side Subsidies 6 Other Factors Influencing the Effectiveness of Subsidies 13 Chapter 2 Benefits of Competitive Voucher Schemes 17 for Health How Does a Voucher Scheme Work? 18 What Are the Advantages of Health Vouchers? 19 PART II INTRODUCING A VOUCHER SCHEME 23 Chapter 3 Conducting Prefeasibility Studies 25 Identifying the Health Sector Problem and the Aims of the Voucher Scheme 26 Justifying the Use of a Voucher Scheme 26 vi | A Guide to Competitive Vouchers in Health Funding the Scheme 27 Understanding the Context in Which Subsidies 29 Will Be Introduced Building Partnerships 29 Checklist for Prefeasibility Assessment 31 Chapter 4 Conducting a Feasibility Assessment 33 Identifying Potential Impediments 34 Identifying Potential Advantages 40 Determining Whether a Voucher Scheme Is Financially Feasible 41 Weighing the Potential Benefits and Obstacles of a Voucher System 42 Checklist for Feasibility Assessment 42 Chapter 5 Designing a Voucher Scheme 45 Choosing a Voucher Agency 46 Establishing Recipient Policies 48 Determining Benefit Policies 54 Determining Price Policies 57 Determining Provider Policies 58 Determining Reimbursement Value Policies 62 Designing the Vouchers and Other Materials 63 Transport and Communications 66 Distributing the Vouchers 66 Payment of Service Providers 68 Developing Information Systems 68 Checklist for Voucher Design 72 Chapter 6 Implementing a Voucher Scheme 73 Tendering for Service Providers 73 Negotiating Contracts 74 Contracting 77 Training 77 Piloting 82 Institutionalizing 84 Scaling Up 84 Contents | vii Chapter 7 Monitoring and Evaluating a Voucher Program 87 Monitoring Service Quality 88 Monitoring Competition among Providers 91 Monitoring to Detect Abuse of the Voucher Scheme 92 Monitoring the Characteristics of Voucher Recipients and Redeemers 96 Monitoring and Evaluating Health Outcomes 97 Monitoring and Evaluating the Impact on Equity and Poverty Reduction 98 Monitoring and Evaluating Cost-Effectiveness 100 References 101 Index 109 Boxes Box 2-1 Varying the Service Package According to the Needs of the Patient in Nicaragua 21 Box 3-1 Prefeasibility Work on a Voucher Scheme for 31 Cervical Cancer Screening Box 4-1 Political Impediments to Voucher Schemes in El Salvador and Honduras 37 Box 4-2 Are Vouchers Stigmatizing? 39 Box 4-3 Assessing the Feasibility of a Tuberculosis Voucher Scheme for Private Practitioners 43 Box 5-1 Who Runs the Voucher Program? 47 Box 5-2 How Can Vouchers Be Made Nontransferable? 53 Box 5-3 Benefit Policies: The Importance of Stating the Limits 57 Box 5-4 Price Policies Used by Different Voucher Schemes 58 Box 5-5 Choosing Service Providers that Meet Program Needs 62 Box 5-6 Designing the Voucher 65 Box 5-7 Facing the Challenges of Transport and Communications 67 Box 5-8 Using Village Leaders to Distribute Vouchers 68 Box 6-1 Tips on Negotiating 75 Box 6-2 Elements in a Model Contract 78 Box 6-3 How Piloting Can Lead to Changes in Strategy 83 viii | A Guide to Competitive Vouchers in Health Box 7-1 Calculating the HHIC 92 Box 7-2 Improving Targeting through Adverse Selection 96 Figures Figure 1-1 Demand-Side Versus Supply-Side Subsidies in Health Care 7 Figure 2-1 How Does a Voucher Scheme Work? 18 Figure 3-1 Decision Tree for Supply-Side Subsidies 28 Figure 3-2 Decision Tree for Demand-Side Subsidies 30 Figure 4-1 Typical Sequence of Activities in a Feasibility Assessment 35 Figure 5-1 Designing Recipient Policies 49 Figure 5-2 Designing Provider Policies 60 Figure 7-1 Monitoring and Evaluating the Voucher Program 89 Tables Table 3-1 Checklist for Prefeasability Assessment 31 Table 4-1 Checklist for Feasibility Assessment 42 Table 5-1 Matching Beneficiaries with Program Aims 50 Table 5-2 Benefit Policies of Selected Voucher Schemes for Health 55 Table 5-3 Checklist for Voucher Design 72 FOREWORD The World Development Report 2004 reviews traditional approaches to pub- lic service delivery and discusses how they have often failed the poor. Whether in health, education, or infrastructure, supply-side subsidy strate- gies to fund inputs--such as staff costs, equipment, and buildings used in delivery--have not improved the access to quality services among the poor. An important question remains for developing countries and the in- ternational development community on how to deliver and target public subsidies in ways that promote efficiency and innovation, increase ac- countability for performance, and leverage public resources with private participation and financing. The output-based aid (OBA) approach seeks to partially answer this question by using a demand-side subsidy delivery strategy. An OBA scheme contracts a private party to deliver services and makes disburse- ix x | A Guide to Competitive Vouchers in Health ment of the public funding contingent upon actual services delivered. Vouchers are one type of OBA approach. They have the potential to target specific segments of the population effectively, stimulate both supply and demand for under-supplied services, and establish a relatively straightfor- ward monitoring mechanism. When the voucher scheme is built on the principle of competition, it can not only further empower clients by al- lowing them to bring their business to providers of their choice but also give incentives for service providers to be innovative, cost effective, and re- sponsive to the clients. Competitive vouchers, however, are one of many types of demand-side subsidy strategies, and only a partial answer to the forthcoming challenges in the health sector. Design and implementation arrangements can affect the effectiveness of the vouchers scheme significantly. This guide aims at providing policymakers and donors with the tools needed to determine the appropriateness of competitive vouchers, as well as information on the design, execution, and monitoring of projects under this type of scheme. We look forward to the use of this guide as a way to facilitate the decisionmaking process regarding alternative options for the delivery of public health services to the poor. Ellis J. Juan Manager Infrastructure Advisory Services Infrastructure Economics and Finance Department July 2004 PREFACE This guide identifies the advantages of competitive voucher schemes in delivering subsidies; describes the circumstances under which they are su- perior to other subsidy mechanisms; and explains how to design, imple- ment, monitor, and evaluate a voucher scheme. It provides a broad out- line of the problems faced by health systems, the rationale for government intervention, and the different ways in which governments and donors subsidize health care. The guide does not advocate greater use of vouchers but simply raises awareness about voucher schemes, offering policymakers guidance on the choices and decisions they need to make. It also highlights some pitfalls of voucher schemes and describes the different formats vouchers can take depending on the health problems being addressed and the objectives policymakers wish to achieve. xi ACKNOWLEDGMENTS This guide was prepared by a team led by Chiaki Yamamoto and Jeff Ruster. Peter Sandiford, Anna Gorter, Micol Salvetto, and Zil Rojas of Instituto Centroamericano de la Salud (ICAS) prepared the guide. Kathy Khuu and Kwadjo Asante also worked on this guide, and Rosario Bartolome provided invaluable administrative support. The team greatly benefited from comments and advice from Abdo Yazbeck, as well as participants of the workshop Competitive Vouchers for Health Care Service Delivery held at the World Bank on April 22, 2004. The team is grateful to Ellis Juan, the manager of Infrastructure Advisory Services, and Hossain Razavi, the director of Infrastructure Economics and Finance, for their support and guidance. PART I WHAT ARE HEALTH CARE VOUCHERS AND HOW DO THEY WORK? CHAPTER 1 DELIVERING PUBLIC SUBSIDIES IN HEALTH CARE Health systems in developing countries face enormous problems. In all of the poorest countries--and even in most rich ones--health outcomes vary widely by socioeconomic group. Governments and insurance companies alike are struggling to meet the costs of ever-increasing public expectations for health services. In many countries, households in which a family mem- ber suffers from a chronic disease are driven into (or kept in) poverty by the catastrophic cost of ongoing medical care. At the same time, vast sums are being wasted on ineffective--or even harmful--interventions, and enormous technical inefficiencies plague the delivery of health services. Worst of all, there is good evidence that public subsidies in health are fail- ing to reach their main intended beneficiaries--the poor and vulnerable. 3 4 | A Guide to Competitive Vouchers in Health WHY DO GOVERNMENT SUBSIDIES DO SO LITTLE TO HELP THE POOR? Government intervention in the health sector has typically been supply- side subsidies delivered through a network of publicly owned and operat- ed health facilities. Some of these facilities serve the entire population; others cover only those unable to afford health insurance and not covered by social security. In many countries, these services have succeeded in re- ducing infant and maternal mortality. But few governments in developing countries have raised sufficient revenue to provide the range of services that meets the public's expectations, and government regulations have re- sulted in serious allocative inefficiencies, with staff often taking prece- dence over equipment and drugs. In the poorest countries, the "public" system is really a mixture of pub- licly funded staff and consumables funded privately through out-of-pock- et spending by patients. In many countries, staff also expect informal "fees" in return for access to health care. As a result, the cost-effectiveness of service delivery is low, and services are consumed by relatively well-off patients with less urgent health needs, undermining both the efficiency and equity of the health system. Supply-side subsidies, which cover some or all of the costs of health services inputs (infrastructure, staff, drugs, equipment, nonmedical con- sumables), provide little incentive to attract patients or increase produc- tivity. As a result, despite relatively low wages, publicly operated services have remarkably high unit costs, and utilization rates are often low. The absence of targeting (restricting benefits to a certain subset of the popula- tion) greatly dilutes the impact of public expenditure on health care. Mid- dle-class people pay less than they can afford, while the poor often pay more. The perception--and reality--of low quality in the public sector allows the private sector to flourish in developing countries. As a result, many countries see an abundance of private providers, not all of whom provide high-quality services, while the population, particularly the poor, under- consume public services. Delivering Public Subsidies in Health Care | 5 WHEN ARE SUBSIDIES FOR HEALTH SERVICES JUSTIFIED? Three factors make subsidizing health care desirable: 1. Inequitable distribution of wealth and health. Society views some redistrib- ution as fair and desirable, particularly to alleviate or eliminate poverty and to give all people the opportunity to enjoy a reasonable standard of health. Health (but not necessarily health care) is generally consid- ered a human right. In the absence of subsidies, as long as the distribu- tion of wealth is inequitable, the distribution of access to health care-- and therefore of health status--will also be inequitable. Skewed distributions of wealth and health are not only unfair in their own right, they also create conditions in which catastrophic health care costs can drive some people into poverty or force them to forgo the health care they need. Catastrophic health care costs can be reduced by risk- sharing strategies, such as health insurance and social security, but even in industrial countries the markets for these services operate imperfect- ly because of asymmetric information. As a result, many people end up falling through the safety net. 2. Presence of externalities. Many health interventions--vaccinations, treat- ment of communicable diseases, removal of vector breeding sites-- benefit not only the person being treated but others as well. Family planning is also associated with positive externalities, or spillover ef- fects, since limiting the number of children reduces a variety of social costs associated with excessive population growth. 3. Incomplete information. Individuals lack knowledge about their future health care needs and their costs. This uncertainty creates a market for pooling risks and sharing costs through health insurance. Because indi- viduals have some information on their likely future health care costs, however, there is an incentive for those with the highest expected needs to purchase more health insurance than those who expect to have rela- tively low expenditures. This problem, known as adverse selection, combined with a tendency to overconsume services that are paid for by a third party (a phenomenon known as moral hazard) can lead to com- plete collapse of the health insurance market or to unaffordable premi- ums for a large section of the population, exposing them to the risk of catastrophic health expenditures. Subsidies for health education and regulation may help address this source of market failure. 6 | A Guide to Competitive Vouchers in Health DEMAND-SIDE VERSUS SUPPLY-SIDE SUBSIDIES Subsidies can be divided into two main groups: supply-side subsidies and demand-side subsidies. Supply-side subsidies are linked to inputs; de- mand-side subsidies are linked to outputs (figure 1-1). Each type of sub- sidy has advantages and disadvantages. The two can be used in combina- tion to take advantage of the benefits each has to offer. Supply-Side Subsidies Supply-side subsidies cover some or all of the costs of the inputs to health services. They fall into two broad categories: cash subsidies and in-kind subsidies. Cash subsidies may or may not specify the inputs they subsidize. They include lump-sum payments and block contracts to provide a set of serv- ices; tax rebates (on the construction of health facilities in underserved ar- eas, for example); and capitated payments based solely on a catchment population. (A capitation payment that depends on the number of pa- tients actually using a provider regularly or a system in which patients lose access to service when they shift to a different provider is a demand-side subsidy, as it is linked to the output of service utilization.) The publicly owned and operated national health systems of many countries are examples of in-kind supply-side subsidies. In-kind subsidies are often provided for a more limited range of goods and services, includ- ing drugs, donations or loans of premises for health facilities, training, and payment of staff salaries. Supply-side subsidies are usually relatively simple to introduce and in- expensive to administer, and they can provide benefits to broad popula- tion groups. They are appropriate where the subsidized good or service can be used only by the target groups. Examples include immunizations; drugs used to treat communicable diseases, such as tuberculosis, and health facility infrastructure and staff costs in poor areas. There are several disadvantages to providing supply-side subsidies: · Difficulty targeting. There is no guarantee that supply-side subsidies benefit those for whom they are intended. One way of restricting ben- efits to target groups is to subsidize the providers they use, but target populations may not use the facilities that receive assistance. Another option is to subsidize inputs that can be used only for specific health Delivering Public Subsidies in Health Care | 7 Figure 1-1. Demand-Side Versus Supply-Side Subsidies in Health Care SUBSIDIES E.g., tax revenue or donation PROVIDER ORGANIZATION PAYMENT ORGANIZATION E.g., Ministry of Health, E.g., voucher agency social security, other. RIGHT TO INPUTS SUBSIDY E.g., salaries, E.g., vouchers, drugs, capitation equipment, payment, fee etc. subsidies Payments PROVIDERS USERS Invoice for subsidies on goods and/or Free or services Redemption of Co-payments Co-payments subsidized the right for services subsidy USERS PROVIDERS E.g., patients SUPPLY-SIDE FINANCING DEMAND-SIDE FINANCING problems, such as immunizations and certain drugs for treating tuber- culosis. The scope of this type of subsidy is limited, however, and may not be sufficient to achieve the desired health system objectives. · Lack of patient empowerment. Supply-side subsidies are often associated with low-quality service. Because assistance flows from the government to the provider rather than to patients, supply-side subsidies create no incentives for service providers to provide good service or offer patients anything beyond the bare essentials in terms of comfort and privacy. 8 | A Guide to Competitive Vouchers in Health · Lack of incentives to improve efficiency. Since supply-side subsidies do not normally link payment to the provision of service, they may be dissi- pated in salary increases and inefficiency rather than used to improve the quality and increase the quantity of services provided. Demand-Side Subsidies The key defining feature of a demand-side subsidy is the direct link be- tween the intended beneficiary, the subsidy, and the desired output (such as access or utilization). The level of funding received by the provider therefore depends on the outputs produced. Demand-side subsidies can be consumer led or provider led. They can be provided before or after service utilization. Consumer-Led Demand-Side Subsidies Provided before Service Utilization Consumer-led demand-side subsidies transferred before service utiliza- tion include cash transfers to patients, contributions to or tax rebates on family medical savings schemes, and vouchers. Features of these are: · Cash transfers provided to potential patients to pay for health care risk being used for other purposes, such as buying food, unless urgent med- ical care is required. · Contributions to or tax rebates on family medical savings schemes re- quire families to deposit a certain percentage of their earnings into ac- counts, which can be used only to pay for medical expenses. (An ex- ample is the Medisave program in Singapore.) Making such deposits tax deductible is one way of subsidizing them. Governments can also make deposits on behalf of the poor. · Vouchers are not always consumer-led demand subsidies (they can also be held by providers). Chapter 2 describes competitive voucher schemes as a form of demand-side subsidies. Consumer-Led Demand-Side Subsidies Provided after Service Utilization Cash refunds to patients represent consumer-led demand-side subsidies that are transferred after service utilization. This type of subsidy is com- monly used by insurance companies, but it could also be used for public subsidies in a well-developed health system. The major concerns with such a subsidy scheme are the need to ensure that claims for refunds are Delivering Public Subsidies in Health Care | 9 legitimate (that is, that service was actually provided) and to avoid moral hazard. The moral hazard problem is reduced if the refund covers only part of the expenses. Provider-Led Demand-Side Subsidies Provided before Service Provision Provider-led demand-side subsidies transferred before service provision include cost-per-case contracts in which the provider receives a fixed sub- sidy for a specified number of services, capitation payments, and referral vouchers distributed by providers that entitle the recipient to goods or services provided by others. · Cost-per-case contracts are normally channelled by third-party pur- chasers, such as district or regional health authorities. In a variant of this mechanism, the cost-per-volume contract, a minimum volume of service is purchased, with the cost changing as the volume increases. · Under capitation a provider receives a subsidy for providing particular individuals with access to care. Capitation payments are demand-side subsidies only if they are tied directly to the number of patients actually served by the provider--as they are in the United Kingdom, where fund- ing follows patients if they change general practitioners. Capitation pay- ments made to providers for covering the residents of a particular zone, irrespective of whether they use a particular provider, are supply-side subsidies. Capitation can be made as payment for providers' own costs and expenses or as payment for providers' purchase of services on behalf of their patients. In the now obsolete British model, general practition- ers received a fixed amount of money per patient (weighted by age, gen- der, and other characteristics), which they used to purchase external services, such as elective surgery, on behalf of their patients. · Referral vouchers can enable providers such as general practitioners to act as gatekeepers for public subsidies, ensuring that the subsidies are directed to the poorest individuals or those most in need of health care. Limiting the number of vouchers received by each provider can control the total volume of subsidies. Making the number of vouchers received by each provider dependent on the number of target patients reached can create an incentive for careful rationing of these subsidies. Such a system makes it possible for medical professionals to retain some of their discretionary power to consider the merits of individual cases, but allowing them to do so also introduces some risk of abuse. 10 | A Guide to Competitive Vouchers in Health Provider-Led Demand Subsidies Provided after Service Provision Provider-led demand subsidies transferred after service provision include fee-for-service subsidy claims and target payments. Under a fee-for-serv- ice subsidy, the provider receives a subsidy from the government for hav- ing provided eligible services to eligible individuals. An example is the New Zealand General Medical Subsidy, in which general practitioners receive a payment from the government for each child consultation they give. Receipt of the subsidy may be made conditional on the provider limiting or eliminating the fee paid by the patient. This form of provider subsidy may be administratively simpler than a voucher scheme, but it can be more difficult to control, as an independent mechanism is required to verify that the service was actually provided. Fee-for-service subsidies have also been criticized for leading to too little service among subsidized groups. They do provide a strong incentive to increase pro- ductivity, as the provider's subsidy income is directly related to the rate at which the services are provided. Target payments are made to providers who achieve certain predefined service targets. In the United Kingdom, general practitioners receive subsi- dies conditional upon achieving specified coverage levels for immuniza- tion. The subsidy is usually not based on individual patients but on some population target, such as vaccination or screening coverage rates. Subsi- dies could also be based on a health outcome among a particular popula- tion, such as the cure rate for tuberculosis patients. The key to successful target payments is the ability to independently (and cheaply) verify their achievement. Advantages and Disadvantages of Demand-Side Subsidies Demand-side subsidies have several advantages: · Output-based remuneration. By tying the receipt of subsidy to the gener- ation of some output, demand-side subsidies create incentives to in- crease those outputs and therefore raise productivity. · Evidence-based practice. The link between subsidies and outputs creates an opportunity to specify what the outputs will be. This is important for two reasons. First, it enables the subsidies to be used explicitly to address health problems for which a solid justification exists. Second, it allows policymakers to specify interventions that are known to be evi- dence based and cost effective. Many health care services have little im- Delivering Public Subsidies in Health Care | 11 pact on health. They are provided partly because of incomplete infor- mation (patients often do not know if an intervention is effective), part- ly because of moral hazard, and partly because of the agency problem-- the fact that patients' ignorance of the impact of health interventions can lead them to allow providers with financial incentives for delivering services to make decisions on their behalf. Demand-side subsidies al- low the agency role to be partially assumed by the donor to ensure that patients receive evidence-based, cost-effective care. · Targeting. Targeting can greatly improve the ability to meet the equity and poverty-reduction objectives of subsidies. It can also increase the consumption of services associated with positive externalities. Al- though targeting can be achieved indirectly with supply-side subsidies, it is possible to directly target individuals only when receipt of the sub- sidy is linked to provision of the service. Targeted consumer-led de- mand subsidies, such as vouchers, are one of the few instruments that allow health planners any degree of certainty that their subsidies are reaching the intended population groups. Perhaps more important, vouchers enable health planners to know who has not been reached, provided that good records are kept on who receives the vouchers and who uses them. This form of targeting is particularly valuable when try- ing to extend services to populations that are difficult to reach (sex workers, drug addicts, indigenous populations). Receiving vouchers in- tended specifically for them may make these groups feel that the serv- ices they receive will cater to their needs. By doing so, they may increase usage. Care must be taken, however, to ensure that targeting does not lead to stigmatization. · Output-based monitoring and evaluation. Monitoring is often thought of as a mundane activity carried out mainly to stop people (including staff) from abusing the system. For traditional supply-side assistance, it often comes down to checking whether or not certain items have been purchased and distributed and whether or not planned activities have been carried out. Monitoring may involve looking at whether the investments are being put to their expected use, but it rarely examines what happens to individuals as a result of assistance. With demand- side subsidies, it is possible to trace exactly who receives services, what services they receive, and even what outcomes are achieved. This in- formation makes it possible to evaluate the system. Being able to iden- 12 | A Guide to Competitive Vouchers in Health tify the fruits of an investment is something that politicians--who must approve subsidies--like very much. The public also likes evalua- tions, which show tangible benefit from their tax contribution. Demand-side subsidies are not without their drawbacks, however: · Higher transactions and administrative costs, because of the need to quantify outputs. Transactions and administrative costs--including the costs of producing and distributing vouchers, contracting providers and moni- toring their performance, reimbursing providers, and establishing sys- tems to avoid abuse of the voucher scheme--can be substantial. · Overservicing. Overservicing can occur because of the direct link be- tween outputs and subsidies, combined with moral hazard and agency problems. · Cream-skimming. Cream-skimming occurs when providers avoid pro- viding care to groups that require more services than others. As under capitation payments and health insurance subsidies, this problem can arise if the subsidy is for providing access to services rather than their utilization. · Lower patient satisfaction. Capitation subsidies create no incentives for providers to make services convenient or comfortable for patients. They can encourage providers to take on more patients than they can adequately handle and to treat them with the bare minimum of con- sumables. The greatest disadvantage of demand-side subsidies is probably the high- er transaction and administrative costs. The choice between supply-side and demand-side subsidies often therefore boils down to whether or not the expected benefits of higher productivity and quality outweigh the higher overhead costs. OTHER FACTORS INFLUENCING THE EFFECTIVENESS OF SUBSIDIES Three other factors--competition, contracting, and corruption--also af- fect the effectiveness of health care subsidies. Delivering Public Subsidies in Health Care | 13 Competition Subsidies are justified because a free market for health care services fails to address such issues as externalities and the distribution of health benefits across different income groups. Subsidies, however, often create other problems--including problems that are even more serious than those they were intended to solve. The common practice of restricting subsidies to publicly owned and operated facilities impairs competition in the mar- ket for health services. This loss of competition alters incentives in a way that often undermines efficiency, quality, and patient satisfaction. Preserving choice and competition in subsidy schemes has the addi- tional advantage of not leading to, or entrenching, parallel or duplicated health infrastructure. Subsidies make use of existing health services, re- gardless of who owns them. In the main cities of many developing coun- tries, the government, the social security institute, and the private sector each own separate hospitals and clinics, all serving the same geographic areas. Competition for subsidies makes it possible to contract services from any of these providers, based on which offers the best prices, quali- ty, or both. Competition for subsidies prevents them from distorting the market for health services in the same way that noncompetitive subsidies do. Competitive subsidies do not prevent new providers from establish- ing services (by unfairly subsidizing others). In fact, by making such investments more profitable, they can encourage the establishment of new facilities in previously underserved areas or the development of new services to address health problems that had been neglected. Demand- side subsidies also do not prevent inefficient providers whom few patients choose from closing down. Finally, demand side subsidies open the public sector, where efficiency and quality are often lowest, up to competition. Do different types of benefits differ in terms of the benefits they yield? Consider the following scenarios: · Subsidies are given in the absence of competition (for example, to a single public sector provider). This is the benchmark scenario, against which the others are compared. For supply-side subsidies, improving servic- es to attract more patients does not alter the amount of subsidy re- ceived. The subsidies therefore create little or no incentive to attract more patients by improving patient satisfaction or lowering prices. For demand-side subsidies, attracting more patients generates a propor- tionate increase in revenue. In the absence of competition, then, 14 | A Guide to Competitive Vouchers in Health demand-side subsidies provide incentives to attract patients, which supply-side subsidies do not. · Subsidies are given to all providers competing in a market. Under this sce- nario, patients have a choice of providers, and a level playing field is established. The competitive environment will create incentives to in- crease efficiency and improve patients' perceptions of quality. There is no financial incentive to improve quality that cannot be perceived by potential patients. Thus, although competition creates an incentive to improve efficiency and patient-perceived quality, supply-side--but not demand-side--subsidies have a tendency to undermine this incentive, particularly if they are large in relation to total provider revenue. · Subsidies are given only to accredited providers. Governments may provide subsidies only to providers who can demonstrate that they meet certain quality standards. This limits choice for patients, but if successful it can ensure that they receive better service. The benefits of including some form of accreditation with subsidies accrue directly and indirectly-- directly by ensuring that the subsidized services meet certain standards, indirectly by creating an incentive for providers to raise their service standards to meet the requirements for accreditation. As under the sec- ond scenario, supply-side--but not demand-side--subsidies have a tendency to undermine the incentive to improve efficiency and patient- perceived quality, particularly if they are large in relation to total provider revenue. · Providers must compete for limited places in a subsidy scheme. This scenario goes a step beyond accreditation, effectively introducing an additional competitive process. Limiting places in a subsidy scheme limits choice for patients, but in many cases it lowers administrative and monitoring costs. This form of competitive subsidy may be used to avoid duplica- tion in the supply of services that require high initial investments and services for which the number of potential beneficiaries is relatively low. Like accreditation, competition for limited places in a subsidy scheme can be used to improve quality in areas that patients are unable to perceive. It can also drive down costs. With supply-side subsidies, there is probably a stronger incentive to compete for places in the scheme than with demand-side subsidies, because once selected the provider is sure that it will receive the subsidy (with demand-side sub- Delivering Public Subsidies in Health Care | 15 sidies, the provider must still compete to attract patients eligible for re- ceiving the subsidy). However, that additional level of competition may be important for maximizing efficiency and patient satisfaction. Even when patient satisfaction is not closely correlated with technical quality, it may still be worth increasing. If patient satisfaction is low, uti- lization of services will be low, reducing the ability of a scheme to meet its objectives. Contracting Contracting can be hugely beneficial, because it forces each party to a con- tract to specify its expectations. Contracts thus force health planners to consider precisely which services they want from providers and to identi- fy them through detailed patient management protocols and quality spec- ifications. At the same time, the contracting process forces providers to un- derstand purchasers' priorities. This helps them reconfigure the care they provide based on purchasers' demands. Contracts can be more explicit about desired outputs and quality of care specifications under demand- side subsidies than under supply-side subsidies. Corruption Leakage--the loss of program funds to people other than the intended beneficiaries--through corruption and theft occurs under both demand- side and supply-side subsidies, but the form that it takes can be quite dif- ferent under the two types of schemes. Vouchers are susceptible to coun- terfeiting and black markets. Fee-for-service and capitation subsidies are vulnerable to submission of fraudulent claims. Supply-side subsidies can be diverted through pilfering, reselling, or using donated inputs (includ- ing staff time) for patients other than intended beneficiaries. Many sub- sidy schemes are susceptible to providers distorting or falsifying the infor- mation they send to donors in order to receive larger allocations than they are entitled to.1 Cross-subsidies--in which some individuals effectively subsidize others, through differen- 1 tial pricing of services, for example--represent a third group of subsidies. They are not con- sidered here. CHAPTER 2 BENEFITS OF COMPETITIVE VOUCHER SCHEMES FOR HEALTH A voucher is a token that can be used in exchange for a restricted range of goods or services. Vouchers tie the receipt of cash to particular goods, pro- vided by particular vendors, at particular times. Health care vouchers are used in exchange for health services (such as medical consultations or lab- oratory tests) or health care consumables (such as drugs). This guide examines voucher schemes as means of subsidizing health care goods and services. It focuses on schemes that involve some form of competition between providers of health goods and services, that provide the bearers of the vouchers with choices, and that involve the private sec- tor. It does not address voucher schemes in which the voucher can be ex- changed for cash as an incentive to use health services. 17 18 | A Guide to Competitive Vouchers in Health HOW DOES A VOUCHER SCHEME WORK? A typical competitive voucher scheme works in the way shown in figure 2-1. For voucher schemes distributing public subsidies, the process begins with the transfer of funds to a voucher agency (1). Vouchers are then pro- duced by a voucher agency and distributed to a target population, either by the agency itself (2a) or by third-party organizations (2b), which in turn distribute them to sections of the target population with which they have close links (2c). Voucher recipients take the vouchers to a health service provider of their choice (3) and exchange them for goods or services (or use Figure 2-1. How Does a Voucher Scheme Work? Government or Donor Agency 1 6 Voucher Agency 2b 4 Voucher redeemed 5 Provider payment Service Voucher Providers 2a Distributor 2c Voucher redemption 3 Voucher Recipients = Vouchers = Cash ($) = Reports Benefits of Competitive Voucher Schemes for Health | 19 them as partial payment for them).1 The service providers return the vouch- ers to the voucher agency (4), along with any other information it requires. The agency then pays the provider a sum, agreed on in advance, for each voucher returned (5). The voucher agency reports the program outputs and outcomes back to the government or donor providing the subsidies (6). WHAT ARE THE ADVANTAGES OF HEALTH VOUCHERS? Health vouchers are a specific type of demand-side subsidy. They have cer- tain advantages over other types of demand-side subsidies, but they often entail higher transactions costs. In deciding whether to use a voucher scheme to deliver subsidies, policymakers must determine whether the ability to deliver subsidies more efficiently or effectively outweighs the ad- ditional administrative costs. They must also consider some of the limita- tions of vouchers. Vouchers Allow Targeting of Beneficiaries Voucher schemes have greater scope for targeting than other demand-side subsidies. Targeting helps policymakers reach a higher proportion of the people they want to subsidize--a feature known as sensitivity. Targeting also helps exclude people who are not in the target group--a feature known as specificity. In many subsidy schemes, it is often difficult at the level of the health facility to separate those whom one wants to assist from those whom one does not. As a result, sensitivity and specificity are low. Voucher schemes can sometimes improve sensitivity and specificity, because the vouchers can be handed directly to target group individuals in the com- munity, where they can often be more reliably identified. This is particu- larly true when intended beneficiaries operate outside the law (for exam- ple, illicit drug users and, in some countries, commercial sex workers and men who have sex with men) or fear stigmatization (for example, people with tuberculosis, leprosy, and HIV/AIDS). Community targeting is also useful when individuals outside the target population can falsely claim they qualify for benefits (for example, nonpoor who receive exemptions from user charges intended only for the poor). Health service providers include clinics, informal practitioners, hospitals, laboratories or 1 other diagnostic services, pharmacies, community care service providers, health promot- ers, ambulances or other transport service providers, and vendors of prostheses. 20 | A Guide to Competitive Vouchers in Health Vouchers Encourage Use of Underconsumed Services Health vouchers may encourage people to visit providers they might not otherwise have seen. They are particularly useful for subsidizing services that tend to be underconsumed from a social welfare perspective, such as family planning, treatment of infectious diseases, immunizations, mental health care, and maternal and child health services. They are also useful when knowledge of the existence of services is poorly disseminated with- in the community. Vouchers Can Be Easy to Administer The administrative cost of voucher schemes is one of their main draw- backs. However, they can be administered more easily than other demand- side subsidies. Requiring the provider to present a voucher in order to re- ceive the subsidy can prevent irregularities and false claims. If designed well, the voucher can serve as a receipt and a data collection form, as well as a token of exchange. Vouchers Reduce Provider-Induced Demand Since they are controlled by the user, vouchers reduce the problems asso- ciated with provider-induced demand. And because they are normally used for a clearly defined and limited service at fixed cost, they probably reduce the risk of subsidies being claimed for more expensive conditions than those actually treated. For example, hospitals sometimes claim sub- sidies based on the patient's diagnostic-related group. Studies have shown, however, that when it is possibile to assign two different diagnostic-relat- ed groups to a patient, hospitals tend to use the one with the higher level of subsidy. Voucher schemes largely avoid this phenomenon, known as diagnostic creep. Vouchers Work Best for Service Packages of Fixed or Predictable Cost Voucher schemes seem to work best when a fixed value can be assigned to the benefits they provide. This makes it easy to reimburse providers, who are given an agreed-upon amount for each voucher they return to the agency. This type of arrangement is possible when the services provided Benefits of Competitive Voucher Schemes for Health | 21 Box 2-1. Varying the Service Package According to the Needs of the Patient in Nicaragua In Nicaragua's HIV/AIDS prevention program, treatment of sexu- ally transmitted infections varies according to the diagnosis. The voucher agency specified the management protocol for each in- fection and provides facilities with the necessary medicines. In special cases, when additional treatment is required, doctors can ask the voucher agency for reimbursement of additional expenses. The same mechanism operates in the adolescent program, which also has standardized management protocols for a range of serv- ices, including counselling, family planning, pregnancy testing, a first prenatal check, and treatment of sexually transmitted infec- tions. Source: Sandiford, Gorter, and Salvetto 2002b. can be specified clearly in advance and each patient receives exactly the same services. It also works when patients receive different services ac- cording to independently verifiable conditions, such as the result of a lab- oratory test (box 2-1). Alternatively, if the proportion of patients receiving different service packages remains constant, then a fixed constant fee (that is, a weighted average of the fees for each cost category) can be used. Prob- lems are likely to arise, however, when the costs to the provider of attend- ing to voucher-bearing patients vary greatly and unpredictably and are im- possible to verify independently. Vouchers Increase Client Satisfaction In a competitive voucher scheme, the bearer of the voucher can usually choose a provider. If the voucher covers the full cost of the services or if the cost charged by all providers is the same, the bearer will usually base the choice on perceptions about which provider offers the highest-quality, most convenient, and most comfortable service. Providers will raise the quality of their services in order to attract voucher-bearing users. PART II INTRODUCING A VOUCHER SCHEME CHAPTER 3 CONDUCTING PREFEASIBILITY STUDIES Before policymakers can proceed with feasibility studies, they must iden- tify the health sector problem and determine whether public subsidies can help solve it. This stage of the process involves five steps: · Identify the health sector problem meriting public subsidies and the aims of the voucher scheme. · Justify the selection of the health sector problem (above others) and the inputs or outputs to be subsidized. · Identify potential sources of funding. · Document key stakeholders and their interests, as well as possible part- nership opportunities. 25 26 | A Guide to Competitive Vouchers in Health · Identify existing service delivery systems that address the problem. Although each of these stages would appear to follow from the preceding one in a consecutive sequence, in reality neither a voucher scheme, nor any other health program, is always designed and implemented in such a logical process. IDENTIFYING THE HEALTH SECTOR PROBLEM AND THE AIMS OF THE VOUCHER SCHEME What problem in the health system are the subsidies going to address? What are the subsidies expected to achieve? Health systems interventions usually attempt to achieve one or more of the following aims: · Increase the health status of the population. · Reduce poverty and socioeconomic inequalities in health outcomes. · Provide services at a lower cost or get more for the same cost. · Increase patient satisfaction. Before conducting a feasibility assessment, policymakers should consider which of these (or any other) aims they expect the subsidies will achieve and assign priorities to each. They should also set targets against which to measure the success or failure of the scheme (see chapter 7). Some pre- liminary data collection might be necessary for a baseline measurement at this stage. JUSTIFYING THE USE OF A VOUCHER SCHEME Why are subsidies needed to address the problem? What will the subsidies be used for? Is there evidence that such expenditures are cost effective? Could the subsidies be used more effectively in addressing other health sector problems? Subsidies are justified when the market fails to achieve social goals--a common phenomenon in the health sector. Externalities such as the social costs of high birth rates and communicable disease, so- cioeconomic inequalities in health, and the risks of catastrophic health care costs are often cited as justifications for subsidies. This is the time to Conducting Prefeasibility Studies | 27 think about whether or not subsidies are really necessary to achieve the aims that have been set. Other forms of government intervention, such as regulation, may be sufficient. Will subsidies introduce other problems? If subsidies can be justified, it is important to write down the basis for this justification before moving on to the feasibility assessment. It is also im- portant to be able to justify the selection of a subsidy scheme over alter- natives by demonstrating that it is likely to make a greater contribution to social goals. If a specific health intervention is proposed, evidence from published international experience that the intervention will yield the expected health outcomes should be cited. Many health sector interventions do not work, and some are even harmful. It is difficult to justify subsidizing in- terventions that are known to be ineffective (although such interventions are regularly subsidized throughout the world). Much work has been done to separate out and promote evidence-based medical practice.1 While there will always be a role for innovation and testing of new ideas and strategies, the bulk of health sector interventions should be evidence based. Figure 3-1 is intended to help guide policymakers in determining whether subsidies are justified and in identifying which type of subsidy is most appropriate to address a given health sector problem. FUNDING THE SCHEME There is little point in proceeding to the feasibility assessment stage unless there is at least a reasonable possibility of finding a source of funds to pay for the subsidies. Where will the money for the subsidy come from? Is this source sustainable? The two most likely sources of funding are govern- ments and donors. Existing public funds could also be modified to fund a new subsidy scheme. 1 See the work of the Cochrane Collaboration, available at www.cochrane.org, which pre- pares, maintains, and promotes the accessibility of systematic reviews of the effects of health care interventions. Figure 3-1. Decision Tree for Supply-Side Subsidies 28 | Deciding whether A to use supply-side Guide subsidiaries to Competitive Is the aim to address Is the Is the Subsidies externalities? aim to increase aim to protect families may not be (e.g., infectious diseases, No equity in health No from catastrophic No justified. family planning)? outcomes? health Voucher cost? Yes Yes Yes sin Health Are Subsidize the Subsidize the there any inputs Does the Are there any supply of supply of that can only be used health problem inputs/services whose those inputs/ those inputs Yes for this health Yes only affect the costs are always Yes services. intervention? poor? catastrophic? No No No Are Is it more there any inputs Supply-side Are there efficient to means that can only be used subsidies areas where all users test input subsidies than to for this health No may not be are poor? means test insurance No intervention? appropriate. subsidies? Yes Yes Subsidize the Introduce supply of all means-tested services in input those areas. subsidies. Conducting Prefeasibility Studies | 29 UNDERSTANDING THE CONTEXT IN WHICH SUBSIDIES WILL BE INTRODUCED Before proceeding to a feasibility study, policymakers need to find out as much as they can about what has been done to address the problem and what opportunities and constraints exist to using a voucher scheme. Some of the questions they need to ask include the following: · If effective interventions exist, why does the problem persist? · What is the existing level of service delivery? · Are services currently being subsidized? How affordable are the services? · Who are the main stakeholders and what are their interests? · Who are the intended beneficiaries of the proposed voucher scheme? · What cultural factors could affect the success of a subsidy scheme? · What legal and regulatory issues could affect the success of a subsidy scheme? · What institutional opportunities and constraints exist? BUILDING PARTNERSHIPS Who are the likely winners and losers from the proposed subsidies? Can their support for the scheme be achieved? In the process of gathering in- formation on the roles and interests of various stakeholders in the health problem, opportunities may arise to identify potential partnerships and to float the idea of a voucher scheme with potential partners. How open is the ministry of health to a scheme that effectively breaks its provider mo- nopoly over the use of subsidies? How willing would private providers be to entering into a partnership with the public sector, to bid for a contract, to accept vouchers as payment, to sign contracts for services, to follow a strict patient management protocol, to participate in an accreditation scheme? Assessing different stakeholders' reactions can provide a valuable input to the next phase, the feasibility assessment. Partners may make different types of contributions. Some can provide services, others technical assistance; some can distribute vouchers, others can lobby donors. It is vital to keep an eye out for institutions that could perform the crucial voucher agency role. 30 | A Guide to Competitive Vouchers in Health Figure 3-2. Decision Tree for Demand-Side Subsidies Deciding whether to use demand-side subsidies. Is the Is the aim to address Is the Subsidies aim to protect families externalities? (e.g., aim to increase may not be from catastrophic No infectious diseases, No equity in health No justified. health cost? family planning)? outcomes? Yes Yes Is there Use population Is it possible Is it possible an existing social based subsidies to set a fixed price to quantify the population security or health (e.g., capitation No for the service covered by the Yes insurance system? payments). package? intervention? Yes Yes Is it Is it difficult to Is independent Introduce feasible to identify the target group confirmation of services a provider identify the target at the health No provided feasible? Yes subsidy claim population? facility? scheme. Yes Yes No Subsidize social No security/health Is it possible to insurance identify the target group No contributions. at the health facility? No Yes Subsidies may Subsidize using not be a voucher feasible. scheme. The analysis of existing service delivery may fail to identify some of the potentially most valuable partners. For example, organizations may be working closely with the target population--and therefore potentially valuable partners for voucher distribution--but not currently providing services that address the chosen health problem. These organizations need to be identified at this prefeasibility stage and some idea gained as to their capacity to contribute to the program. Policy makers should keep in mind that partnerships are only tentative at this stage and should be formalized during the design phase when contracts are negotiated and signed. Conducting Prefeasibility Studies | 31 CHECKLIST FOR FEASIBILITY ASSESSMENT Before moving on to the next stage in the process, policymakers need to check that they have completed all of the steps in the prefeasibility assess- ment (table 3-1). An example of such an assessment is shown in box 3-1. Table 3-1. Checklist for Prefeasibility Assessment Yes No Has the health sector problem been selected and justified (including target groups)? Have sources of potential funding been identified and explored? Has the full range of possible interventions been considered and one or more selected on the basis of expected cost-effectiveness? Have existing programs and services to address problems been documented? Have all other contextual aspects to addressing the problem been recorded? Has the full range of possible interventions been considered and one or more selected on the basis of expected cost-effectiveness? Have key stakeholders been identified and potential partnerships considered? Box 3-1. Prefeasibility Work on a Voucher Scheme for Cervical Cancer Screening Problem: Mortality from cervical cancer, the single greatest killer of adult women in Nicaragua. Aims of voucher scheme: Increase the uptake of screening among poor and high-risk women, improve quality of cervical cytology, ensure follow-up and effective treatment of precancerous lesions. Justification for choice of problem and use of public subsidies: Externali- ties associated with social costs of orphaned children if disease is not detected and/or treated on time, socioeconomic inequalities in continued 32 | A Guide to Competitive Vouchers in Health Box 3-1. Continued mortality rates, catastrophic health costs of treatment (and some- times even screening), market failure from inability of women to assess quality of cytology. Screening with Pap smear and prompt treatment of precancerous lesions are known to be cost effective. Potential source of funding: Donors and government; Ministry of Health to provide political support, identify priority areas, pur- chase vouchers and/or request donations, and treat patients with invasive lesions. Context: · Poor women and those at high risk of cervical cancer are missed, while younger and better-off women are screened more often than necessary. · Quality of cervical cytology is poor. No formal training or re- training for cytologists. Inadequate internal and no external quality assurance. · Inadequate follow-up and treatment, particularly in the public sector. Tendency to overtreat low-grade lesions (possible over- servicing). · Organization capable of serving as voucher agency exists. · Good network of public and private sector clinics able to take Pap smears. · Community health workers and civil society institutions are ca- pable of distributing vouchers and following up with women. Potential partnerships: The National Cytology Institute can assist in establishing external quality assurance scheme and accreditation system for participating clinics. CHAPTER 4 CONDUCTING A FEASIBILITY ASSESSMENT The purpose of the feasibility assessment is to determine whether it is pos- sible to structure the delivery of the proposed subsidies in the form of a voucher scheme. Feasibility assessment also involves making a judgment as to whether doing so is likely to have significant advantages over alter- native ways of delivering the subsidies. It may also be necessary to conduct ad hoc studies or pilot tests to determine whether a voucher scheme is fea- sible and desirable. The objectives of the feasibility assessment include the following: · Determine whether the proposed subsidy scheme has characteristics that make voucher schemes difficult to implement. · Determine whether the proposed subsidy scheme has characteristics that make vouchers a particularly effective delivery strategy. 33 34 | A Guide to Competitive Vouchers in Health · Weigh the pros and cons of providing the subsidy with a voucher scheme. · Decide whether to proceed with the design of a voucher scheme or to use an alternative strategy. Figure 4-1 illustrates a typical sequence in carrying out the feasibility as- sessment. The order in which the activities are performed is not critical, however. IDENTIFYING POTENTIAL IMPEDIMENTS Several potential impediments stand in the way of implementing a vouch- er scheme. Legal or Regulatory Impediments Some countries prohibit private medical practice. Lack of private doctors would severely constrain a competitive voucher scheme (although it might still be possible to promote competition among public sector providers). Other countries prohibit public sector providers from receiving cash, which limits the use of monetary incentives. In these cases, public providers cannot receive extra payment from outside or private funding sources, which becomes an issue if, for example, a private foundation were trying to fund the subsidies. Legal requirements can also undermine the flexibility the voucher agency needs to get a scheme up and running. If the scheme is based in a public sector institution, such as the ministry of health, laws and regula- tions regarding the contracting of staff or externally provided services may be rigid and bureaucratic. The need to obtain signatures from high-level functionaries for minor expenditures may make a voucher scheme unten- able. There may also be regulatory and even legal impediments to evidence- based, cost-effective best practice, which undermine one of the main ad- vantages of demand-side subsidies. Ministries of health frequently have norms for treating health problems--norms that do not always represent evidence-based best practice or the most cost-effective way of addressing a problem. In some cases, the norms may not have kept up with new tech- nologies. Ministerial norms do not always apply to the private sector. Al- Conducting a Feasibility Assessment | 35 Figure 4-1. Typical Sequencce of Activities in a Feasibility Assessment From prefeasibility work Consider the potential disadvantages of a voucher scheme · Lack of political support? · Cultural barriers to service use? · Insufficient spare capacity to provide services? · Monopolies in service provision? · No institution capable of performing voucher agency role? · High administrative and transaction costs? · Weak transport and communications networks? Consider the potential advantages of a voucher scheme · More accurate targeting of subsidies · Ability to stimulate demand for under consumed services · Administrative simplicity compared with other demand-side subsidies · Potential to reduce provider-induced demand · Scope for increasing patient/client satisfaction · Potential use for rationing subsidized high-cost services Assess the financial feasibility of a voucher scheme · Determine available resources · Estimate unit input costs · Estimate voucher redemption rates · Estimate costs per output produced · Compare with alternatives Do any of the disadvantages Yes make a voucher scheme a nonstarter? No Is the voucher scheme likely to achieve the aims No that have been set? Yes Is there any alternative way to achieve these aims Yes at a lower cost? No Voucher Will the voucher Voucher scheme is scheme create scheme is To design not feasible. Yes distortions whose ill No feasible. effects outweight its benefits? 36 | A Guide to Competitive Vouchers in Health though it may be possible to negotiate exceptions for public sector providers, this should not be assumed in advance. Lack of Political Support The importance of political support often depends on the role that the ministry of health plays in the scheme. Even if the voucher scheme is im- plemented by a private organization or a nongovernmental organization (NGO), the attitude of the ministry of health toward the scheme is likely to be important. For one thing, many donors require the prior agreement of the ministry of health. If government funding is used, the scheme may be introduced and promoted by the ministry of health. Its support may also be required if the voucher will be used at government-owned clinics or hospitals. Experience has shown that ministry of health officials (and even some donors) often feel that subsidies should be restricted to public sector providers, because they see their role as one of running and improving government hospitals and clinics rather than the broader one of sustain- ing and improving the health of the population. Government-owned clin- ics tend to be short of drugs, staff, equipment, and other necessary sup- plies. It is therefore easy to make a case for prioritizing their rehabilitation before purchasing services from private providers. Ministry of health offi- cials may also view voucher schemes as loosening their control over health sector development funding. Some will probably even have ideological objections to working with the private sector, because they feel that tax- payers' money destined for health care should not end up in private hands. Others may believe that private sector services inevitably cost more than publicly provided services. Political issues may affect support for the scheme. Voucher schemes can result in poor and underprivileged groups receiving better-quality services than the general population, which can cause resentment. And voucher schemes provide little opportunity for politicians to gain visibility by opening new government-owned facilities. Box 4-1 illustrates what can happen when political factors influence the design and implementation of voucher schemes. If the ministry of health is to play a significant role in promoting, de- signing, and running the voucher scheme, other obstacles will also emerge--unless the idea for the voucher scheme is put forward by the highest level of the ministry itself. Conducting a Feasibility Assessment | 37 Box 4-1. Political Impediments to Voucher Schemes in El Salvador and Honduras Political support from the highest level encouraged a donor to support the establishment of a voucher scheme for screening cer- vical cancer in El Salvador (Calero 2003a). Although it had been made clear that the program was to be run by an independent NGO according to agreed-upon strategies, once the project was ready to be implemented, the technical level at the Ministry of Health expected total control and involvement in running the program. The United Nations agency contracted by the donor to disburse the funding was unwilling to relinquish control, so the Ministry of Health changed the strategy to such an extent that the project became little more than a replication of what the ministry had already been doing in its own screening program. In particu- lar, the ministry restricted the provision of smear-taking services and voucher distribution to its own network of facilities and im- posed its own management protocol for patients with abnormal smears--against the expert advice of the project's consultants. The ministry's actions raised the costs of the voucher scheme and di- minished the opportunity to prevent deaths. In a donor-sponsored voucher scheme designed for the pre- vention of HIV/AIDS in Honduras, failure to secure political sup- port from the outset meant that although funding was secured, the project was eventually abandoned in favor of a traditional supply-side subsidy approach. Source: Instituto Centroamericano de la Salud. Voucher schemes for health are virtually unheard of in most countries. Legitimate concerns about potential abuses of the scheme (black markets, collusion between service providers and distributors, counterfeiting) can take on exaggerated proportions. Experience in setting up voucher schemes in diverse settings suggests that it is probably better to begin with small schemes and gradually win support within the ministry of health for expansion, once all of the "teething" problems have been addressed and 38 | A Guide to Competitive Vouchers in Health the program is running effectively. Once officials can see that a scheme leads to improved services for poor and vulnerable groups and that com- petition between the public and private sector can benefit both, it be- comes easier to win support for program expansion. Sociocultural Barriers to Service Use The mere distribution of vouchers may not provide recipients with suffi- cient incentive to overcome sociocultural barriers to using a service. Efforts to encourage vasectomy in men, for example, may be thwarted by a mis- perception that the intervention leads to impotence. Combining the voucher scheme with social marketing or an information, education, and communication campaign may help remove these barriers. For some health problems, vouchers may be stigmatizing (box 4-2). People may be reluctant to redeem vouchers--for abortions, emergency contraception, treatment of sexually transmitted infection or tuberculo- sis--because of the social stigma associated with these conditions and services. In such cases, it is important that the scheme make discretion and respect for patients' confidentiality as priority. Publicity campaigns in the mass media may be counterproductive; alternative ways of promoting the scheme to target groups should be employed. Lack of Capacity to Provide Services If service providers are working at full capacity, they will have little incen- tive to participate in a voucher scheme unless they are paid more than they currently are. With ambulatory care services, there is usually spare ca- pacity or the ability to increase capacity at relatively low cost (by employ- ing another doctor or working longer hours, for example). Capacity limi- tations are most likely to be encountered for secondary and tertiary care services, such as those requiring the use of operating theaters. These con- straints may prove critical. Capacity problems may also be important for services provided by only a few trained specialists; procedures for which there is a lack of inputs (for example, organ transplants); and procedures for which the equipment required is costly and already fully utilized (for example, magnetic resonance imaging). Conducting a Feasibility Assessment | 39 Box 4-2. Are Vouchers Stigmatizing? One of the concerns about Nicaragua's voucher program for the prevention of HIV infection and the treatment of other sexually transmitted infections in sex workers and their clients was that the stigma associated with these vouchers might discourage benefici- aries from keeping and using the voucher (for fear of being dis- covered by a relative, for example) (Sandiford, Gorter, and Salvet- to 2002b). Although this remains a concern, so far it seems that the benefits patients perceive they are receiving from using the voucher are greater than the stigma associated with being singled out as a sex worker or client. Since the scheme began in 1995, care has been taken to avoid publicizing the program, for fear that do- ing so might increase stigmatization. In a patient-led partner referral program for control of sexually transmitted infections in the Central African Republic (Koumans and others 2003), vouchers were offered to infected patients to give to their (often asymptomatic) partners for treatment. Al- though one might have expected some reluctance among patients to refer their partners, well over 90 percent of patients accepted the referral vouchers and 40­50 percent of program participants successfully referred at least one partner. Lack of a Competitive or Contestable Market for the Services Provided A voucher scheme does not require competition, but competition can greatly help improve patient satisfaction and productivity. A competitive market is one in which goods or services are produced or sold by a num- ber of different providers. Even in markets with many sellers, perfect com- petition may not exist. Collusion can allow sellers to fix prices. 40 | A Guide to Competitive Vouchers in Health Lack of Institutional Capacity to Perform the Voucher Agency Role The key to a successful voucher scheme depends to a great extent on the abilities and transparency of the voucher agency. To avoid any conflict of interest, the voucher agency should not be part of an organization that provides the services subsidized by the vouchers. It must have the remit, skills, and capacity to negotiate and contract health service providers. Above all, it must be an organization that can be relied on to act honestly and transparently. The voucher agency can be from the public or the pri- vate sector (profit-making or nonprofit), and there can be more than one voucher agency for a scheme (more than one agency is often necessary in large countries). A critical criterion is that the agency has the capacity to perform its role and to do so honestly. Financial transparency can be en- sured by regular audits, but it is equally important that the organization is--and is perceived to be--objective in negotiating and awarding con- tracts. Administrative and Transaction Costs Voucher schemes can be administratively onerous compared with tradi- tional supply-side subsidies, and there are significant transaction costs as- sociated with negotiating and monitoring contracts. A detailed discussion of this topic appears in chapter 5. Lack of Adequate Transport and Communications Voucher schemes require transport and communications networks devel- oped enough to ensure the reliable distribution of vouchers and the time- ly exchange of information between the voucher agency and other organ- izations involved. A detailed discussion appears in chapter 5. IDENTIFYING THE POTENTIAL ADVANTAGES Voucher schemes are potentially superior to other subsidy delivery strate- gies because they can more accurately target beneficiaries, stimulate de- Conducting a Feasibility Assessment | 41 mand for underconsumed services, reduce provider-induced demand, in- crease patient satisfaction, and ration high-cost services. They are also eas- ier to administer than other demand-side subsidies (see chapter 2). In addition to these advantages of voucher schemes, policymakers should also consider the advantages of demand-side subsidies generally, if the alternative to the voucher scheme would be a traditional supply-side subsidy. These advantages include the fact that remuneration is based on productivity; that subsidies can be restricted to evidence-based, cost-effec- tive services; that output-based monitoring and evaluation is possible; and that the incentives to improve productivity and client satisfaction are greater for demand-side subsidies than for supply-side subsidies. DETERMINING WHETHER A VOUCHER SCHEME IS FINANCIALLY FEASIBLE In the prefeasibility phase, potential sources of funding are identified and explored. In the feasibility phase, policymakers assess whether sufficient resources are available to cover the costs of the voucher scheme, achieve its aims, and do so at a lower cost than alternative subsidy delivery strategies. Costing a voucher scheme can be difficult without having implemented it, but it should be possible to at least draw up a budget. The costs can be di- vided broadly into the redemption value of the vouchers; those for the voucher agency itself; and those for the logistics required to run the scheme, including marketing, distribution costs, training, production of the vouchers, information system development and maintenance, and ex- ternal audit. A major source of uncertainty in costing a voucher scheme is the pro- portion of recipients who redeem their vouchers. If too few recipients use their vouchers, the resulting health or other benefits may be smaller than expected. Furthermore, the lower the redemption rate, the higher the dis- tribution costs in relation to the outputs achieved. The best way to esti- mate the likely redemption rate is to perform a pilot test (see chapter 6). 42 | A Guide to Competitive Vouchers in Health WEIGHING THE POTENTIAL BENEFITS AND OBSTACLES OF A VOUCHER SYSTEM Ultimately, weighing the benefits of a voucher system and obstacles to im- plementation become a matter of judgment. It may be helpful to structure this process by answering the following questions: · Are any of the impediments great enough that they make it impossible to implement the voucher scheme? · Is the voucher scheme likely to achieve its aims? · Is there an alternative way to achieve these aims at a lower cost (or to a greater extent for the same cost)? · Will the voucher scheme introduce distortions into the health system whose effects might outweigh the benefits of achieving these aims? Box 4-3 illustrates how a feasibility assessment might be conducted for a tuberculosis voucher scheme in the private sector. CHECKLIST FOR FEASIBILITY ASSESSMENT Before moving on to the next stage in the process, policymakers need to check that they have completed all of the steps in the feasibility assess- ment (table 4-1). Table 4-1. Checklist for Feasibility Assessment Yes No Have all the potential obstacles, disadvantages, and potential benefits of implementing a voucher scheme been considered? Is there a need to conduct any additional studies or pilot tests to reduce areas of uncertainty about the feasibility of a voucher scheme? Has a judgment been made as to the feasibility of a voucher scheme? Conducting a Feasibility Assessment | 43 Box 4-3. Assessing the Feasibility of a Tuberculosis Voucher Scheme for Private Practitioners Prompt detection and treatment of tuberculosis yields positive ex- ternalities, in the form of higher economic productivity of the in- fected person and the preservation of the health of people who otherwise would have been infected. Because tuberculosis dispro- portionately infects the poor, treatment subsidies also increase in- come equity in health outcomes. Since the costs of treatment are high, subsidies can also prevent patients from being driven into poverty. What are the potential benefit and obstacles of using a vouch- er scheme to treat tuberculosis? Potential benefits Potential obstacles · Vouchers can target tuberculosis patients · Doctors may be reluctant to who are difficult to reach. join a scheme that limits their ability to use nonstandard di- · Vouchers offer potential for greater privacy agnostic tests and treatment and higher quality of diagnosis and treat- protocols. ment. · National tuberculosis pro- · Demand-side subsidies give private practi- gram staff may believe that tioners greater incentive to test and treat tu- only public sector clinics berculosis. Vouchers simplify the adminis- should manage tuberculosis tration of these subsidies. patients. · Demand-side subsidies allow standardized, · Many private practitioners evidence-based, best-practice management are traditional healers, who protocols to be enforced in the private sec- would not be appropriate for tor. treating tuberculosis patients. · Providing vouchers for tuberculosis drugs · Costs are difficult to estimate (rather than the drugs themselves) reduces in advance. wastage from expired stocks. · Vouchers for diagnostic testing create a strong incentive for laboratories to increase screening of patients. · A voucher scheme facilitates information exchange between private practitioners and national tuberculosis control programs. CHAPTER 5 DESIGNING A VOUCHER SCHEME A voucher scheme can be described in terms of five key policy areas: · Recipient policies (who is eligible to receive or use the voucher) · Benefit policies (what services can be paid for with the voucher) · Price policies (how much the recipient pays to use the voucher) · Provider policies (which providers can participate in the voucher scheme) · Value policies (how much the provider receives for each voucher). The aims of the voucher scheme are specified and prioritized during the prefeasibility phase. In the design phase, a set of policies that is consistent 45 46 | A Guide to Competitive Vouchers in Health with the stated aims is determined. The various options available are de- scribed, and guidelines on which are suitable for achieving different aims are provided. CHOOSING A VOUCHER AGENCY Choosing the right voucher agency can make the difference between suc- cess and failure. The voucher agency is responsible for producing the vouchers, negotiating contracts with service providers, and reimbursing service providers on presentation of vouchers. It is also usually responsi- ble for monitoring the quality of the services or goods provided. In addi- tion, the agency may be in charge of distributing the vouchers, although it may subcontract or delegate this role to one or several other organiza- tions. If the distribution of vouchers is carried out by other organizations, the voucher agency is usually responsible for identifying these organiza- tions, negotiating with them, and paying for their services if they charge for them. In choosing a voucher agency, policymakers should place more em- phasis on competence and ability to deliver than on political or academic credentials. Furthermore, there need not be just one voucher agency. In- deed, in large countries, it may not be practical to have a single agency, un- less the organization has branches nationwide. Another alternative would be to subcontract out the voucher agency role, with different entities work- ing in different areas. Time-limited contracts could be awarded through competitive tender, making the markets for the agencies' services con- testable, if not fully competitive. Box 5-1 provides examples of the differ- ent types of institutions that have served as voucher agencies. The voucher agency can be a public sector institution, a private sector institution, or a parastatal organization, which has more independence and autonomy than traditional government departments. Whichever type of institution is chosen, the voucher agency must have four attributes: · The voucher agency must be neutral. It should not have links to any po- tential service providers, as such links could create conflict of interest. Thus, if ministry of health clinics are participating in the scheme as providers, it may not be appropriate for the ministry of health to serve as the voucher agency. Other public sector institutions, such as the min- istry of finance, could be sufficiently independent of service providers. Designing a Voucher Scheme | 47 Box 5-1. Who Runs the Voucher Program? Many of the reports on using vouchers in health care fail to docu- ment the institutional arrangements for the voucher agency role. But the information that is available suggests that a wide range of alternatives appear to have been used. The Local Initiatives Pro- gram in Kolchata, India, was implemented by the Child in Need Institute, a child health and nutrition non-governmental organi- zation. The Nyeri Youth Health Project in Kenya is run by the Fam- ily Planning Association of Kenya. The Taiwan (China) coupon scheme for intrauterine devices was administered by the Maternal and Child Health Association, which subsequently became the Planned Parenthood Federation affiliate. In Zambia, an emer- gency contraception scheme was set up primarily as a research project and run by the University Teaching Hospital (Skibiak, Chambeshi-Moyo, and Ahmed 2001). Ministries of health have also operated voucher schemes. In Indonesia, the Safe Mother- hood project, which made the services of private midwives afford- able for poor women, was run by the district health authority. · The voucher agency must have a good reputation. The agency is responsible for ensuring the transparency and accountability of the scheme and preventing abuse or misuse of vouchers. Opportunities may arise for collusion between the voucher agency and service providers. External audits may not be sufficient to ensure transparency. In theory, public organizations are more accountable to voters and taxpayers, but they may be just as susceptible to corruption as private agencies. Whichever type of agency is chosen, it makes sense to publish the criteria used for selecting providers and the agreements on fee rates (if they are not es- tablished by benchmarks). · The voucher agency must have the appropriate range of skills and experience. Needed skills include the ability to negotiate and contract with service providers and the ability to monitor performance of those contracts, in- cluding any quality specifications they may stipulate. General account- 48 | A Guide to Competitive Vouchers in Health ing and administrative skills are also required. In most cases the agency will also need skills in the specific area of health being tackled. This raises a problem, as the organizations with these skills tend to be serv- ice providers themselves or those with ties to service providers. Such skills can be brought into an organization relatively easily. Similarly, if the voucher scheme is providing services to a particular target popula- tion, such as commercial sex workers or people with disabilities, the agency should be familiar with beneficiaries' problems and needs. Some skills can be developed within institutions through training (see chapter 6). · The voucher agency must have sufficient autonomy to be able to handle fi- nancial management and contract providers. This is unlikely to be a prob- lem in private organizations, but regulations for government agencies are often strict. Sometimes these regulations can be circumvented by creating parastatal organizations with wider powers. ESTABLISHING RECIPIENT POLICIES Recipient policies clearly define the target beneficiaries, the geographic limits of the scheme, or both. Generally, policymakers define the target groups, while the voucher agency works out the operational definitions of eligibility. In an HIV/AIDS prevention scheme, for example, the govern- ment could contract with an organization that works with high-risk pop- ulations, which would distribute the vouchers to all of its clients. Figure 5- 1 shows a decision tree for designing recipient policies. Establishing Criteria for Voucher Eligibility Inclusion criteria that can be used in determining the recipient policies in- clude age; occupation (for example, miners, commercial sex workers, mi- grant workers); location; exposure to specific disease risk factors (for exam- ple, smoking, family history, contact with infected carriers); income; ethnic group; disease or health status (for example, tuberculosis patients, pregnant women); gender; sexual orientation; and previous health service usage. How do policymakers decide who should be eligible for vouchers? Tar- get beneficiaries need to correspond to the aims of the subsidy scheme (table 5-1). Designing a Voucher Scheme | 49 Figure 5-1. Designing Recipient Policies Think about who should receive vouchers. Establish eligibility criteria that include those who Was will most benefit from improving health the intervention and a priority? Yes exclude those who won't benefit. No Was targeting Establish eligibility criteria the poor a that include the poor or priority? Yes exclude the nonpoor. No Is the Establish the program restricted geographic limits georgraphically? Yes to eligibility. No Does the Decide on ways to voucher need to make vouchers be nontrans- Yes nontransferrable. ferrable Consider the sensitivity and specificity of your recipient policies. Continue Will a significant group of the target with No population be excluded? benefit Will a significant group of the nontarget policies. population be included? A program that seeks to increase equity and reduce poverty, for exam- ple, should target recipients who are poor. Geographic targeting is one of the simplest ways to reach the poor, but doing so risks including some people who are not poor, but who reside in poor areas, and excluding some poor people who do not reside in poor areas. It also excludes some 50 Table 5-1. Matching Beneficiaries with Program Aims | Program aim Target beneficiaries Country Reference A Guide Prevent HIV/AIDS by preventing and Sex workers, their partners, and their clients; Nicaragua Sandiford, Gorter, and treating sexually transmitted infections men who have sex with men; Salvetto (2002b) adolescent glue sniffers to Competitive Increase use of sexual and reproductive All poor adolescents 12­20 years of age in Nicaragua Sandiford, Gorter, and health services among adolescents selected regions Salvetto (2002b) Increase screening and treatment of All poor women 30­59 years of age from El Salvador Calero (2003a) women with preinvasive cervical villages or provinces in areas selected for Nicaragua Sandiford, Gorter, and Voucher abnormalities their high levels of poverty Salvetto (2002b) Increase use of village midwives by the Poor women who are pregnant or have a Indonesia Knowles (2000) sin poor child less than one year old Health Increase access to emergency Young women in need of emergency Zambia Skibiak, Chambeshi- contraception among adolescents contraception Moyo, and Ahmed (2001) Increase use of reproductive and child All female residents and children of selected India Mookherji (2003) health services in slum areas urban slums Reduce child mortality from malaria All pregnant women who receive prenatal Tanzania Marchant and others health services (2002) Uganda Root (2003) Reduce incidence of sexually Partners of patients with sexually transmitted Central African Koumans and others transmitted infections diseases Republic (2003) Designing a Voucher Scheme | 51 of the poorest people, who may not have an identifiable home at all. Oth- er ways to identify the poor, such as means testing, can be more accurate, but they are more costly. Another alternative is to delegate responsibility for identifying the poor to such groups as nonprofit organizations, com- munity health workers, and faith-based organizations that work with im- poverished groups. This strategy allows policymakers to increase the accu- racy of targeting by varying the number of vouchers distributed to each organization (or even to individuals within the organizations) based on poverty indicators recorded at the time the vouchers are redeemed. For example, a policymaker gives agency A and agency B 100 vouchers each to distribute among their respective recipient group. Each time some- one uses a voucher, his or her income level is recorded. Assume that the groups distribute all 100 and everyone who received a voucher used it. Af- ter the first round, the record shows that 80 percent of the voucher re- deemers who received vouchers from Agency A are poor, but only 20 per- cent are poor among those who received vouchers from Agency B. This means that only 100 out of 200 voucher recipients were poor. The next time, the policymaker gives 150 vouchers to Agency A and 50 to Agency B. This time, 130 out of 200 voucher recipients will be poor. Tradeoffs between Sensitivity and Specificity Sensitivity--reaching a higher proportion of specific groups--can be in- creased only at the cost of reduced specificity. Policymakers thus need to decide which is more important, not missing potential beneficiaries or ex- cluding people who do not belong to the target group. In a program in- tended to provide vouchers to the poor, excluding some poor people may seem unfair. Moreover, if too many poor people are excluded, the impact of the program will be limited. But including too many of the nonpoor wastes scarce program resources and subjects the program to criticism. Policies can be fine tuned to improve the targeting of a program, but im- proving sensitivity and specificity is costly. Transferability of Vouchers Vouchers can be transferable (that is, redeemable by someone other than the original recipient) or nontransferable. The initial temptation is always to make them nontransferable, in order to prevent people other than the 52 | A Guide to Competitive Vouchers in Health original recipients from receiving the benefits of the voucher. But making vouchers nontransferable may be costly. To prevent voucher transfer, one needs to identify the original recipient and have the service provider con- firm his or her identity. This can be done in several ways (box 5-2), but they add administrative complexity and cost to the scheme. Policymakers need to consider whether the cost is worth the increase in specificity. Even if it is possible to prevent people other than the voucher recipient from using the voucher, it may not always make sense to do so. The indi- rect recipient may well be as poor as the original recipient; where vouch- ers are targeted according to health need rather than income, indirect voucher bearers may even be in greater need of the service than the origi- nal recipients. Furthermore, if one assumes that a high proportion of the direct recipients who give their vouchers to indirect recipients would not have used them, transfer of vouchers may help keep redemption rates high, thereby raising the efficiency of the scheme at producing its outputs. An alternative to making vouchers nontransferable is to monitor the frequency of their use by indirect recipients and examine the proportion of these recipients that falls into the target group. Doing so provides in- formation upon which to base a subsequent decision to enforce transfer- ability or not. At the end of a consultation, service providers could simply ask bearers whether their voucher was given to them by a friend or a voucher distributor and whether or not they had to pay for it. Collecting this information makes it possible to model the effect of voucher transfer- ability on sensitivity and specificity. If the proportion of indirect recipients becomes too high, or if a black market develops, the possibility of making the voucher nontransferable could be considered. Policymakers need to ensure that the original recipients are fully aware of the benefits and do not pass on their vouchers simply because they per- ceive them to be of no value. This is something that should be monitored closely through studies of nonredeemers. Voucher distributors can play a key role in ensuring that recipients are well informed. Their efforts can be reinforced by a clear, well-targeted information, education, and commu- nication campaign. In some cases, ensuring that vouchers are not transferred is important. Some vouchers entitle the bearer to an expensive service, and considerable effort has been put into identifying suitable recipients. Restricting transfer is particularly important if the goods or services received have a mar- ketable value. This is most likely to be the case where the benefits of the vouchers can be used for diagnosing or treating illnesses other than those Designing a Voucher Scheme | 53 Box 5-2. How Can Vouchers Be Made Nontransferable? There are a number of ways to make vouchers nontransferable. Here are a few: · Write the recipient's name on the voucher and ask the provider to confirm it by asking to see a photo identification card. To en- sure that the provider checks the recipient's identification, the provider can be required to fill in the number of the identifica- tion card, which the voucher agency can check against a register kept at the time the voucher is distributed. This process may be costly, and many poor people may not have personal identifica- tion. · Have the recipient sign the voucher at the time it is distributed and again when it is redeemed. If a high proportion of the pop- ulation cannot write, a thumb print can be used instead. The provider can compare the signatures or thumb prints to ensure that they belong to the same person. This is a relatively low-cost technique, but it may make the voucher less attractive to recipi- ents (whether it does so could be checked in a pilot test). Providers may need to be trained to read and compare thumb prints. · Attach a photograph to the voucher. Photos are expensive, but if the redemption value of the voucher is high, this measure may be justified. This method does not prevent providers from ac- cepting vouchers from people other than those whose photo- graph is on the voucher, but few voucher bearers will try to use a voucher that carries a photo other than their own. The honesty of the provider can be monitored periodically by the use of "mystery patients" (people with vouchers carrying photos of someone else). · Take digital images and send them electronically to providers. Digital photos are far cheaper to produce than printed photos. continued 54 | A Guide to Competitive Vouchers in Health Box 5-2. Continued They can also be sent electronically. A digital photo could be taken at the time the voucher is distributed and sent to the provider by e-mail or posted to a secure Web site. For addition- al certainty, providers could be asked to make a digital image of the patient, which could be compared with the original. The ob- vious limitation of this method is the need for digital cameras and universal Internet access by providers. For some target groups, such as commercial sex workers or men who have sex with men, this method may be unacceptable. the scheme is focusing on (for example, broad-spectrum antibiotics). Nontransferability is more important when the aim of the voucher scheme is to increase equity in health than when the aim is to control communi- cable disease, since the success of the scheme can be undermined if the voucher is sold to someone better off than the original recipient. Prevent- ing the transfer of vouchers is also important when a core group for the spread of an infectious disease (such as commercial sex workers) is being targeted, since, assuming that the original recipient would otherwise have redeemed the voucher, transferring it to someone not in the core group undermines the scheme's effectiveness in controlling the disease. DETERMINING BENEFIT POLICIES Benefit policies define what the voucher entitles its bearer to receive. The benefit can be a specific health service or a package of services. Table 5-2 illustrates the wide range of health services that have been provided through voucher schemes. Benefit policies are particularly important if health gain/effectiveness is a key objective of the voucher scheme. Benefit policies allow policymakers to define a package of services or a patient management protocol that is derived from evidence-based, cost-effective best practice. The more effec- tive the services covered by the benefit policy, the greater the health gain produced by the scheme. The process of determining the benefit policy Table 5-2. Benefit Policies of Selected Voucher Schemes for Health Aim Benefit Location Reference Provide female contraception through Insertion of IUDs Taiwan Cernada and Chow insertion of intrauterine devices (IUDs) (China) (1970) Improve family planning and pre- and Package of maternal health care and Indonesia World Bank (2000) postnatal maternal health services family planning services through village midwives Prevent sexually transmitted infections Package of sexual health services, including Nicaragua Gorter (2003) and HIV, and treat sexually transmitted voluntary counseling and HIV testing infections among sex workers and other vulnerable groups Improve the uptake and quality of sexual Package of services, including counseling, Nicaragua Gorter (2003) and reproductive health care for family planning, pregnancy tests, prenatal care, adolescents and treatment of sexually transmitted infections Increase access to and improve quality Package of services including screening and Nicaragua Sandiford, Gorter, and of cervical cancer screening for poor treatment for precancerous lesions Salvetto (2002a) Designing women El Salvador Calero (2003a) Provide young people with access to Package including diagnosis of sexually Kenya Erulkar (2003) sexual and reproductive health care transmitted infections, male circumcision, a and family planning services Voucher Provide girls with access to emergency Dose of emergency contraceptive pill Zambia Skibiak, contraception through a variety of Chambeshi-Moyo, Scheme suppliers and Ahmed (2001) Provide residents of urban slums with Package of reproductive and child health India Mookherji (2003) |5 reproductive and child health services services 5 56 | A Guide to Competitive Vouchers in Health may therefore involve a search of the literature or the employment of a technical expert. Ideally, one should draw up a detailed patient manage- ment protocol, perhaps in the form of a flowchart, that covers every con- ceivable contingency. If management depends on clinical diagnoses or the outcome of laboratory tests, this should be reflected in the protocol. The benefit policy must be clear to both the voucher bearer and the service provider. If bearers do not know what is excluded, they may have false expectations, especially if a copayment is required (see box 5-3). If providers do not know what is included and excluded, they will not be able to price their services accurately during the tendering process. It is im- portant to specify whether the voucher entitles the bearer only to diagno- sis of a disease or to treatment as well. When treatment is included, it should be clear what the treatment is and at what point in the disease process the patient is entitled to it. If, for example, operations for invasive tumors or antiretroviral treatment for people with HIV are excluded, it must be clear. It is customary to print the benefit policies on the voucher itself. If there is insufficient space on the voucher, it may be more practical to print the policies on an accompanying flyer or packaging. It is also im- portant to accurately describe the benefit policies in any social marketing or information, education, and communication campaigns that accompa- ny the voucher scheme. In most situations, it is wise to establish an expiration date for the vouchers. Setting an expiration date helps limit the financial risks assumed by the voucher agency (which could otherwise be liable for payments af- ter the scheme had officially ended), and it can ensure that the contract with the provider remains in force. Expiration dates also encourage recip- ients to use the vouchers. Without expiration dates, or with long validity periods, recipients may be tempted to save their vouchers. During this time, they may lose or forget about the vouchers. Having an expiration date also helps in the process of evaluating the uptake strategy. If a short validity period is set, the voucher agency has to be confident that the vouchers can be distributed with plenty of time in which recipients can use them. DETERMINING PRICE POLICIES The price refers to what the voucher bearer pays to the provider at the time the voucher is redeemed. This amount can range from nothing, for a fully Designing a Voucher Scheme | 57 Box 5-3. Benefit Policies: The Importance of Stating the Limits The cervical cancer program in Nicaragua offers a package of serv- ices exclusively for screening and treating precancerous lesions. Although the scope of the benefits was clearly specified both on the voucher itself and during the social marketing campaign, some women expected and requested a full gynecological consul- tation and breast screening. Had the limits of service provision not been clearly stated on the voucher, this confusion may have led to serious problems and bad publicity for the scheme. subsidized scheme, to the cost of the services (see box 5-4).1 The price pol- icy can take the form of a percentage discount off the normal price or a fixed monetary discount (for example, a discount of $5 off the consulta- tion price). The price paid by the voucher bearer can be variable, with dif- ferent discounts given to different groups. Thus very poor people may pay nothing for services, while others pay 50 percent. Price policies are particularly important where the main aim of the voucher scheme is to increase equity or reduce poverty. In these cases, the voucher bearer must be fully subsidized or pay a price that is well below the market price. Together with the recipient policy, the price policy deter- mines how effective the scheme will be in achieving its equity and pover- ty reduction aims. Price policies can also affect the efficiency of the scheme. Costs that can be recovered from patients will lower the total cost of the scheme to the donor or government (or enable it to cover a larger population). But costs borne by the voucher bearer, no matter how nominal, will deter some people from using their vouchers. Asking voucher recipients to bear some of the cost also introduces administrative costs that fully subsidized schemes do not face. In some voucher schemes, the bearer is paid for using the voucher. These schemes, which 1 have been used in drug rehabilitation programs, are not considered here. 58 | A Guide to Competitive Vouchers in Health DETERMINING PROVIDER POLICIES Provider policies determine who can provide the benefits of the vouchers and under what conditions. They are important for three of the main aims of subsidy schemes, for different reasons. For health gain, provider poli- cies make it possible to select providers with the highest technical quality and therefore the greatest effectiveness of treatment. For technical effi- ciency, the policies affect how much the scheme costs. For patient satisfac- tion, a free provider participation strategy can be used that allows patients to decide whom to see. Box 5-4. Price Policies Used by Different Voucher Schemes Voucher schemes can take different approaches to prices. Box Table 5-1. Price Policies of Voucher Schemes Program Location Reference Price policy Female contraception Taiwan Cernada and 50 percent discount on program (China) Chow (1969) the cost of inserting intrauterine device National Insecticide- Uganda Root (2003) Discount of the impregnated Bednet equivalent of $1 program for purchase of net Randomized trial for United Total exemption from mammography States fee or reimbursement of screening for women 80 percent of subsidized with Medicare fee, provided patient meets Medicare deductible Safe motherhood Indonesia Knowles Village midwives can program (vouchers (2000) charge only supplemental subsidize services fees for consumables provided by village purchased on open midwives) market, such as drugs La Clínica, the voucher agency and provider of some services in a voucher scheme for migrant farm workers in Wisconsin, has a fixed third-party provider reimbursement schedule (box table 5-2). continued Designing a Voucher Scheme | 59 Box 5-4. Continued Box Table 5-2. Fees Charged by La Clínica (1992) Service Fee Outpatient care Office visit $15 maximum Prescription $5 maximum Laboratory services $15 maximum Dental visit $35 maximum Emergency room visit 75 percent of total cost X ray 75 percent of total cost per X ray X ray interpretation 75 percent of total cost per X ray Inpatient care Hospital charges 60 percent of charges per admission, up to $500 maximum Physician charges 50 percent of charges per admission, up to $250 maximum One-day surgical procedures Hospital charges 60 percent of charges per admission, up to $400 maximum Physician charges 50 percent of charges per admission, up to $200 maximum Source: Slesinger and Ofstead (1996). The most liberal provider policy allows any provider to participate in the voucher scheme. Such a policy can be a good idea when it is important to make the services as widely available as possible and when the admin- istrative costs associated with each provider can be kept low. The down- sides of a liberal provider policy are that administrative costs are to some extent proportionate to the number of participating clinics and that such a policy does little to control the technical quality of service provided (see figure 5-2). If the number of providers is restricted, the process of selection should be transparent, so that as many providers as possible take part in the se- lection process. 60 | A Guide to Competitive Vouchers in Health Criteria for Selecting Providers A number of criteria can be used to select clinics for participation in the scheme. The first is cost. Providers can be asked to bid; those offering the lowest bids can be selected. Alternatively, a benchmark price can be estab- lished, and clinics offering to provide services to a larger population at that price can be selected. Figure 5-2. Designing Provider Policies Provider policies Consider again the objectives of the scheme. Is there a Draw up a list of verifiable quality justification quality requirements for for restricting the Yes particiaption in the scheme. numbers of providers? No Is there an administrative Decide how many providers should be cost justification for allowed to participate and define restricting the Yes transparent selection criteria. numbers of patients? NO No Decide on other criteria used to select Continue providers. with For example, prices for the services, reimbursement minimum opening hours, staffing levels, value presence of basic or specialized policies. equipment, and so on. Designing a Voucher Scheme | 61 A second criterion for selecting providers is quality. All participating clinics can be required to meet certain minimum standards, including the following: · Minimum opening hours · Staffing levels (for doctors, nurses, receptionists) · Basic or specialized equipment · Communications, such as e-mail · Registration with the ministry of health · Willingness and ability to send samples to laboratories and patient records to the voucher agency · Performance of staff in proficiency tests · Average patient waiting times · Patient satisfaction indices Some illustrative examples can be found in box 5-5. Encouraging Providers to Participate It may be necessary to spend time with providers to convince them of the benefits of participating in the voucher scheme. Where voucher schemes are new, providers may be reluctant to participate. They may wonder whether they will actually be reimbursed for the vouchers they return. Some may find the contracting or accreditation processes intrusive or threatening. Others may object to the loss of autonomy brought about by the imposition of a fixed patient management protocol, details of which they may not agree with. Providers are also sometimes worried that the so- cioeconomic profile of voucher-bearing patients may undermine the im- age of their clinics as places attended by the wealthy. It is difficult to allay all of these fears until providers see the scheme op- erating in practice. Once the program is operating successfully, some providers who originally refused to participate may express interest in do- ing so. The scheme needs to create opportunities for such providers to par- ticipate. 62 | A Guide to Competitive Vouchers in Health Box 5-5. Choosing Service Providers that Meet Program Needs The cervical cancer screening program in El Salvador needed to contract with a laboratory to provide cytology results, but none of the laboratories tested passed the voucher agency's proficiency test. The lack of a provider encouraged a few professionals who were skilled at reading Pap smears to form a laboratory, which was subsequently contracted by the voucher agency. In the Zambian emergency contraception voucher scheme, all public maternal and child health units had to be excluded from the list of service providers because of their restricted working hours (Monday to Friday only). Since, to be effective, emergency contraception needs to be taken within 72 hours of unprotected sex, any outlet that could potentially be contracted had to be open 24 hours a day, 7 days a week. The agency eventually chose private pharmacies and hospital outpatient departments. Pharmacies have the additional advantage of anonymity, which is particularly well suited to a sensitive issue such as emergency contraception. DETERMINING REIMBURSEMENT VALUE POLICIES The value of the voucher is the amount the provider receives upon return- ing the voucher to the voucher agency. The simplest mechanism is one in which providers agree in advance to a fixed value for the voucher and all providers receive the same amount. This method is used if a benchmark price is used to select providers. An alternative is to use the providers' actual tendered price and pay dif- ferent providers differently. This policy should be pursued with caution, as it can lead to accusations of lack of transparency. Even more sophisticated policies, which generate incentives for providers, are possible. Payments can be made to voucher distributors based on the number of target group patients treated, on the redemption rate they achieve with their vouchers, or on the socioeconomic profile of patients treated. Designing a Voucher Scheme | 63 In many situations, the service package received by the patient varies, based on clinical diagnoses, laboratory tests, or other reasons. If this is the case, it is important to independently establish verifiable criteria for values that are above the minimum in order to prevent what has been called "di- agnostic creep" (the tendency for providers to select higher-value packages). One way to do so is to maintain separate contracts for providers of labora- tory diagnostic services or services that merit higher reimbursement values. Value policies can affect effectiveness of a voucher scheme if they are used as incentives for providing additional services (if, for example, extra payments are made for referring patients for voluntary counseling and testing for HIV). DESIGNING THE VOUCHERS AND OTHER MATERIALS Once overall design of the voucher scheme is decided, the voucher itself needs to be designed to maximize the redemption rate of the voucher. Creating an Attractive Graphic Design The appearance of the voucher often has an important effect on whether or not the recipient uses it. The voucher should be attractive to recipients and give the appearance that it represents something of value. A modest investment in the services of a graphic designer can pay off handsomely. If the voucher scheme is promoted through social marketing, it is sensible to use the same slogans or images on the voucher as on the promotional material. Preventing Counterfeiting A wide range of measures can be taken to prevent counterfeiting. These measures vary in cost as well as the degree of security they confer. Policy- makers must assess the risk of counterfeiting and weigh it against the cost of different security measures. Low-cost measures that can be employed to prevent counterfeiting include the following: · Individually number the vouchers (including check digits). · Use watermarks. 64 | A Guide to Competitive Vouchers in Health · Use several fonts. · Use colored ink stamps. · Use bar codes for numbering. · Use self-adhesive seals · Laminate the vouchers or package them in cellophane. Of these measures, the first is perhaps the most important, since it enables one to quickly identify duplicate vouchers or vouchers with numbers that have not yet been distributed. Check digits are a more sophisticated meas- ure, based on making one digit of the voucher number the result of a mathematical function of some or all of the other digits. A computer pro- gram can quickly detect counterfeit numbers by performing a check as the number is introduced, but this can add to printing costs, since the num- bers will not be a simple sequence. Most of the other measures simply add to the cost of producing counterfeit vouchers. Unless counterfeiters think they can easily recover this cost, they are unlikely to make the investment. Expiration Date If the vouchers are valid only for a limited time, it is important that the ex- piration date be immediately visible, especially if the vouchers are in a closed package. If the vouchers come in a closed package, it may be useful for voucher distributors to ask recipients to open the packet upon receipt. Multiple-Section Vouchers Sometimes it can help to create a voucher with multiple tear-off or cut-off sections, in which one part of the voucher remains with the bearer, while other parts are returned to the voucher agency or sent to a laboratory with the specimens (box 5-6). It can also be useful to leave a space in the vouch- er for the patient's signature, to ensure that patients actually receive the health service they were supposed to have received. This signature can be compared with one obtained at the time the voucher is distributed. Doing so can help monitor transfer of vouchers to indirect recipients and detect fraud by providers. Designing a Voucher Scheme | 65 Printing Because unit printing costs drop dramatically as the number of vouchers printed increases, it is usually worth producing a large number of vouch- ers at a time, enough for several rounds of distribution. Doing so, howev- er, makes it difficult to change policies. Distributing the voucher with a cheaply produced insert that explains the voucher policies in detail, in particular the benefit policies, allows the policies to be changed after the voucher is printed. It also allows the voucher itself to be small and portable. Box 5-6. Designing the Voucher GinecoBONO, the voucher used in the cervical cancer program in Nicaragua, is sequentially numbered, with the expiration date stamped on top and the services offered clearly stated both inside and outside the packaging, which is a packet or booklet. One part of the voucher, which includes a space to write the appointment for the second visit (to collect test results), is retained by the pa- tient. The other part of the voucher is retained by the clinic. The patient is asked to sign this part of the voucher, which is then re- turned to the voucher agency as proof that service was rendered. In the Taiwan (China) female contraception program, the voucher consisted of three detachable parts, each of which served a different purpose. The first part remained with the field worker as a record of distribution, the second and third parts were given to the patient, who handed them over to the physician. The physi- cian held on to the third part as a proof of services delivered. He or she sent the second part by registered mail to the county nurse, who forwarded it to the voucher agency. This part of the vouchers was used for reimbursement of services as well as evaluation and monitoring purposes. 66 | A Guide to Competitive Vouchers in Health TRANSPORT AND COMMUNICATIONS Voucher schemes require reliable communications--between the voucher agency and voucher distributors and between the voucher agency and service providers. The voucher agency needs to be able to distribute the vouchers. Providers need to be able to return them, along with any infor- mation required, to the voucher agency. Patients sometimes need to be transported to specialist care providers. The logistics involved in organiz- ing reliable transport and communications can be complex (box 5-7). To ensure that they work properly, they should be pilot tested. If providers are required to send samples to a central laboratory, systems to transport the samples need to be established, along with reliable ways to ensure that providers, patients, and usually the voucher agency receive the results. Producing double (or triple) copies of the results can help. DISTRIBUTING THE VOUCHERS Vouchers can be distributed by the voucher agency, by an external agency contracted by the voucher agency, or by providers. In most cases, vouchers can be handed directly to beneficiaries. It is also possible to make the vouchers available at various dispensing points (such as municipalities), where beneficiaries pick them up. This strategy can significantly lower costs. Box 5-8 describes a scheme in which community leaders were re- sponsible for voucher distribution. Whoever distributes the vouchers, they must reach the intended bene- ficiaries. If providers distribute the vouchers, beneficiaries will be restrict- ed to those who visit their clinics, and providers may be tempted to give them to patients other than those for whom they are intended, especially if recipient characteristics are not monitored (see chapter 7). It is important to keep track of whom the vouchers have been given to. The voucher agency should record the serial numbers of the vouchers giv- en to distributors. These simple measures facilitate evaluation of the effica- cy of the voucher distribution strategy. They also provide a check on coun- terfeiting and prevent distribution of vouchers at the point of service delivery. Designing a Voucher Scheme | 67 Box 5-7. Facing the Challenges of Transport and Communications Transport logistics can be complex--especially in a region in which some of the areas served are accessible only by plane or boat in the wet season, as is the case in Nicaragua's cervical cancer program. That program has four transport systems in place: · Health promoters from the voucher agency and private clinic nurses distribute the vouchers. Service providers organize and pay for distribution, which is coordinated and supervised by the voucher agency. · Medical records, Pap smears, and laboratory results are sent to the voucher agency in the capital by planes, buses, a courier service, and some of the clinics' own vehicles. The cost is shared by service providers and the voucher agency (on behalf of the subsidy provider). · Community health workers follow up on patients with high- grade lesions who fail to return to the clinic for their results. The workers visit the women and advise them that they need treatment. If this fails, a promoter from the voucher agency is sent. If necessary, the health promoter provides the woman with the money required for transport to the clinic. · The voucher agency (on behalf of the subsidy provider) pays the expenses of women requiring specialist treatment in the capital if they cannot afford the bus ticket or overnight accom- modation. PAYMENT OF SERVICE PROVIDERS In the simplest cases, service providers bring or mail in the vouchers and the voucher agency calculates the amount owed, writes a check, has the provider sign a receipt, and records the numbers of the vouchers that have been returned. If a computerized information system is in place, it is not difficult to have it automatically generate receipts for each provider, with 68 | A Guide to Competitive Vouchers in Health Box 5-8. Using Village Leaders to Distribute Vouchers Village leaders may be better able to identify the poor than outside agencies, and they can make qualitative judgments that can be more valid than quantitative indicators. Village leaders may be in- clined to identify their relatives, friends, and political supporters as beneficiaries, however (Gwatkin 2000). To avoid this, formal methods for identifying the poor that involve village leaders can be used. One such method is the participatory wealth ranking, which uses a community's own definitions and perceptions of poverty and employs rigorous cross-checking methods to ensure consistency and accuracy of results (Simanowitz, Nkuna, and Kasim 2000). A safe motherhood program in Indonesia has experimented with using village leaders to identify the poor and distribute vouchers to them (Knowles 2000). Under the program, women are eligible if their family eats two or fewer meals per day, their husband has lost his job, some children in the family are not at- tending school for economic reasons, or the community health center is too far away to be easily accessed from the village. Whether the program has succeeded in targeting beneficiaries re- mains to be seen. the numbers of the vouchers recorded on them. A decision has to be made as to how often providers are paid for their services. This is a compromise between the wishes of the providers (who like to be paid as soon and as frequently as possible) and those of the agency, for which frequent pay- ment adds to administrative costs. DEVELOPING INFORMATION SYSTEMS Design and development of the information system is best carried out once all other design aspects of the scheme have been worked out and the information needs have become clear. Voucher schemes need only infor- Designing a Voucher Scheme | 69 mation that contributes to decisionmaking or determines actions. There is no point in collecting or recording information that will never be used. The development of information systems for voucher schemes is facilitat- ed by the fact that each voucher represents a self-contained unit, which can be recorded as a single "case" in a computer database. Data Collection Forms One of the first steps in designing the information system is designing the data collection forms. Some of the forms and registers that may be used include the following: · Registry of voucher distributors (individuals, organizations, or both). The reg- ister should record the serial numbers of the vouchers given to each dis- tributor. The register will help maintain control over voucher distribu- tion and utilization. · Registry of voucher recipients. For each voucher, it is useful to record the date and place of distribution. It can also be useful to obtain the name, address, date of birth, and signature of the recipient in order to moni- tor transfer of vouchers and to facilitate studies of nonredeeming voucher recipients. · Registry of service providers. Keeping a registry of service providers facili- tates both quality control and reimbursement. The registry should in- clude the dates the providers began participating in the program. · Clinical and diagnostic records. Clinical and diagnostic test records can have an important impact on public health policy and planning. · Registry of patients requiring follow-up. It is important to track patients carefully to minimize loss to follow-up. If feedback can be obtained once treatment is complete, it should be recorded. · Record linkage. Ideally, all registries and records should be linked in a common database. It can be helpful if the system is able to recognize the same patient with two different voucher numbers and recall the clinical history from previous voucher use. This can assist in determin- ing appropriate patient management, but it may be difficult to make this information available to service providers without a sophisticated computer network. 70 | A Guide to Competitive Vouchers in Health Design of data collection forms is something of an art. The forms should be unambiguous and nonredundant, and they should include adequate space for responses (especially addresses). The check boxes and other spaces for responses should be aligned along the right-hand margin of the form, to facilitate data entry. Wherever possible, the form should occupy a single sheet of paper, even if this means using a larger sheet, a small font, or both, in order to prevent multiple sheets from becoming separated. Perhaps the most common flaw in form design is making them too long. This wastes time and makes those filling out the forms less careful. Sometimes clinics gather more data than are required for the purposes of the scheme. These data need not be incorporated in the forms or reported in the information system. If possible, only data that will affect decision- making about the scheme should be solicited. The forms should be tested and retested before finalizing them and printing large quantities . As a general rule, the voucher agency should receive a copy of all forms. Service providers will probably wish to keep a copy of the form for their records. Sometimes diagnostic service providers will require copies of at least parts of the clinic record. Multiple copies of forms can be produced cheaply using NCR (no carbon required) paper that prints through to the sheet below with the pressure of a pen. Data Entry Each form should be entered into a computerized database linking forms with voucher numbers. The data entry program can mimic the exact for- mat of the form, with drop-down menus for multiple-choice responses, and it can perform consistency checks as the date is entered. The data en- try phase can provide an additional opportunity to detect transfer of vouchers to indirect beneficiaries. Comparing names and surnames with those recorded at the time of distribution can be a first screen, but varia- tions in spelling and the inclusion or exclusion of middle names can make it difficult to automate this process. If information such as the recipient's date of birth was recorded when the vouchers were distributed, it may be a better first screen for voucher transfer. Designing a Voucher Scheme | 71 Using the Database to Run and Monitor the Program What decisions and actions should the database assist? Many of these are related to monitoring and evaluation, discussed in chapter 7. In addition, the database should provide insights into the following: · Distribution strategies. By providing voucher redemption rates for each voucher distributor, the information system identifies which are the most effective. This information can be used to create productivity in- centives. The information system can also identify the distribution sites from which vouchers are most likely to be redeemed, and it can identi- fy where they achieve the scheme's aims most cost-effectively. · Provider policies. The database can identify which providers voucher bearers favor. This information can be used to drop ineffective providers by not renewing their contracts. If voucher bearers are ex- pected to make more than one visit to a clinic, the proportion who fail to make a second visit may be used as an indicator of poor-quality care. Providers who provide such care can be dropped. · Provider remuneration. The system can automatically generate monthly receipts for payments to providers. · Abuse. The system can alert staff to possible counterfeiting, black mar- ket formation, collusion, or other types of fraud. The system can be programmed, for instance, to accept only voucher numbers that have been produced and distributed. · Reporting to funding agencies providing the subsidies. The system can pro- duce summaries of patient outcomes or service delivery outputs. If the voucher scheme is designed as an output-based assistance project, these outcomes or outputs will form the basis for invoicing the funding agency. · Storing medical records. Where questions are raised about the manage- ment of individual patients, computerized databases can be used to quickly access records. An information system can also allow more than one staff member to access records at the same time. · Tracking cases and follow-up. The information system is extremely useful as a way of flagging patients requiring follow-up or identifying individ- uals in response to queries from providers regarding their manage- ment. 72 | A Guide to Competitive Vouchers in Health CHECKLIST FOR VOUCHER DESIGN It can be helpful to use a checklist when designing vouchers, to ensure that all of the issues surrounding the vouchers and their use have been consid- ered and handled. Table 5-3 shows such a checklist. Table 5-3. Checklist for Voucher Design Yes No Has the voucher agency been selected or created? Has a set of rules been drawn up to define who receives the vouchers, what the voucher entitles its bearer to, what the bearer must pay to the provider to use the voucher (if anything), who can participate in the scheme as service providers (and under what conditions), and what each provider will receive in payment from the voucher agency (and how payment is to be determined)? Have the voucher and promotional material been designed? Have the logistics systems (especially transport and communication) been fully worked out? Has a strategy for distributing the vouchers been selected? Have the mechanisms for paying contracted providers been determined? Has the information system been designed? CHAPTER 6 IMPLEMENTING A VOUCHER SCHEME Successful implementation is not simply a matter of distributing vouchers and having them redeemed. It requires achieving the aims for which the subsidy scheme was created. TENDERING FOR SERVICE PROVIDERS If the number of service providers is limited--either because of the ad- ministrative costs involved or because of the desire to achieve potential economies of scale--some degree of competition can be retained by awarding the contract (or contracts) by public tender. Tendering for serv- ice providers allows the voucher agency to make a selection based on price and quality. In drawing up the announcement, it is essential to make the 73 74 | A Guide to Competitive Vouchers in Health request for services as specific as possible to enable providers to price their services accurately and to avoid having to deal with a large number of un- suitable providers. To maximize the benefits of competition, it is impor- tant to encourage as many providers as possible to participate in the ten- dering. Advertising the tender, either in national newspapers or specialized periodicals, is one way to publicize the tender, but advertising can be ex- pensive and some providers may miss the advertisement. Some providers may be reluctant to serve subsidized population groups for fear of lower- ing their status among their nonsubsidized clientele. Others may have pre- conceptions about slow and low payments (perhaps based on previous experience with the public sector). For these reasons, alternative strategies, such as mailings to clinics registered with the ministry of health or to in- dividual providers, should also be considered. In certain geographic settings or for certain highly specialized services, choosing providers may not be possible, and formal tendering is not ap- propriate. The existence of the subsidy scheme could encourage new providers to establish service provision in the area. If it does not, the voucher scheme can either use the only service provider available or con- tract services from providers outside the immediate area in which vouch- er recipients live in order to generate some competition. Before contract negotiations begin, the provider policies established in the design stage need to be applied to the selection process. The process of selecting a provider must be totally transparent: all providers who ex- pressed interest in participating must be told why they were or were not selected. Doing so will encourage them to participate in future tenders. NEGOTIATING CONTRACTS Negotiation is a process of communication in which two parties seek to arrive at a mutually satisfactory result on a matter of common concern. A mutually satisfactory result is necessary because each side depends on the other and both must live with the agreement. Box 6-1 provides some use- ful tips on negotiation. Price Price is one of the key points to be negotiated in a competitive voucher scheme. The voucher agency wants to keep the price as low as possible Implementing a Voucher Scheme | 75 Box 6-1. Tips on Negotiation · Listen carefully, be sensitive, and be quick to adapt to changing situations and to assess the impact of changing situations on the negotiation objectives. · Avoid confrontation. Disagree with people politely. Use humor to defuse tension. Tolerate conflict while searching for agree- ment. Attack problems and sticking points, not individuals. · Analyze and try to understand the interests and expectations of the other party. · Prepare for the negotiating table by thinking through what your bottom line is (that is, the highest price you are prepared to pay and the terms of the contract that are not negotiable). · Know the market well. Be aware of alternatives. If possible, pre- pare a draft contract for discussion. · Sell your position. · Where possible, deal with those who have the real decision- making power. · Speak confidently and in a businesslike manner. · Be honest and project honesty. Ploys and tricks may win the negotiation, but contract performance is likely to suffer. · As a team, present a unified stance. · Strive for long-term mutual satisfaction. · Be prepared to make concessions and sacrifices that do not substantially affect the overall objectives but that demonstrate cooperation. Be creative, open minded, and flexible. · Win results, not arguments. · Apply objective standards to assess negotiation results. · Emphasize win/win results of negotiation. Never suggest to the other party that it lost the negotiation. 76 | A Guide to Competitive Vouchers in Health without compromising service quality. One option is to accept the lowest price tendered. Another is to pay all providers a maximum benchmark price. A third is to attempt to negotiate lower prices from providers whose price tendered was close to the benchmark or from those who can offer the scheme something special (for example, a key location). Quality Specifications Technical and human quality specifications should be made clear; if ac- creditation is required (or offered), it should be discussed at this stage, be- fore entering further into the negotiations. Lower- and middle-income countries may not have quality assurance schemes in place to guarantee a high and consistent level of quality of care. Guidance from the ministry of health may be available, but it may be inadequate and not up to date. In this case, policymakers should consider introducing their own accredita- tion scheme, by seeking technical assistance and building on the experi- ence of other countries. The ministry of health may consider that accredi- tation is its prerogative, even if it is not currently offering it. Offering to extend the accreditation scheme to the public service could help assuage doubts and leave the entire health system with added value. Minimum standards of care and patient management protocols should also be agreed on at this stage. Evidence-based management protocols-- selected by experts in the voucher agency or by external consultants-- should be the preferred choice. Some health service providers may not be up to date on these issues. Ownership and Access Rights It is important to establish and agree on ownership of laboratory speci- mens, X rays, and medical records; on who is to store them and for how long; and on who may have access to them in the future. It may be desir- able for the voucher agency to retain intellectual property rights of all data for all studies and publications. Patient confidentiality must be respected at all times, but the voucher agency may want to request access to medical records for supervision or study purposes. If the voucher scheme incorpo- rates a research component involving human subjects, it is essential to ob- tain approval from the appropriate bodies and, if necessary, informed consent from the patients. Implementing a Voucher Scheme | 77 CONTRACTING Contracting is the process of formalizing an agreement that has been reached by negotiation. A written contract gives both parties the opportu- nity to express expectations and agree on outcomes. Sound contract de- sign is a key element in a successful relationship with providers and there- fore a prerequisite to the success of the program. Policymakers should not hesitate to include clauses in the contract that address issues of potential concern, provided that they do not compromise the scheme's flexibility. Box 6-2 provides an outline of the content of a typical contract. Health care providers may not be accustomed to working under written contracts and may find the process threatening. It is therefore important to clarify that the main purpose of contracts is to set out and agree upon a set of rules that suit both parties. A contract should set out not only the provider's duties and commitments but also the purchaser's. For instance, the voucher agency may commit itself to train the provider's staff, to reim- burse providers in a timely manner, to follow up specific cases, and so forth. Contracts can be very simple, and they can evolve and change as the scheme goes on. It may make sense to establish a relatively short contract, especially for the first contract with a new provider, which can be renewed periodically if both sides are satisfied with each other's performance. Al- ternatively, the voucher agency may establish a trial period or make the contract conditional upon satisfaction with the provider's performance in the piloting phase. Once the methodology and logistics have been tried, the service provider may feel more relaxed about committing itself to a written contract, and the voucher agency has a better idea of the suitabili- ty of the provider. TRAINING All new programs involve a training component. In the case of voucher schemes, training can be needed for voucher agency staff, for contracted service providers, and for voucher distributors. What are the training needs at each level? 78 | A Guide to Competitive Vouchers in Health Box 6-2. Elements in a Model Contract 1. The Agreement 1.1 The parties. Specifies who is the purchaser and who is the provider. 1.2 The offer. The provider says that he is prepared to provide the goods or services outlined in the contract (and its annexes) at a certain price. The annexes may include the provider's original offer, submitted as a bid during the tendering process. The offer is signed by the provider 1.3 The acceptance. The purchaser (the voucher agency) agrees to pay the provider according to the terms in the contract. The acceptance may also stipulate when the agreement comes into effect. The acceptance is signed by the voucher agency. 2. General Provisions 2.1 Definitions. Provided for words used as shorthand for longer titles (for example, the purchaser can be defined as whatever organization is purchasing the services) or for terms that may have an ambiguous interpretation (for example, force majeure). 2.2 Law. The parties agree to have the contract governed by the law in a particular jurisdiction. 2.3 Communications. Stipulates the mode (such as in writing) and language of communications to be used between the contracting parties. 3. The Purchaser 3.1 Purchaser's representative. Names the person authorized to represent the purchaser in dealings with the provider. 4. The Provider 4.1 Provider's representative. Names the person authorized to rep- resent the provider in dealings with the purchaser. 4.2 General obligations. Can set down the provider's obligations with regard to the inputs it will need to complete the con- tract and their subsequent ownership. Implementing a Voucher Scheme | 79 Box 6-2. Continued 4.3 Subcontracting. Can be used to restrict providers' rights to subcontract service provision to other entities. 5. Duration of Contract 5.1 Duration. Specifies the date at which the contract ceases to be in force. 5.2 Extension. Specifies the manner in which the duration of the contract can be extended. 6. Remedies 6.1 Remedying errors. Used to assign responsibility (usually to the provider) to make good any faults due to poor equip- ment, consumables or standards of care, and so forth 6.2 Quality control. Gives the purchaser the right to conduct checks on the quality of the services being delivered. The cost is usually assigned to the purchaser unless errors or faults attributable to the provider are uncovered. 7. Variations Establishes procedures and rules for dealing with, for example, patients who do no fit the standard management protocol. Can allow for additional investigation and extra payments to the contractor. Can be used to give the voucher agency the flexibil- ity to change the management protocol. 7.1 Right to vary. Specifies who can make changes to the agreed upon services and under what conditions. 7.2 Value of variations. Specifies what the provider will be paid for variations from the protocol. 7.3 Early warning. Stipulates the warning that the parties must give each other for variation. 7.4 Right to claim. Allows providers to claim for costs due to changes imposed by the voucher agency. 8. Contract prices and payments 8.1 Schedule of fees. Gives the fees that the voucher agency agrees to pay the provider for specific services. continued 80 | A Guide to Competitive Vouchers in Health Box 6-2. Continued 8.2 Payment schedule. States how often the voucher agency will make payments to the provider. 8.3 Payment conditions. Can stipulate conditions the provider must meet before being eligible to receive payment (for example, patient must sign receipt of laboratory results). 9. Default 9.1 Default by the purchaser. What happens in the event the pur- chaser fails to meet its obligations 9.2 Default by the provider. What happens in the event the provider fails to meet its obligations. 10. Risk and Responsibility 10.1 Contractors' responsibilities. Obligations, for example, to treat patients according to accepted ethical standards and to accept the consequences of medical negligence. 10.2 Force majeure. Unavoidable and unpredictable events not due to the other party. Medical misadventure may enter into this definition; the clause should establish what hap- pens in these cases. 11. Resolution of Disputes 11.1 Adjudication. Names someone who can adjudicate a dispute that cannot be settled amicably and sets out how the adju- dication will be performed. 11.2 Notice of dissatisfaction. Gives parties the opportunity to dis- agree with results of adjudication before it becomes bind- ing and to take the process to arbitration. 11.3 Arbitration. Can provide an alternative mechanism for set- tling disputes without recourse to legal action. 12. Appendices Provides detail on the goods or services to be provided (for ex- ample, the detailed patient management protocols). The con- tract should state that these form an integral part of the agree- ment. The appendices can also set out detailed rules for arbitration, adjudication, and other aspects of the contract. Implementing a Voucher Scheme | 81 Training within the Voucher Agency For the voucher agency, the training required will depend on the existing skills mix of the staff. Training should be tailored to the specific tasks the agency is required to perform. These will normally include tendering, ne- gotiating contracts with service providers, monitoring contract perform- ance, designing and running information systems, evaluating the pro- gram, accounting and paying contractors, training providers and voucher distributors, designing promotional and client information materials (in- cluding the voucher itself), and dealing with complaints and queries from voucher recipients. Some of these tasks (such as the design of promotion- al materials) can and probably should be contracted out to organizations that specialize in them. It will often, however, be necessary to have at least one staff member with a good knowledge of the health problem being ad- dressed and the interventions used. Some schemes place restrictions on service providers, requiring that only staff who have taken part in training workshops can care for voucher patients, for example, or that only female staff can conduct gynecological examinations. If other voucher schemes are in operation in the country, it would be beneficial for them to provide some training, as there is no substitute for hands-on experience. Training Service Providers Training for service providers could include instruction on how to fill out the forms correctly and how to return vouchers for payment, as well as ex- planations of management protocols, contractual obligations, and pay- ment procedures. If there are aspects of the management protocol that are novel or technically challenging, it may be necessary to provide training in these areas, too. The training of receptionists should not be overlooked. The reception- ist is usually the first service provider the voucher bearer meets. The valid- ity of the voucher needs to be acknowledged immediately to avoid any embarrassment for the patient. The agency may have negotiated preferen- tial treatment for voucher-bearing patients, which is something the recep- tion staff often control. The training of receptionists is in the interests of the provider, as a bad experience of one patient may deter attendance by others. Emphasizing the importance of treating voucher bearers with re- spect and courtesy and maintaining their privacy can play a key role in preventing voucher bearers from feeling stigmatized. 82 | A Guide to Competitive Vouchers in Health Training Voucher Distributors Voucher distributors are usually the first point of contact between the scheme and the target population. They therefore play a key role in ex- plaining what benefits the voucher offers (and does not offer), in con- vincing the recipient of its legitimacy, and in answering any questions or doubts patients may have about the scheme. Some of the most difficult- to-reach groups (indigenous populations, drug users, commercial sex workers) tend to be suspicious of outsiders and require sensitive han- dling. Social anthropologists, who are trained in these skills, are well- placed to train voucher distributors. It is important to remember that training is an ongoing activity. Staffs change, new service providers and voucher distributors enter the scheme, and policies change. Periodic training is therefore worthwhile. PILOTING Piloting is a crucial step in implementing any program. It is especially im- portant for voucher schemes, where several different parties and a wide range of activities are often involved. Piloting provides a chance to try out all strategies and their logistics and to modify them based on the results. It is important to resist the temptation of excluding small logistical or or- ganizational steps from piloting. All components of a program need not be tested at the same time. Pi- loting of the voucher design, for instance, can be done at the beginning of the program, before any other stage is commenced. The scheme's educa- tional material should also be tested early on, as well as its information systems. Time limits should be set on piloting, so that it does not contin- ue indefinitely in an attempt to perfect the system. Restricting the piloting to a geographic area and to a specific number of vouchers can also help make the process more efficient. Piloting can result in the exclusion of certain providers and the inclu- sion of more appropriate ones. It can also result in dramatic changes in strategies. Such change need not be viewed as a failure but as a chance to rethink different processes to ensure the program's success. The example in box 6-3 describes how piloting led to a significant change in the design of a voucher when a scheme from one country was tested in another. Implementing a Voucher Scheme | 83 Piloting is not complete until the results have been formally assessed and acted on. A time needs to be set aside to consider which aspects of the scheme's design worked and which did not. This is the time to trans- late poorly performing strategies into changes to the scheme. It is vital to resist committing the program to a fixed set of strategies, even if the fund- ing agency applies pressure to do so. Once again, flexibility is the key to success. Box 6-3. How Piloting Can Lead to Changes in Strategy The El Salvador cervical cancer program was to implement many of the same strategies successfully tried out in the Nicaraguan pro- gram, including the advertising material. However, when the pro- ject's logo was tested in El Salvador, program designers discovered that what was considered feminine and discrete in Nicaragua was interpreted as offensive in El Salvador. The nude woman in the Nicaraguan logo became a dressed woman in the Salvadoran logo. Original logo in Nicaragua Logo adapted for El Salvador The voucher distribution in El Salvador was going to make use of traditional birth attendants, in the hope that their relationship with older women would facilitate the distribution. However, when this component of the program was pilot tested, traditional birth attendants managed to distribute only a handful of vouch- ers. The strategy of voucher distribution had to be rethought. Currently, vouchers are being distributed by community health workers. 84 | A Guide to Competitive Vouchers in Health INSTITUTIONALIZING Because of their novelty, voucher schemes are often introduced with sig- nificant levels of external technical assistance. Institutionalizing is about making the transition from an experimental initiative, or "project," to a sustained, ongoing health program. Part of this institutionalization process is about weaning off of technical assistance. Another part is about establishing norms for the various activities (such as contracting new clin- ics or paying for redeemed vouchers) that are repeated. A third element is a conscious search for ways to minimize ongoing costs. Experience has shown that the institutionalization phase must be tak- en as a conscious step. It is otherwise all too easy to allow the ad hoc ap- proach that typifies (and indeed helps) new initiatives to continue indefi- nitely. Although it is not always possible, ensuring that trained staff stay on the job is a key element to ensure a smooth transition, even if it means increasing their salaries. A danger in institutionalizing a program can be the fact that, in an ef- fort to simplify and streamline all aspects of the scheme, the flexibility of the program is compromised. Remaining project staff must be left some room to maneuver, conceptually and operationally, in order to respond adequately to new threats and opportunities faced by the voucher scheme. Retaining access to ongoing but occasional technical assistance in these instances could be budgeted for. SCALING UP Scaling up is a process that can occur at least twice in the implementation of a voucher scheme. The first occasion occurs when the pilot testing ends and the project is applied to the full target population. This full imple- mentation phase does not necessarily imply nationwide coverage. Vouch- er schemes often begin as small innovative projects aimed at serving the needs of relatively small and geographically limited population groups. If the scheme is perceived as having been successful, there is therefore often a subsequent stage or stages of scaling up to cover a much wider propor- tion of the population--by expanding either geographically or in scope to cover new target groups. Scaling up can greatly improve the efficiency of a voucher scheme, since there are certain relatively fixed administrative costs inherent in the use of Implementing a Voucher Scheme | 85 vouchers, such as the need for a voucher agency. But when voucher schemes begin as small projects, they often start up in the most favorable settings, where there are a wide range of potential service providers willing to compete for contracts. As they expand, the settings may become less conducive to competitive schemes, the costs of transport and communi- cations tend to rise, and the quality of services may decline. These factors can easily undermine the efficiency gains of greater scale. As the scale increases, the number of institutions that become involved tends to increase. Until voucher schemes form part of the orthodoxy of health development assistance, merely securing agreement to introduce a voucher scheme can meet resistance. The distribution of tasks within the voucher agency tends to become more specialized as a scheme is scaled up. Although this may mean that specific tasks are performed more efficiently, there remains a need for someone--perhaps the scheme's coordinator--to retain a broad vision of how the scheme is operating. As the program expands, some of the systems initially established by the voucher agency--computer databases, communications, contract ne- gotiation skills, transport--may begin to fail as they reach the limits of their capacities. These problems can be overcome by upgrading and ex- panding technologies and training, but the process of doing so can be time consuming. CHAPTER 7 MONITORING AND EVALUATING A VOUCHER PROGRAM Monitoring is the process of routinely gathering information to determine whether something is meeting expectations. Evaluation, in its broadest sense, is the process of assessing something or making a judgment. House (1980) sees evaluation as a process that leads to a judgment about the worth of something, a settled opinion that usually leads to a decision to act in a certain way. Clearly, there is overlap between monitoring and eval- uation, but a distinction can be made. Monitoring is an ongoing activity, whereas evaluation is carried out periodically. Evaluation necessarily in- volves judgments, whereas monitoring can be merely observation. Moni- toring is normally carried out by people responsible for running the sys- tem. Evaluation is often conducted by people who do not have direct responsibility for running the system. 87 88 | A Guide to Competitive Vouchers in Health Monitoring and evaluation of the voucher scheme begin before the first voucher is distributed. It starts at the design stage, by careful consideration of what activities and outputs need to be monitored, what information will be needed to monitor them, and how that information can be gath- ered in the most economical and timely fashion without compromising quality. One of the beauties of voucher schemes is the ease with which moni- toring and evaluation can be performed. This is because the vouchers themselves define discrete units for which individual processes and out- comes can be traced and measured. Monitoring and evaluation involve the following tasks: · Determine what aspects of the voucher scheme need to be monitored. · Determine how and when these aspects should be monitored. · Establish the necessary systems for conducting the monitoring. · Conduct the monitoring. · Determine the basis upon which the voucher scheme should be evalu- ated. · Determine how and when evaluation should be carried out. · Conduct the evaluation. Figure 7-1 lists the eight areas that should be addressed in developing monitoring systems and evaluating the scheme. They are not presented in any particular order. MONITORING SERVICE QUALITY Quality includes technical quality, which affects the scheme's ability to improve health, and human quality, which relates to patients' satisfaction with the services rendered. In both cases, monitoring involves three steps: setting and reviewing verifiable standards, measuring quality, and review- ing results. In setting standards, policymakers have at least three options. One is to allow providers to set the standards themselves. A second is to impose ex- ternally set standards on providers. The third is for the voucher agency and providers to negotiate and agree on a set of standards. The advantage of Monitoring and Evaluating a Voucher Program | 89 Figure 7-1. Monitoring and Evaluating the Voucher Program Monitoring costs Monitoring service quality Monitoring competition between providers Monitoring to detect abuse of the voucher scheme Monitoring the characteristics of voucher recipients and redeemers Monitoring and evaluating health outcome Monitoring and evaluating the impact on equity and poverty reduction Monitoring and evaluating cost-effectiveness the first strategy is that it gives providers a stake in the process, which can make them more receptive to improving quality. The advantage of the sec- ond option is that standards can be chosen that are known to be adequate for achieving the expected health gains or patient satisfaction. The third strategy, which is one that voucher schemes and other contract-based pro- grams lend themselves to, should have the advantages of both of the first two options. Monitoring Technical Standards To be objective, the measurement of quality must be neutral and well publicized. If the voucher scheme incorporates accreditation, the process needs to be repeated periodically. Certain health services need more 90 | A Guide to Competitive Vouchers in Health supervision than others, particularly those for which performance is known to vary widely (such as sputum microscopy for tuberculosis and cervical cy- tology). One policy is to require accreditation of all new providers. For lab- oratory-based care, samples can be regularly doubled-checked at a different institution and comparisons of performance made over time. Monitoring the Human Quality of Care Measurement of the human quality of care poses less of a problem. Pa- tients can bear testimony to the quality of care they receive. They can pro- vide objective information (such as waiting times) and subjective indica- tors of satisfaction. Patients can be interviewed as they leave a facility (sometimes it may even be sufficient to interview non-voucher-bearing pa- tients), but the presence of an interviewer outside the clinic may annoy the provider or change his or her treatment of patients. Tracing patients at the point at which their vouchers were originally distributed may be a better strategy and one that allows people who had not used their vouchers to be interviewed. Respondents who have used the voucher are normally willing to share their experience and take the time to participate in the evaluation. "Mystery Patients" Another alternative for monitoring human quality (and even some as- pects of technical quality) is the use of "mystery patients." These are peo- ple employed by the voucher agency who pose as voucher-bearing pa- tients. Providers can be warned in advance that they can expect such visits as part of quality monitoring (and the use of mystery patients can be writ- ten into the contract). Mystery patients can report back on all aspects of care: whether the premises were clean and tidy, the receptionist was polite, waiting times were acceptable, doctors and nurses discriminated against the patient, the procedures performed were fully explained, and so forth. Mystery patients can be very useful if given thorough guidance in advance, but their ability to assess the technical quality of care is usually quite lim- ited, unless they are health professionals themselves. Field Visits and Facility Inspections Voucher agency staff should make periodic unannounced visits to the health facility and speak with staff to gather their views. These visits can be Monitoring and Evaluating a Voucher Program | 91 used to ensure that all equipment is still available and working properly and that forms are being filled out correctly. The staff member can accom- pany a voucher bearer to observe the full process of care. Interpreting Findings Random fluctuations and normal error in measurement can account for some variation in results. One should therefore take action only when these results lie outside the limits of what might be expected as a result of such variation. The frequency with which one reviews the quality indica- tors is a matter of judgment. If monitoring is done too infrequently, one may miss some serious problems with service quality that could affect the impact of the scheme. Monitoring too frequently increases the adminis- trative burden, however, as well as the random variation. MONITORING COMPETITION AMONG PROVIDERS One of the more useful process indicators that can be monitored is the de- gree of competition in the scheme. If there is little competition among providers, many of the benefits of the voucher scheme, such as improved quality and cost containment, may not materialize. The degree of compe- tition within the voucher scheme will change over time, as providers enter and leave and as providers' shares of the total number of vouchers change. A widely accepted measure of the degree of competition is the Herfin- dahl-Hirschmann index (HHI) of concentration (box 7-1). HHI levels above 1,000 indicate moderate market concentration; HHI levels above 1,800 are considered highly concentrated (U.S. Department of Justice and Federal Trade Commission 1997). One must use these indices with some caution, for several reasons. First, there may be competition simply to enter the scheme, but providers who are not selected do not enter into the calculations. Second, even with a large number of providers, each with a small market share, competition may not be perfect. Geographic location, to name just one factor, will also affect competition, because voucher bearers may prefer to go to a clinic of worse quality if it means less traveling. Third, in fully subsidized schemes, providers are not competing on price for their share, since the price has al- ready been negotiated with the voucher agency. Notwithstanding these limitations, the HHI does illustrate the importance to competition of hav- 92 | A Guide to Competitive Vouchers in Health Box 7-1. Calculating the Herfindahl-Hirschmann Index of Concentration The Herfindahl-Hirschmann index is calculated using the follow- ing four steps: 1. Write down the names of all providers in the first column. 2. In the next column, write down the proportion of the total num- ber of vouchers redeemed with each provider during the period monitored. 3. In the third column, write down the square of each of those per- centages. 4. Add up all of the numbers in the third column. This is the HHI. Sample Calculation of the HHI of Concentration Clinic Share of vouchers (percent) Score A 30 900 B 25 625 C 20 400 D 15 225 E 10 100 HHI 2,450 ing as many providers as possible and a relatively even distribution of market share. This should be borne in mind when making decisions about adding or dropping providers from the scheme. MONITORING TO DETECT ABUSE OF THE VOUCHER SCHEME Voucher schemes can be abused or misused in a variety of ways, including counterfeiting, collusion between providers and voucher bearers or dis- tributors, black market sales, bribery and kickbacks, overservicing, Monitoring and Evaluating a Voucher Program | 93 provider moral hazard, adverse selection, and cream-skimming (cherry- picking). Counterfeiting The risk of counterfeiting and development of a black market for vouchers is highest if the vouchers have a high monetary value or if they subsidize a highly sought service. This risk should have been considered during the design phase and measures taken to keep it to a minimum. However, in some cases it may be better (and cheaper) simply to monitor the scheme carefully to detect abuse or misuse as soon as it occurs. If a computerized information system has been developed, it can auto- matically flag redeemed vouchers that have not yet been printed or dis- tributed (or ones for which the check digits do not tally). In addition, the system can match the voucher number with the geographic area in which it was distributed or with the name of the patient to whom it was given (if this information was collected and recorded). It may also be desirable to monitor for counterfeiting by providers. It is sensible to share at least the obvious anti-counterfeiting measures with providers and ask them to check vouchers for these. The less obvious checks for counterfeiting should be kept secret by the voucher agency for its own use. This is perhaps most important where the voucher has little or no monetary value, since the only people who could then readily convert them into cash would be serv- ice providers. Collusion A slightly different situation arises when there is collusion between the voucher distributor and health service providers: voucher distributors may give in to the temptation to come to an agreement with one or more serv- ice providers to give them vouchers in exchange for part of the earnings. The service provider could then ensure that all (or the majority) of the vouchers are "redeemed" at his or her clinic and could return them with- out providing services. This fraud may be detected if there is a sudden change in the proportion of vouchers returned from a particular provider. Visiting and interviewing a proportion of patients can also uncover such scams. It is unlikely, however, that many health service providers would stoop to such actions. Apart from the ill repute they would earn if caught-- as well as the possible criminal charges--they would still have to ensure 94 | A Guide to Competitive Vouchers in Health that the names of recipients submitted by the voucher distributors match those on the forms they send to the voucher agency. Inventing patients re- quires even more work, not to mention the difficulty of obtaining labora- tory specimens if these are required. Black Markets If the benefits provided by the voucher are of high value and the vouchers are widely demanded, a black market for them may develop. It may be worth monitoring to detect the existence and size of such a market. Until it appears, or unless it becomes significant in scope, it may not be neces- sary to do more than monitor the situation. If black market trading be- comes a serious problem, some of the other measures to enforce non- transferability may be necessary (see box 5-2). Printing "no monetary value" on the voucher, as well as mentioning that in any educational or social marketing campaign, may also help. Dishonesty within the Voucher Agency The voucher agency or its staff may receive bribes in return for contracting with certain providers or agreeing to higher than necessary voucher pay- ments. The risk of this form of abuse can be reduced by having strict rules regarding contracting with providers and by setting a single price for all providers. Such rules can undermine the flexibility of the voucher scheme, but they may be necessary to yield adequate transparency. An important monitoring measure that should be taken is external financial audit. Overservicing Where the voucher agency pays providers different sums based on patient characteristics, providers may be tempted to shift patients from lower- to higher-paying categories, even if it means providing services the patients do not merit according to the management protocol. This possibility, called "moral hazard" (or "overservicing"), is well-known to health insur- ance schemes. Monitoring costs is one way of detecting or at least screen- ing for it. Another way to monitor for provider moral hazard is to period- ically review a sample of cases from each provider that qualified for the higher payments, to ensure that it meets the established criteria. Monitoring and Evaluating a Voucher Program | 95 Adverse Selection Adverse selection is most likely to occur where vouchers incorporate some element of risk spreading. For example, if a voucher covering the cost of both screening and treatment for a disease is sold, those who are free of the disease usually subsidize those who are not. This means that there is an incentive for patients diagnosed elsewhere to purchase a voucher and receive the treatment at a subsidized cost. This raises the average cost per voucher and can make the screening and treatment vouchers uncompeti- tive with screening alone. Thus individuals will tend to pay to be tested elsewhere and will buy the voucher only if they are found to be positive. The result can be a collapse of the scheme. To avoid this eventuality, indi- viduals who know that they have a disease could be prohibited from pur- chasing the voucher, but in many cases it is impossible to enforce. What could help in this case is making it clear to the voucher distributor (or sell- er) that vouchers should not be given away to people who are known to have the disease. If the main aim of the voucher scheme is to detect and treat a disease, to subsidize the poor, or both, adverse selection need not be a major con- cern. Indeed, it may represent an improvement in technical efficiency if those most at risk of a disease are most likely to use their voucher. Where diagnosis outside the voucher scheme is not reliable, adverse selection be- comes less of a risk, because many of those seeking prepaid treatment may not actually have the disease and will therefore not require treatment (see box 7-2). Cream-Skimming (Cherry-Picking) Cream-skimming, sometimes known as cherry-picking, is a form of provider adverse selection in which voucher bearers whose costs are like- ly to exceed the payment received from the voucher agency are excluded or obstructed from redeeming their vouchers while patients whose costs are likely to be lower are actively encouraged to redeem their vouchers. Instead of competing to attract voucher-bearing patients, providers com- pete to attract low-cost patients and off-load high-cost patients onto oth- er providers. Cream-skimming is relatively easy to detect by examining the profile of patients at each of the clinics in relation to the services (and hence costs) they receive from the providers. It can be addressed by weighting pay- 96 | A Guide to Competitive Vouchers in Health Box 7-2. Improving Targeting through Adverse Selection A scheme in Nicaragua sold vouchers with prepaid screening for and treatment of cervical cancer. The scheme allowed women who had previously been diagnosed by Pap smear as having a high- grade lesion to purchase the voucher, because the quality of smear diagnosis outside the scheme was generally so poor that many of these patients did not actually have these lesions. However, since there was considerable variation in the quality of cytological screening outside the scheme, some patients did benefit from ad- verse selection. If the quality of screening nationwide rose to the standards within the program (which is possible, since it initiated an external quality assurance scheme for all cytologists), then ad- verse selection may become more of a concern. In contrast, for ful- ly or highly subsidized schemes, adverse selection helps improve targeting and therefore the program's efficiency at achieving its aims. ments according to the patient profile (individually or across the facility as a whole), but care must be taken to ensure that this differential payment strategy does not create provider moral hazard. MONITORING THE CHARACTERISTICS OF VOUCHER RECIPIENTS AND REDEEMERS Knowing who received the vouchers is useful for measuring the efficiency of the distribution system at targeting priority groups. This information can also be used to identify the best distributors. Monitoring can reveal who distributed the most vouchers and who best targeted the priority groups. Together with information on who has redeemed vouchers, this information can reveal who has the highest redemption rates. Distributors who distribute the most vouchers may not be the ones who have the most vouchers redeemed at service providers. The difference may depend on how good they are at explaining what the vouchers are for and why they Monitoring and Evaluating a Voucher Program | 97 should be used. This information can be used as the basis for providing incentives to distributors. For instance, a prize or special recognition could be offered to the distributor with the highest redemption rate among pri- ority groups. Monitoring these characteristics can also serve another important pur- pose. It is often valuable to understand why some people choose not to use their vouchers. This knowledge can help the voucher agency design strategies to improve redemption rates. Small changes in the way the scheme is designed or promoted may greatly improve redemption rates. In particular, the voucher agency will want to know whether recipients are fully aware of the benefits the voucher offers. Of course, it may be that nonredeemers are a self-selected subgroup of low-priority individuals. Qualitative as well as quantitative data must be collected to understand why recipients are not redeeming their vouchers. Keeping a record of the names and addresses of voucher recipients makes it possible to go back to interview samples of those who chose not to redeem their vouchers and thus identify any unforeseen barriers to service uptake. MONITORING AND EVALUATING HEALTH OUTCOMES Outcomes include reducing inequalities in health, compensating for the presence of externalities, and addressing the distortions caused by incom- plete and asymmetrical information. In the case of externalities, the prob- lem is often one of underconsumption of services that (for example) lead to lower population growth rates or reduce the transmission of commu- nicable disease. The impact of subsidies should therefore be measured in terms of how well they achieve specific outcomes, such as lower birth rates and transmission of infectious disease. Program evaluation is often complicated by the difficulty of determin- ing whether changes in outcome indicators are attributable to the pro- gram or to other extraneous factors. In this regard, voucher schemes have some advantages over other ways of delivering subsidies, because one can pinpoint who has been a beneficiary of the program and who has not. It therefore becomes relatively simple to establish what goods or services each beneficiary (or at least a sample of them) received. Evaluating the scheme's impact becomes a matter of determining what the outcome would have been for these individuals had they not received vouchers. Three different approaches can be used. One is to estimates how these in- 98 | A Guide to Competitive Vouchers in Health dividuals would have fared had the voucher scheme not been introduced. Another is to compare outcomes among voucher recipients and nonrecip- ients. The third is to ask voucher recipients what they would have done had the voucher scheme not existed and to speculate on what the out- come would have been. None of these approaches is perfect, but all give a better indication of impact than merely measuring outcomes among voucher recipients alone. Subsidies are justified for the treatment and prevention of communica- ble disease because avoiding or reducing the duration of an infection in one person reduces the risk of others contracting the disease. This makes it difficult to determine the true impact of subsidies. It is difficult to know how many infections are avoided for every one treated or how many are prevented by immunization among those not immunized because of herd immunity. Sophisticated mathematical models are being developed to help answer these questions, but they are not yet used regularly in pro- gram evaluation. In practice, one is often limited to quantifying the im- pact on voucher recipients and making assumptions about the down- stream benefits of these outcomes. MONITORING AND EVALUATING THE IMPACT ON EQUITY AND POVERTY REDUCTION Ideally, one would like to know the number or proportion of voucher re- cipients whom the scheme protects from becoming impoverished or the reduction in income inequalities in health outcomes. This is known as poverty impact analysis. In practice, these outcomes can be very difficult to measure because of the need to simultaneously measure changes in the health status of subgroups within a population as well as within the pop- ulation as a whole and to demonstrate that differences among subgroups were caused by the program. Benefit incidence analysis seeks to quantify how program subsidies are distributed across socioeconomic classes.1 It seeks to find out how well the voucher scheme has targeted resources to the poor and whether it has done so more effectively than existing subsidy schemes. The first of these aims can be achieved by documenting the socioeconomic profile of The term "incidence" is used here in a quite different sense from that used in epidemiol- 1 ogy, where it is a measure of the frequency of occurrence of a disease event. Monitoring and Evaluating a Voucher Program | 99 voucher scheme beneficiaries and comparing it with the socioeconomic profile of the general population. The second requires data on the exist- ing benefit incidence of government spending on health. These data may be obtained by documenting the socioeconomic profile of a sample of health service users weighted by spending on services to each (for exam- ple, hospital users consume a much higher level of government subsidies than ambulatory care users and may have quite a different socioeconom- ic profile). An even simpler way of evaluating a program is to examine the service outputs (for example, receipt and utilization of vouchers) in terms of the socioeconomic profile of the beneficiaries. Doing so avoids the need to quantify the value of these outputs in terms of the level of subsidy going into each. If the same socioeconomic indicators have been measured in other subsidy schemes or for government health expenditure generally, a judgment can be made about the equity impact of the scheme. Irrespective of the type of analysis, monitoring and evaluation of equi- ty must compare differences in outcomes by socioeconomic status. With- out a consistent summary measure, confusion can arise, because differ- ences or changes over time can be caused by either variation in the socioeconomic status of the populations being compared or by variation in the distribution of health outcomes among them. This issue can be ad- dressed by using appropriate summary measures, such as concentration coefficients or the index of relative inequality (Wagstaff, Paci, and Van Doorslaer 1991). Some measurement of income or socioeconomic status among pro- gram participants is required. If one wishes to make comparisons with the population as a whole, the measures or indicators of socioeconomic sta- tus should be ones that have been applied in censuses or surveys nation- wide. Income is perhaps the best indicator of socioeconomic status, but measuring it in developing countries is fraught with difficulties and rarely practical. Measuring household expenditure is often more reliable, but it is also too time consuming for routine data collection. A better alternative is to record the same proxy indicators of socioeconomic status used in censuses and representative population surveys. These may include educa- tional level, house construction materials, number of household mem- bers per bedroom, and certain key household assets. These data can be collected either at the time the vouchers are distrib- uted or at the point of service provision. The advantage of collecting the data when the vouchers are distributed is that it permits the socioeco- 100 | A Guide to Competitive Vouchers in Health nomic status of voucher redeemers and nonredeemers to be compared. Moreover, if distribution is done house to house, the distributor can ob- serve or verify house construction materials and asset ownership, if these are used as indicators. People may be suspicious or embarrassed to dis- close details regarding their housing or income status to a young person who may be one of their neighbors, however. MONITORING AND EVALUATING COST-EFFECTIVENESS Evaluation of cost-effectiveness of the program is one of the most diffi- cult enterprises in public health. Vouchers simplify this process, however, often making it possible to accurately measure both costs and project- specific outcomes. If costs and outcomes are monitored, monitoring cost- effectiveness becomes simply a matter of calculating the ratio of positive outcomes to the cost of the program. Problems arise, however, when one wishes to compare alternative subsidy schemes or the status quo, be- cause accurate information on costs and outcomes is much harder to come by. REFERENCES Angrist, J., E. Bettinger, E. Bloom, E. King, and M. Kremer. 2001. Vouchers for Private Schooling in Colombia: Evidence from a Randomized Natural Experiment. Boston: National Bureau of Economic Research. --. 1995. "Sexual Health and STDs: An Avenue to HIV Prevention Services." AIDS Information Exchange (U.S. Conference of Mayors) 12: 6­8. Astin, A.W. 1992. "Educational `Choice': Its Appeal May Be Illusory." Sociology of Education 65(4): 255­260. Bennell, P. 1997. Vocational Education and Training in Zimbabwe: The Role of Private Sector Provision in the Context of Economic Reform. Institute of Development Studies Research Report, University of Sussex, United Kingdom. Bertsch, E.F. 1992. "A Voucher System that Enables Persons with Severe Mental Ill- ness to Purchase Community Support Services." Hospital Community Psychiatry 4311: 1109­1113. Birkhead, G.S, C.W. Lebaron, P. Parsons, J.C. Grabau, L. Barr-Gale, J. Fuhrman, S. Brooks, J. Rosenthal, S.C. Hadler, and D.L. Morse. 1995. "The Immunization of Children Enrolled in the Special Supplemental Food Program for Women, In- fants, and Children (WIC): The Impact of Different Strategies." Journal of the American Medical Association 2744: 312­316. Blanck, P., L. Clay, J. Schmeling, M. Morris, and H. Ritchie. 2002. "Applicability of the ADA to `Ticket to Work' Employment Networks." Behav Sci Law 206: 621­636. BONOSOL. 2003. Available at www.Bonosol.Bo. Bono ISAPRE 2003. Available at www.andueza.Cl. Bono AUGE 2003. Available at www.Fonasa.Gov.Cl. Borghi, J., A. Gorter, P. Sandiford, and Z. Segura. 2003. "The Cost-Effectiveness of a Voucher Scheme to Reduce Sexually Transmitted Infections in Sex Workers and Their Clients: The Case of Managua, Nicaragua." Paper presented at the Fourth World Congress of the International Health Economics Association, San Francisco. 101 102 | A Guide to Competitive Vouchers in Health Bosselaar, H., and R. Prins 2001. Return to Work of Disabled Persons: The Dutch Rein- tegration Voucher. Meccano, Netherlands: European Employment Observatory. Calero, C. 2003a. El Salvador: Cervical Cancer Prevention for Poor Rural Women. In- stituto Centro Americano de la Salud. Managua, Nicaragua. --. 2003b. Questionnaire for the World Bank Study on Voucher Schemes for Health. In- stituto Centro Americano de la Salud. Managua, Nicaragua. Caram, M. 2002. "Voucher Program on Reproductive Health: Profamilia Project Report." Profamilia. Santo Domingo, Dominican Republic. Cernada, G., and L.P. Chow. 1969. "The Coupon System." American Journal of Pub- lic Health 59 (12): 147-166. --. 1970. "The Coupon System." In The Taiwan Family Planning Reader, ed. George P. Cernada, 147­166. Taichung: Chinese Center for International Training in Family Planning. Cohen, D.K., and E. Farrar. 1977. "Power to the Parents? The Story of Education Vouchers." Public Interest 48 (1): 72­97. Culyer, A.J., J.W. Posnett, A.J. Culyer, A. Maynard, and J. Posnett, eds. 1990. Com- petition in Health Care: Reforming The NHS. Vol. 1, Hospital Behavior and Compe- tition. London: Macmillan. Dole, G.F., and P.A. Carney. 1998. "Characteristics of Underserved Women Who Did and Those Who Did Not Use A Free/Low-Cost Voucher as Part of a Mam- mography Screening Program." Journal of Cancer Education 13 (2): 102­107. Du, K., K. Zhang, and S. Tang S. 1999. "Summary of `A Draft Report on a MCHPAF Study in China.'" Poverty Net Library. Available at poverty.worldbank.org/li- brary/. Friedman, M. 1962. Capitalism and Freedom. Chicago: University of Chicago Press. Friedman, J., and D.H. Weinberg. 1982. The Economics of Housing Vouchers. New York: Academic Press. Galasso, E., M. Ravallion, and A. Salvia. 2001. "Assisting the Transition form Work- fare to Work Argentina's Proempleo Experiment." Poverty Net Library. Avail- able at poverty.worldbank.org/library/. Gauri, V. 1998. School Choice in Chile: Two Decades of Educational Reform. Pitts- burgh, PA: University of Pittsburgh Press. Gauri, V., and A. Vawda. 2003. "Vouchers for Basic Education in Developing Coun- tries, A Principal-Agent Perspective." Policy Research Working Paper 3005, World Bank, Washington, DC. References | 103 Gorter, A. 2002. A Voucher Scheme for Adolescents in Nicaragua to Improve the Uptake of Reproductive Health Services. Project report for DfID (Department for Interna- tional Development), Instituto CentroAmericano de la Salud, Managua, Nicaragua. --. 2003. "Review Paper on Evidence for Using Competitive Voucher Schemes and Related Methods for Ensuring Young People Have Access to Health Service In- terventions Designed to Prevent or Provide Care for HIV/AIDS." Background paper prepared for the Consultation on the Health Services Response to the Prevention and Care of HIV/AIDS among Young People, organized by WHO with UNICEF, UNFPA, UNAIDS, and Youthnet, Montreux, Switzerland. Gorter, A., Z. Segura, and P. Sandiford. 2002. "Competitive Voucher Schemes for Better Health for Vulnerable Populations and Poor." Invited clinic and presen- tation at the Discussion Forum on Private Provision of Health Services in De- veloping Countries, World Bank, Washington DC. Green, J., W. Howell, and P. Peterson. 1997. "An Evaluation of the Cleveland Scholarship Program." Occasional Paper, Harvard University, Program in Edu- cation Policy and Governance, Cambridge, MA. Available at http://www. schoolchoices.org/roo/cleveland1.htm. Green, J.P., P.E. Peterson, and J. Du. 1997. "Effectiveness of School Choice: The Milwaukee Experiment." Occasional Paper 97, Harvard University, Program in Education Policy and Governance, Cambridge, MA. Gwatkin, D.R. 2000. "The Current State of Knowledge about Targeting Health Pro- grams to Reach the Poor." World Bank, Washington, DC. Hammett, T.M., D.C. Des Jarlais, W. Liu, D. Ngu, and N. Son. 2002. "A Cross-Bor- der HIV Prevention Intervention for Injection Drug Users (IDUs) in Ning Ming Country Guangxi Province, China and Lang Son Province, Vietnam." Paper pre- sented at the 14th International AIDS Conference, Barcelona, Spain. Higgins, S.T., A.J. Budney, W.K. Bickel, F.E. Foerg, R. Donham, and G.J. Badger. 1994. "Incentives Improve Outcome: Outpatient Behavioral Treatment of Co- caine Dependence." Arch Gen Psychiatry 51: 568­576. Higgins, S.T., S.M. Alessi, and R.L. Dantona 2002. "Voucher-Based Incentives: A Substance Abuse Treatment Innovation." Addict Behav 276: 887­910. House, E.R 1980. Evaluating with Validity. Beverly Hills, CA: Sage. House of Commons. 1996. Proceedings of the British Parliament. November 13. London. Hsieh, C.­T., and M. Urquiola. 2002. "When Schools Compete, How Do They Compete? An Assessment of Chile's Nationwide School Voucher Program." 104 | A Guide to Competitive Vouchers in Health Occasional Paper No. 43, National Center for the Study of Privatization in Ed- ucation, New York. Kent County Council Education Department. 1978. Education Vouchers in Kent: A Feasibility Study for The Education Department of The Kent County Council. Kent County Council, Maidstone, United Kingdom. Kiefe, C.I., S.V. Mckay, A. Halevy, and B.A. Brody. 1994. "Is Cost a Barrier to Screen- ing Mammography for Low-Income Women Receiving Medicare Benefits? A Randomized Trial." Arch Intern Med 15411: 1217­1224. King, E., P. Ozarem, and D. Wohlgemuth. 1998. "Central Mandates and Local In- centives: The Colombia Education Voucher Program." Paper No. 6, Working Paper Series on Impact Evaluation of Education Reforms, World Bank, Devel- opment Research Group, Washington, DC. Knowles, James C. 2000. Consultant's Report of Technical Assistance Provided to the BDD Sustainability Component of the Safe Motherhood Project. Draft consultant's report to the World Bank for the Indonesian Safe Motherhood Project. Koumans, E.H., K. Barker, M. Massanga, R.V. Hawkins, P. Somse, K.A. Parker, and J. Moran. 2003. "Patient-Led Partner Referral Enhances Sexually Transmitted Disease Service Delivery in Two Towns in the Central African Republic." Int J STD AIDS 106: 376­382. Luft, H.S. 1984. "On the Use of Vouchers for Medicare." Milbank Memorial Fund Quarterly / Health and Society, 62, No.2: 237­250. Marchant, T., J.A. Schellenberg, T. Edgar, R. Nathan, S. Abdulla, O. Mukasa, H. Mponda, and C. Lengeler. 2002. "Socially Marketed Insecticide-Treated Nets Improve Malaria and Anemia in Pregnancy in Southern Tanzania." Trop Med Int Health 72: 149­58. Migrant Health Programs Dick Bohrer Paper. 2001. Observations and Recommenda- tions on Migrant Health Programs to the Office of Migrant Health: Version 5 Draft Incorporating the Observations, Critiques, and Contributions of the MHP Action Team. Office of Migrant Health, Bureau of Primary Health Care, Bethesda, Maryland. Mookherji, S. 2003. Questionnaire for the World Bank Study on Voucher Schemes for Health. Instituto Centro Americano de la Salud, Managua, Nicaragua. Mushi, A., J. Armstrong Schellenberg, H. Mponda, and C. Lengeler. "Targeted Sub- sidy for Malaria Control with Treated Nets Using a Discount Voucher System in Southern Tanzania." Health Policy and Planning, 18: 163-171. Oberlander, J. 1998. "Remaking Medicare: The Voucher Myth." Int J Health Serv 281: 29­46. References | 105 O'Donnell, L.N., A. Sandoval, R. Duran and C. O'Donnell. 1995. "Video-Based Sexually Transmitted Disease Patient Education: Its Impact on Condom Acqui- sition." American Journal of Public Health 85: 817­822. O'Neill, D.M 1977. "Voucher Funding of Training Programs: Evidence from the GI Bill." Journal of Human Resources 124: 425­45. Patrinos, H., and D. Ariasingam 1997. Decentralization of Education: Demand-Side Financing. World Bank, Washington, DC. Robey, B. 1987. "Community-Based Contraceptive Distribution: A Korean Success Story." Asia-Pacific Population and Policy 4: 1­4. Roda, A. 2003. Questionnaire for the World Bank Study on Voucher Schemes for Health. Instituto Centro Americano de la Salud, Managua, Nicaragua. Root, G. 2003. Questionnaire for the World Bank Study on Voucher Schemes for Health. Instituto CentroAmericano de la Salud, Managua, Nicaragua. Sanderson, D., M. Place, and D. Wright. 2000. Evaluation of the Powered Wheelchair and Voucher Scheme Initiatives. York Health Economics Consortium, York: UK. Sandiford, P., A. Gorter, and M. Salvetto. 2002a. "Use of Voucher Schemes for Out- put-Based Aid in the Health Sector in Nicaragua: Three Case Studies." Paper presented at the World Bank Workshop on Output-Based Aid, Frankfurt, Janu- ary 24­26. --. 2002b. "Vouchers for Health: Using Voucher Schemes for Output-Based Aid." Public Policy for the Private Sector, Viewpoint No. 243. Sandiford P., M. Salvetto, Z. Segura, and A. Gorter 2000. "Clinics for Sex Workers in Managua." In Public Services through Private Enterprise: Micro-Privatization for Improved Delivery, ed. M. Harper. Pages 228-244. London: IT Publications. Seidman, S.N., C. Sterk-Elifson, and S.O. Aral. 1994. "High-Risk Sexual Behavior among Drug-Using Men." Sexually Transmitted Diseases, May-June, 21(3): 173­180. Silverman, K., M.A. Chutuape, G.E. Bigelow, and M.L. Stitzer. 1996. "Voucher- Based Reinforcement of Attendance by Unemployed Methadone Patients in a Job Skills Training Program." Drug Alcohol Depend 413: 197­207. Simanowitz, A., B. Nkuna, and S. Kasim. 2000. "Overcoming the Obstacles of Identifying the Poorest Families: Using Participatory Wealth Ranking, the CASHPOR House Index, and Other Measurements to Identify and Encourage the Participation of the Poorest Families, Especially the Women of Those Fam- ilies." Paper presented at the Microcredit Summit Meeting of Councils June 24­26. 106 | A Guide to Competitive Vouchers in Health Shaw, R.P. 1999. "New Trends in Public Sector Management in Health." Paper pre- pared for the World Bank Institute's Flagship on Health Sector Reform and Sus- tainable Financing, Washington, DC Skaer, T.L., L.M. Robinson, D.A. Sclar, and G.H. Harding. 1996." Financial Incen- tive and the Use of Mammography among Hispanic Migrants to the United States." Health Care Women Int 174: 281­291. Skibiak, J.P., M. Chambeshi-Moyo, and Y. Ahmed. 2001. Testing Alternative Chan- nels for Providing Emergency Contraception to Young Women. July. Population Council, Washington, DC. Slesinger, D.P., and C. Ofstead. 1996. "Using a Voucher System to Extend Health Services to Migrant Farm Workers." Public Health Rep 111(1): 57­62. Sorensen, J.L., C.L. Masson, and A.L. Copeland. 1999. "Coupons/Vouchers as a Strategy for Increasing Treatment Entry for Opiate Dependent Injection Drug Users." In Motivating Behavior Change among Illicit Drug Abusers: Contemporary Research on Contingency Management Interventions, eds. S. Higgins and K. Silver- man, 147­161. Washington, DC: American Psychological Association Press. Steel, W. 2002. Small Business Development Services: Kenya Voucher Program. Seminar presented at the World Bank, Washington, DC. Presentation to Pro-poor Mar- ket Development Seminar, Series Social Capital Thematic Group by William F. Steel, Senior Adviser, Private Sector, Africa Region, World Bank, May 14, 2002. Steuerle, E.C. 2000. "Common Issues for Voucher Programs." In Vouchers and the Provision of Public Services, eds. E.C. Steuerle, V.D. Ooms, G.E. Peterson, and R.D. Reishaver, 3­39. Washington, DC: Brookings Institution Press. Stoner, T.J., B. Down, W.P. Carr, G. Maldonado, T.R. Church, and J. Mandel 1998. "Do Vouchers Improve Breast Cancer Screening Rates? Results from a Random- ized Trial." Health Serv Res 331: 11­28. Turner, S. 1997. "The Vision and Reality of Pell Grants: Unforeseen Consequences for Students and Institutions." Paper presented at the 25th Anniversary Pell Grant Conference, November 13­14, Washington DC. U.S. Department of Justice and Federal Trade Commission. 1997. "Horizontal Merger Guidelines." Washington, DC. Available at http://www.usdoj.gov/atr/ public/guidelines/horiz_book/toc.html. Wagstaff, A., P. Paci, and E. Van Doorslaer. 1991. "On the Measurement of In- equalities in Health." Social Science and Medicine 33: 545­557. Weinrich, S., D. Holdford, M. Boyd, D. Creanga, K. Cover, A. Johnson, M. Frank- Stromborg, and M. Weinrich. 1998. "Prostate Cancer Education in African American Churches." Public Health Nursing 15 (3): 188­95. Weinrich, S.P., M.C. Weinrich, J. Priest, and C. Fodi. 2003. "Self-Reported Reasons References | 107 Men Decide Not to Participate in Free Prostate Cancer Screening." Oncology Nursing Forum 301: E12­E16. West, Edwin G. 1996. "Education Vouchers in Practice and Principle: A World Sur- vey." HCO Working Paper 64, World Bank, Human Capital Development and Operations Policy, Washington, DC. Wortman, P.M., and R.G. St. Pierre. 1997. "The Educational Voucher Demonstra- tion: A Secondary Analysis." Education and Urban Society 9(4): 471­92. World Bank. No date. Reading List for Incidence Analysis: Poverty Impact Analysis. Washington, DC. --. 1995. Investing in People: The World Bank in Action. Washington, DC: World Bank. --. 2000. Indonesia: Evaluating A Pilot Pro-Poor Safe Motherhood Project. Poverty Net Library. Available at http//:www.poverty.worldbank.org/library/. INDEX A checklists feasibility assessment, 42 abuse of voucher schemes, 37­38, 71 prefeasibility assessment, 31 acceptance, 78 voucher design, 72 access rights, 76 cherry-picking. See cream-skimming adjudication, 80 client satisfaction, 21 administration, 20 clinical records, 69 adverse selection, 96 collusion, monitoring for, 93­94 monitoring for, 95 communications, 66, 67, 78 agency role, lack of institutional lack of, 40 capacity, 40 competition, between providers, 13­15, 39, appendices, 80 91­92 arbitration, 80 absence of, 13­14 competitive voucher schemes, 17­21 B context in which subsidies will be intro- benefit incidence analysis, 98­99 duced, 29 benefit policies contracting, 15, 77 determining, 54­56, 57 contractors' responsibilities, 80 examples, 55 contracts limits, 57 elements, 78­80 benefits vs obstacles, 42 negotiating, 74­76 black markets, monitoring for, 94 tips, 75 contributions, 8 C corruption, 15 cost-effectiveness, 4 capacity, lack of, 38 monitoring and evaluation, 100 capitated payments, 6, 9, 12 cost-per-case contracts, 9 cases, tracking and follow-up, 71 costs, 4, 5 cash administrative, 12, 40 refunds, 8­9 fixed or predictable, 20­21 subsidies, 6 transaction, 40 transfers, 8 voucher-bearer-borne, 57 cervical cancer screening counterfeiting limits, 57 monitoring for, 93 prefeasibility assessment, 31­32 prevention, 63­64 provider selection, 62 cream-skimming, 12 voucher design, 65 monitoring for, 95­96 109 110 | A Guide to Competitive Vouchers in Health D output-based, 11­12 see also monitoring data collection forms, 69­70 evidence-based practice, 10­11 design, 70 exclusions, 56 data entry, 70 expectations, 56 definitions, 78 expiration date, 56, 64 demand, provider-induced, 20 extension, 79 demand-side subsidies, 6, 7, 8­12 externalities, 5 decision tree, 30 supply-side vs, 7 F design vouchers, 63­65 facility inspections, 90­91 voucher scheme, 45­72 feasibility assessment, 33­43 diagnostic records, 69 advantages, 40­41 discounts, 57 benefits vs obstacles, 42 dishonesty, within agency, monitoring for, checklist, 42 94 financial feasibility, 41 disputes, resolution, 80 objectives, 33­34 dissatisfaction, notice of, 80 sequence of activities, 35 distribution, 66, 67, 68 tuberculosis voucher scheme for private registry of distributors, 69 practitioners, 43 strategies, 71 feasibility studies, 25­32 distributors, training, 82 see also prefeasibility studies doctors, lack, 34 fee-for-service subsidies, 10 duration, 79 fees, 4 schedule of, 79 E field visits, 90­91 financial feasibility, 41 early warning, 79 follow-up, registry of patents requiring, 69 efficiency force majeure, 80 consumer-led, after service utilization, funding, voucher scheme, 27 8­9 consumer-led, before service utilization, H 8 incentives to improve, supply-side health care, access to, 5 subsidies, 8 health outcomes, monitoring and evaluat- provider-led, after service provision, 10 ing, 97­98 provider-led, before service provision, 9 Herfindahl-Hirschmann index, 91­92 eligibility, 48­49, 51 Honduras, political impediments to vouch- El Salvador, cervical cancer program er schemes, 37 limits, 57 piloting, 83 I political impediments to voucher impediments, identifying, 34, 36 schemes, 37 implementation, 73­85 provider selection, 62 information equity, monitoring and evaluation, 99­100 health care needs and costs, 5 errors, remedying, 79 systems, developing, 68­72 evaluation, 87­100 institutional capacity, lack of, 40 cost effectiveness, 100 institutionalizing, 84 equity and poverty reduction, 98­100 intervention, specific, 27 health outcomes, 97­98 Index | 111 J patient empowerment, supply-side subsidies, 7 justification, voucher scheme, 26 patient satisfaction, 12, 21 payments, 79­80 L conditions and schedule, 80 law, 78 piloting, 82­83 leakage, 15 policy legal impediments, 34, 36 areas, 45 recipients, 48­54 M political impediments to voucher schemes, 37 market, lack of, 39 political support, lack of, 36 medical records, storing, 71 poor, identifying, 51 ministry of health officials, 36 population, 4 monitoring, 87­100 poverty reduction, monitoring and evalua- abuse of system, 92­96 tion, 98­99 cost effectiveness, 100 prefeasibility assessment database and, 71 cervical cancer screening, 31­32 equity and poverty reduction, 98­100 checklist, 31 health outcomes, 97­98 prevention strategies, 5 human quality of care, 90 price policies interpreting findings, 91 determining, 56­57 output-based, 11­12 various voucher schemes, 58­59 providers, competition between, 91­92 pricing, 73­74, 79­80 service quality, 88­89 printing, 65 technical standards, 89­90 private providers, 4 voucher recipients and redeemers, char- problem, identifying, 26 acteristics, 96­97 program running and monitoring, database mystery patients, 90 and, 71 provider policies, 71 N designing, 60 national health systems, 6 determining, 58­62 Nicaragua selection criteria, 60­61, 62 service package vs needs, 21 providers voucher design, 65 accredited, 14 competition between, 91­92 O participation, 61 payment for service, 67­68 obligations, general, 78 registry, 69 offer, 78 remuneration, 71 output-based representative, 78 monitoring and evaluation, 11­12 public system, poorest countries, 4 remuneration, 10 purchaser's representative, 78 overservicing, 12 monitoring for, 94 Q ownership, 76 quality control, 79 P quality of care, 4 monitoring, 90 parties, 78 partnerships, building, 29­30 112 | A Guide to Competitive Vouchers in Health R T receptionists, training, 81 Taiwan, voucher design, 65 recipient policies, 48­54 targeting of beneficiaries, 11, 19 designing, 49 adverse selection and, 96 recipients program aims and, 50 indirect, 52 supply-side subsidies, 6­7 registry of, 69 target payments, 10 record linkage, 69 tax rebates, 8 referral vouchers, 9 technical standards, monitoring, 89­90 regulatory impediments, 34, 36 tendering for service providers, 73­74 reimbursement value policies, determining, training, 77 62­63 service providers, 81 reporting to funding agencies, 71 voucher agency, 81 responsibility, 80 voucher distributors, 82 right to claim, 79 transferability, 51­54 right to vary, 79 how vouchers can be made risk, 80 nontransferable, 53­54 risk sharing, 5 transport, 66, 67 lack of, 40 S tuberculosis, feasibility assessment, 43 scaling up, 84­85 V sensitivity, specificity vs, 51 service packages, 20­21 variations, 79 needs and, Nicaragua, 21 value of, 79 services voucher agency outputs, evaluation, 99 examples, 47 quality, 88­89 requirements, 46­48 underconsumed, 20 scaling up, 85 sociocultural barriers, 38, 39 vouchers, 17­21 socioeconomic status, monitoring and administration, 20 evaluation, 99­100 advantages, 19 specifications, 76 design checklist, 72 specificity, sensitivity vs, 51 design, 63­65 standards of care, 76 how the scheme works, 18­19 stigma, 38, 39 multiple-section, 64 subcontracting, 79 printing, 65 subsidies voucher scheme, design, 45­72 delivering, 3­15 justification, 5 W problems, 4 wealth, inequitable distribution, 5 supply-side subsidies, 4, 6­8 advantages and disadvantages, 10­12 Z appropriate use, 6 decision tree, 28 Zambia, provider selection, 62 demand-side vs, 7 disadvantages, 6­8 incentives to improve efficiency, 8 patient empowerment, 7 targeting, 6­7 C ompetitive vouchers are one of many types of an output-based aid approach to health care, which links public funding to delivery of basic services. A Guide to Competitive Vouchers in Health presents tools that can help determine the appropriateness of competitive vouchers to meet different policy objectives. The book explains potential advantages and disadvantages of this approach and discusses ways to deliver public health care subsidies that promote efficiency, innovation, and accountability. A Guide to Competitive Vouchers in Health is directed to international donors and health care providers and policymakers in developed and developing countries. ISBN 0-8213-5855-3