A case study from Reducing Poverty, Sustaining Growth--What Works, What Doesn't, and Why A Global Exchange for Scaling Up Success Scaling Up Poverty Reduction: A Global Learning Process and Conference Shanghai, May 25­27, 2004 Papua New Guinea and Sri Lanka: Scaling Up Health Interventions Shiladitya Chatterjee, Principal Poverty Reduction Specialist Poverty Reduction and Social Development Division (RSPR) Regional and Sustainable Development Department (RSDD) Asian Development Bank (ADB) 6 ADB Avenue, Mandaluyong City 1550, Philippines Phone: (632) 632-5983 e-mail: schatterjee@adb.org Leo Deville, Managing Maryse Dugue, Colandavelu Brahm Prakash, Director Manager, Malaria Narayansuwami, Director Health Research for Medicines and Supply Managing Director Poverty Reduction and Action Services C.V. Nam and Social Development Laarstraat 43 World Health Associates Division (RSPR) B-2840 Reet, Belgium Organization 3/48 Hornebush Road, Asian Development 20 Avenue Apia, CH- Strathfield, N.S.W Bank (ADB) 1211 Geneva 27, 2135, Australia Phone: (632) 632-6646 Switzerland Phone: 612-97644718 Fax: (632) 636 duguem@who.int cvnam@ozemail.com.au 2360/2356 bprakash@adb.org Implementing agency contact: Dr. Puka Temu, Minister for State Enterprises and Information, Papua New Guinea. Development partner: Asian Development Bank The findings, interpretations, and conclusions expressed here are those of the author(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 cannot guarantee the accuracy of the data included in this work. Copyright © 2004. The International Bank for Reconstruction and Development / THE WORLD BANK All rights reserved. The material in this work is copyrighted. No part of this work may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or inclusion in any information storage and retrieval system, without the prior written permission of the World Bank. The World Bank encourages dissemination of its work and will normally grant permission promptly. PAPUA NEW GUINEA AND SRILANKA: SCALING UP HEALTH INTERVENTIONS Executive Summary The Papua New Guinea (PNG) and Sri Lanka experiences in some ways provide useful comparisons and hold valuable lessons for developing countries attempting to scale up health sector interventions to achieve their millennium development goals (MDGs). Although there are obvious differences between the two countries, there are common features in the health systems of both. The differences include the nature of the countries, for example: PNG is sparsely populated and the terrain makes communication difficult; Sri Lanka is more densely populated and has better communications infrastructure. Human resources development indicators were higher in Sri Lanka. However, both countries had decentralized health care delivery systems, although PNG started much later and was still in the throes of institutionalizing decentralization in the 1990s. In both cases, initial efforts were to develop the primary health care infrastructure to extend health care coverage to the countryside and rural population; while later the emphasis turned to improving the quality of health care services. Both countries were faced with low financial allocations to the health sector. Although PNG was able to step up allocations significantly in the 1990s, health allocations remained low as a share of GDP and on a per capita basis. However the outcomes are quite dissimilar. Sri Lanka has reached or is well on its way to reaching the Millennium Development Goal targets in health, while PNG is very far behind. In trying to find the causes of the somewhat unsatisfactory results in the PNG case, a comparison of the two countries' experiences where the Asian Development Bank (ADB) played a significant supporting role provides valuable insights. An important conclusion is that while funds are obviously important, they are not sufficient to guarantee delivery of services. PNG increased health allocations significantly and yet outcomes remained unsatisfactory. Sri Lanka was able to make do with low allocations and appeared to have delivered health services quite cost-effectively. One must note, however, that the adequacy of expenditures on health depends greatly on country circumstances. Sri Lanka may have been able to make its health expenditures go further in view of better complementary communications infrastructure and human resource development attainments overall. In addition, because of public budgetary constraints, there is growing private sector involvement in Sri Lanka. The quality of governance is obviously a major determinant. While both countries exhibited strong leadership and political will at the central level, in PNG this did not permeate down to lower decentralized levels to the extent needed. There were very apparent problems in coordination, both among departments and between central and decentralized levels. Inadequate monitoring and supervision has also been cited as a factor in the PNG case. In addition there is the issue of lack of good governance, PNG still needs to address adequately. A special governance-related matter is the proper implementation of decentralization and its implications for developing countries worldwide. Many developing countries in Asia are still grappling with this problem--including Pakistan, Indonesia, and the Philippines--and there are no easy solutions. Decentralization is essential for public participation but the administrative readiness for 1 CASE STUDIES IN SCALING UP POVERTY REDUCTION delivering basic services at the local levels remains questionable in many cases. This has major implications for achievement of the MDGs and is not a problem related to the health sector alone. The PNG case illustrates the problem very well. Issues that the PNG case highlights, and that need to be addressed, include (i) capacity building at the regional level as well as central level for monitoring, supporting, and coordinating decentralized efforts; (ii) improved fiduciary standards; and (iii) effective personnel policy that provides incentives for performance at all levels. An important requirement for effective delivery of quality health services is human resources development and management in the health sector and this has been underscored through both the Sri Lanka and PNG cases. In Sri Lanka, the first health project had not focused sufficiently on this issue and the deficiencies became apparent soon. This was effectively rectified in the second project. In PNG, management of human resources remains a major problem in the public sector in general and health services in particular. Much more attention needs to be focused on human resources issues if health services are to be delivered effectively. While focus on public health services will continue to need attention, support can be provided through alternative means such as the private sector (if that is developed) or nongovernmental organizations (NGOs). While PNG has made attempts to use the services of some NGOs, such as the church, some regional examples hold promise. Cambodia's recent success with contracting out health services to NGOs is an example. 2 PAPUA NEW GUINEA AND SRILANKA: SCALING UP HEALTH INTERVENTIONS Introduction Health is an important dimension of poverty. In many societies, poverty and bad health are synonymous--a person with poor health is poor. As Amartya Sen (1999) explains: "The usefulness of wealth lies in the things that it allows us to do, the substantive freedoms it helps us to achieve [...]. And among the most important freedoms that we can have is the freedom from avoidable ill-health and from escapable mortality."Better health leads to an immediate welfare gain for the poor and access to essential health services is a basic human right. It is not surprising that the Millennium Development Goals (MDGs) place health and nutrition at their center. The last few decades have seen substantial gains in the health of populations throughout the world. However these gains have not been shared by many of the poor and often the poorest and most vulnerable groups of society have been left behind. This has led to the so-called "vicious circle" of bad health and poverty and growing inequality in health status in many countries (Bhushan, Bloom, Nguyen, and Nguyen, 2001; Bloom, Canning, and Jamison, 2004). The central issue in designing health sector interventions is to make health services work for the poor. Improving the access to primary health care services to the poor is a well-documented policy prescription in developing countries. Unless well designed, health services provided by Government tend to be distributed in favor of the non-poor at the expense of the poor (Gwatkin 2000; Jha and Mills 2002; Wagstaff 2000). Strategies to better target the poor within the broad scope of health sector reform are difficult, at best, especially when fees are charged to help offset the cost of services (Abel-Smith and Rawal 1992; Besley and Kanbur 1993). At the same time, user fees and private participation can strengthen client power and enhance the effectiveness of services provided (World Bank, 2003). Increasingly, it is being recognized that enhancing development effectiveness and reaching the poor requires more than simply increasing spending. It requires a two tiered approach that focuses both on (1) generating incentives for the health systems to function efficiently and encouraging both public and private providers to operate and provide pro-poor services ("creating an investment climate"), and (2) increasing the equity and pro-poor orientation of public policy in the health sector ("social inclusion"). There is no single panacea or "one size fits all" approach that will improve the health status of the poor in different countries through the world. Rather countries need to learn from each other, experiment, and scale up approaches that work the best, given each country's context. This paper attempts principally to draw lessons from Papua New Guinea (PNG) where ADB has had long term involvement in the health sector. In doing so, it makes selective references to health sector experiences of two other countries where ADB has also been involved, namely Cambodia and Sri Lanka. While there are considerable differences in overall economic and social achievements among the three countries (see Table 1), there are many lessons that can be drawn from a comparative study to explain the health outcomes in PNG. The paper begins (Section II) by describing the implementation process of ADB health sector investments in these three countries, including their rationale and objectives and the context of the health sector operations. It then (Section III) looks at the impact that these interventions have had on health, particularly health of the poor. In Section IV it 3 CASE STUDIES IN SCALING UP POVERTY REDUCTION seeks to explain factors behind the observed impacts and provides a generalized set of lessons learned in Section V. The basis of the PNG study is a special staff assessment conducted for the purposes of this Conference, a health sector review conducted in 2002 and a study done on the Health Sector Development Program Experience . The Cambodia results are based on the National Health Survey 1 of the National Institute of Statistics, Phnom Penh, 2000. The Sri Lanka experience is based on post evaluation studies2. ADB's Health Interventions and Implementation Experience in Papua New Guinea (PNG), Cambodia, and Sri Lanka The Asian Development Bank (ADB) has had long-term involvement in the health sectors of Papua New Guinea (PNG), Cambodia and Sri Lanka. In this section a brief description is provided of the respective countries' health sectors and ADB's health sector programs. Papua New Guinea ADB's program for improving public service delivery included assistance to the health and water supply and sanitation sectors, with ADB playing a major role in both. Modalities included investments, technical assistance and policy dialogue to develop national strategies and to guide the sector's development program. In the health sector, ADB funded the development of the 1996-2000 national health plan that was used to help officials make informed policy and investment choices. Assistance to the health sector in PNG had focused in the 1980s and the early 1990s on improving access to rural health services by investing mainly in infrastructure (aid posts, health centers, staff housing) through three rural health projects. Difficult terrain and far-flung communities had prevented health services to be effectively delivered and initial interventions thus focused on building up delivery infrastructure. Evaluations confirmed the improved physical accessibility to health facilities in many remote parts of the country, but did show that extending facilities did not solve the problems of closure or limited operation of facilities due to shortage of staff, insufficient drugs, malfunctioning equipment, and poor maintenance of buildings. A loan focusing on support for population and family planning was also approved in 1993. After 1994, the health sector strategy was refocused from expansion of the service delivery network to quality improvements, and eventually resulted in the Health Sector Development Program (HSDP). While considerable health gains had been made in the first decade after PNG's independence in 1975, progress in a number of key indicators has slowed since the early 1990s. The deterioration of 1TA 3762: Papua New Guinea: Health Sector Review; and Moving Toward a Sector-wide Approach: PNG the Health Sector Development Program Experience: John Izard and Maryse Dugue, ADB Pacific Studies Series August 2003. Several project RRPs have also been consulted including RRP on Proposed Loan and TA grant to PNG for the Third Rural Health Services Project, August 1991, and February 1997. 2Project Performance Audit Reports: Health and Population Project in Sri Lanka (1993); Second Health and Population Project in Sri Lanka ( September 2003). 4 PAPUA NEW GUINEA AND SRILANKA: SCALING UP HEALTH INTERVENTIONS health services has many causes including the macro-economic crisis of 1994­1995. The economic recovery package (ERP) by the International Monetary Fund (IMF) and the World Bank (WB), assisted by both Australia and Japan, in 1995­1996 provided for government policy to redirect spending to the social sector, particularly in health and education. These events established a framework conducive to an ADB health sector program loan as part of the overall economic recovery effort. Among Pacific countries, PNG's Human Development Index (HDI) is the lowest (rank 133 in 2002), and the poverty incidence, as measured by the Human Poverty Index (HPI), the highest (rank 62 in 2002). Poverty and low levels of human development have an important gender dimension. Life expectancy, income, and educational achievements of women are universally lower than those of men. Although men and women have equal rights under the constitution, gender inequality remains a severe impediment to development. PNG's total burden of disease is high. It is calculated at 21,000 DALY3 per 100,000 compared to 16,000 in the Western Pacific as a whole. In the communicable, maternal and perinatal category, the difference is striking with 12,301 DALY per 100,000 compared to 3,900 in the Western Pacific. Maternal mortality is 370 per 100,000 births placing it second highest in the Western Pacific. Infant mortality rates are also very high ranging from 33 to 87 per 1,000 live births and averaging 73. The under-5 mortality rate is recorded as 102 per 1,000 live births. Obviously, PNG is lagging far behind the 2015 MDG targets. These pose a major challenge to policy makers. At the point of care, health services are often not delivered effectively. Many facilities are closed or not properly staffed, equipped, and supplied. Only half of scheduled outreach clinics4 (immunization, maternal and child health) are actually held, resulting in low immunization rates and contributing to the high maternal and child mortality. One indicator of the ineffectiveness of delivery of health services is that only 57 percent of pregnant women receive any prenatal care (DoH, 2003a). Childhood immunization is less than 60 percent (DoH, 2003a) resulting in epidemics of preventable childhood communicable diseases such as measles. In comparison with its Pacific country neighbors, PNG spent in 1997 only 2.8 percent of GDP on health in the public sector (falling from 3 percent in 1990), the lowest as proportion of GNP and also the lowest level of health expenditure per capita. Within this overall allocation to health, there are also significant variations between provinces on health resources. These variations reflect past investment decisions as well as current perceived provincial priorities. However, there was a significant step-up of expenditures on health after 1997 with the public health expenditure going up to 3.9 percent of GDP in 2001. Prior to HSDP, the public health system had tried to cope with its poor performance essentially without any change in its management style. A series of reforms and policy initiatives were identified as critical to reverse the declining trend of health services. Focusing on the whole sector with a major emphasis on performance of rural health service delivery, the HSDP marked a 3Disability Adjusted Life Years (DALY) is a measure of morbidity and mortality that calculates the number of years of healthy life lost each year. These results are from the 2002 PNG Burden of Disease Study, Hiawalyer and Spohr. 4PNG health indicators from the NHIS. 5 CASE STUDIES IN SCALING UP POVERTY REDUCTION new approach to assisting the health sector. Under the strong leadership and high commitment of the Secretary for Health issues of sector performance and decentralization were being addressed (though with variable degrees of success) through new partnerships between the central and provincial levels. The HSDP was the first sector wide assistance program in health in PNG designed to support the National Health Policy, which was developed through a broad consultative process in 1995. The Policy focused on improving health services to the rural majority and the need to adopt health promotion and preventive health strategies to ensure improved health status. It also addressed issues of management reform in all areas and all levels of the public health hierarchy. The policy gave shape to the National Health Plan, a cohesive document that presents a comprehensive set of objectives and strategies in the public health sector to be pursued over five years. The objectives and strategies described under HSDP, consolidated as benchmarks for the purpose of evaluating performance, were drawn from the government's National Health Plan 1996­2000. The policy-based loans under the HSDP were also designed to support the reform process embodied in the Organic Law on Provincial Governments and Local Level Governments5 passed in 1995, adjusting and completing the process of decentralization begun after Independence. Implementation of HSDP began in 1998, coinciding with the progressive implementation of the National Health Administration Act (1997). A technical assistance project was attached to HSDP to provide advisory support on the objectives of the development program. The overall goal was to provide services to the majority of people and, in particular, to the rural poor. Consequently, assistance focused on rural areas and on primary and maternal and child health care. Apart from ADB, other development partners have also been active. In fact, AusAID is the biggest donor in the health sector. Australia has provided general budgetary support of approximately A$300 million per year to the government of PNG. In a change of policy, budgetary support has been phased out in favour of project aid between 1995 and 2000. Early in this transition, AusAID focused on hospital management and operations as well as providing assistance in the training of health professionals. The Health Sector Support Program, which provides comprehensive assistance to the National Department of Health (NDOH) and targets six provinces, came online in 1998-1999. The Women and Children's Health Project was designed to improve vaccination coverage and women and children's health extension activities nationwide. Other development partners that are active in the public health Sector are New Zealand, Japan, USAID, EC. Cambodia While terrain, geography and habitat pattern initially posed major constraints in PNG, in Cambodia it was the near total destruction of administrative machinery and human resources as a result of war and political strife. From near non-existent to extremely low levels only ten years ago, the Cambodian private and public sector health system has developed rapidly and has had a number of important achievements during this brief period that have been encouraging. In terms of rebuilding the health system, the Government developed a national health policy that includes major financial reforms 5The Organic Law on Provincial Governments and Local Level Governments (New Organic Law), enacted in 1995, provided the framework for a greater devolution of powers to the provinces and local level governments, with clearer definitions of the division of administrative and service functions between the three levels of government. 6 PAPUA NEW GUINEA AND SRILANKA: SCALING UP HEALTH INTERVENTIONS (e.g., budgetary reforms, introduction of user fees) and a national system of primary health care coverage. The country has also made substantial progress in other critical areas of public health, including HIV/AIDS prevention and in the introduction of modern birth-spacing methods. Notwithstanding the progress in the last decade, health indicators in Cambodia are still among the worst in the Asia and Pacific region. Average life expectancy at birth is estimated at only 56 years--54 years for men and 58 years for women. The infant mortality rate is estimated to be 95 per 1,000 live births, while the mortality rate under the age of 5 is 124, and the maternal mortality ratio is 437 per 100,000 live births (Ministry of Planning, 2003). Rates of malnutrition are the second highest in Southeast Asia, with 56 percent of children under 5 affected by chronic malnutrition. War and political upheaval has left Cambodia with almost non-existent health care infrastructure, especially in rural areas. There were sufficient paramedical and management staff, but training and quality of care were inconsistent and morale was low (Bhushan, Keller and Schwartz 2002). The primary health care system was not able to deliver an adequate level of services. For example in 1998 only 39 percent of children 12-23 months of age were fully immunized (National Institute of Statistics 2000). Since Cambodia began the current process of reintegration in 1991, ADB has had three projects which have targeted interventions in health: the Basic Skills Project (approved 1995), the Basic Health Services Project (approved in 1996) and the on-going Health Sector Support Project (approved in 2002). The Basic Skills Project had primary health care training as one of its two main components. The objective of the Basic Health Service Project is to assist with the development and implementation of a coverage plan modeled according to WHO guidelines to restructure and broaden the primary health care system. The plan included the construction or rehabilitation of health centers designed to provide services to a population of about 10,000 people, and the creation of operational districts with populations with an average of about 150,000 people. The coverage plan also defined a minimum package of activities for health centers consisting of basic preventive and curative services including immunization, birth spacing, antenatal care, provision of micronutrients, and simple curative care for diarrhea, acute respiratory tract infections, and tuberculosis. One innovative part of the Project was to pilot contracting of health services to NGOs in some operational districts6. The Health Sector Support Project (HSSP) builds on the Basic Health Services Project. It was designed in the context of the Government's Health Sector Strategy, which was also approved in 2002. This Strategy is the health sector's main input into the National Poverty Reduction Strategy and fully reflects the MDGs in its monitoring framework. It has received widespread support from stakeholders and builds on previous policies in the health sector, including the minimum package of services. To support the implementation of the Strategy, ADB, the World Bank, and the United Kingdom's Department for International Development (DFID) entered into a partnership to develop and implement a joint project. Although HSSP is not a SWAP in the classic sense of the term, it does have many of the characteristics of SWAP. The partners coordinate their activities closely and, when 6The Cambodia contracting experiment is discussed in detailed in the World Development Report (World Bank, 2003) and in Bhushan, Keller and Schwartz 2002. 7 CASE STUDIES IN SCALING UP POVERTY REDUCTION feasible, share resources and have adopted the same technical approach for their activities. Given the proven success of contracting (see below), HSSP supported the expansion of the model to 11 districts. Sri Lanka Compared to PNG and Cambodia, the health system in Sri Lanka presents almost a model of a functioning health care system if not a well functioning system. The Government of Sri Lanka has traditionally made substantial efforts in human resource development, including health care and family planning services. This was reflected in Sri Lanka's 2000 indicators, which were better than those of its neighboring countries in South Asia. In 1980, the Government signed the Charter for Health Development and endorsed the global strategy of Health for All by year 2000. The charter promoted the utilization of primary health care (PHC) as the main health delivery system. A new National Health Policy was adopted in 1992 which aimed at strengthening human resources development and management in the health sector stressing managerial capacity; education research and training; improving capacity for monitoring and control of diseases; and encouragement of the private sector for providing health services. A special emphasis was also placed on integrating primary health care services with family planning. Thanks to the long-standing support of the government to health sector over the decades, even though Sri Lanka's health expenditures are low as a share of GDP in comparison to PNG and Cambodia, its health sector indicators are quite impressive compared to several comparator Asian countries. Table 1 below illustrates this. ADB has been involved in the Sri Lanka health sector since the early 1980's. Support was first provided under the Health and Population Project, which was designed to help implement the Government's Health for All initiative. The loan which was approved in 1982, aimed at strengthening the primary health care (PHC) delivery system. It was considered in post-evaluation to have been satisfactorily implemented, achieved its objectives and rated generally successful. ADB followed up its first health and population loan with a second project which essentially supported the implementation of the new National Health Policy of 1992. The key issues that required assistance in the implementation of the Policy were in the areas of (i) PHC services that needed strengthening in terms of equipment, communication, and transport facilities, and upgrading of physical infrastructure to enable functioning and delivering of quality services; (ii) support for human resource development (HRD) and training; (iii) modern hospital management methods to improve the efficiency and cost effectiveness of hospital services; (iv) strengthening the referral system between small hospitals in rural areas with larger, better equipped hospitals; (v) maintaining the high acceptance of family planning methods that experienced setback due to the civil unrest, meeting the demand for clinical contraceptives, and training the providers of these services; (vi) setting up a computerized management information system (MIS) to monitor the decentralized health care delivery system, particularly at the PHC level; and (vii) preparing a proposal for promoting health insurance. The Policy was also in line with ADB's focus in the health and population sector at that time, which was to upgrade existing assets and improve efficiency through human resource and institutional development. The Project which was approved in 1992 aimed to directly benefit the rural population, particularly the poor, elderly women, and children. The post evaluation report found this 8 PAPUA NEW GUINEA AND SRILANKA: SCALING UP HEALTH INTERVENTIONS project, like the first one, also successful in meeting its objectives. Apart from ADB, other development partners who have been active in the sector include the World Bank, Japan, United States, Finland and the UN system. ADB's activities were coordinated with these partners, particularly with UNICEF, WHO and the World Bank. Impact Analysis Papua New Guinea Health Expenditure Between 1996 and 2001, there was a 65 percent increase in real public sector expenditure on health in PNG. When population increases are taken into account, there has been an increase of 42 percent in the overall availability of resources for the health sector (SMRG 2002b). Over this period, per capita domestic resources for health have risen by 4 percent and this has been brought about by increasing the proportion of the national budget spent on health from 4.8 percent in 1996 to 6.2 percent in 1999. Foreign assistance has provided the bulk of additional resources to the health sector. Due to these efforts, public health expenditures rose to 3.9 percent of GDP in 2001. Public resources are supplemented to a significant extent by NGOs (principally church and missionary run facilities which run about half of rural centers) and to a lesser extent by the private sector. Initial concerns about lack of adequate financial resources in the health sector led to prioritizing financial allocations for health. With better financial resources availability deeper structural issues were unmasked (as discussed later). Under the HSDP's first policy reform area "Shift Emphasis from Urban to Rural Health Services," the first of the benchmarks reflecting the government's priorities was to increase financial allocations for rural health services by not less than 10 percent every year in 1997, 1998, and 1999 over the 1996 level of expenditure. Analysis of the annual public health sector expenditure revealed rural health services expenditure increased by more than 10 percent for each of the years 1997­1999 over the 1996 level of expenditure, specifically: (1997) 44.1 percent, (1998) 22.4 percent, and (1999) 16.2 percent. The large increase in 1997 is due to the inclusion of AusAID project aid for the first time. Two more benchmarks were exceeded. One was to "increase the value of drug supplies to health centers and aid posts to K3.0 per capita in the area served." In 1998, the value was K4.29 per capita; in 1999, it was K6.00 per capita. The second was to "increase the allocation for drugs and medical supplies to 25 percent of NDOH's budgetary allocation for 1998." In 1998, the allocation of drugs and supplies as a percentage of total NDOH expenditure was 28 percent. In 1999, the value increased to 31.8 percent. More importantly, the nature of the contribution has resolved the persistent problem of inadequate drug supplies reaching the lowest level of the public health system, the aid posts. Output Indicators Although most important program benchmarks were achieved under the HSDP, and financial resources for health increased considerably, health outcomes did not improve over those achieved in the mid 1990s. Poverty is reported to be increasing in both urban and rural areas. A review of the 9 CASE STUDIES IN SCALING UP POVERTY REDUCTION National Health Plan 1996­2000 revealed deterioration in several indicators such as a drop in immunization coverage, a rise in mortality from malaria, an increase in malnutrition, and widespread shortages of medicines. While there have been gains shown in the demographic indicators over the last decade, with an increase in life expectancy from around 52 years in the 1990 census to around 54 years in 1996, which however remains low by regional standards, the infant mortality rate (IMR) had risen from 72 to 77 per 1,000 births over the same period, the only country in the Pacific to record an increase in IMR. There are significant urban and rural differences in IMR, with figures ranging from 33 to 86 per 1,000 lives births respectively. Importantly, there are indications that PNG may soon face a serious HIV/AIDS epidemic. While inputs (measured by financial resources) to the health sector have increased significantly between 1995 and 2001, output indicators have shown only a marginal increase over the seven years and suffered a significant fall between 1999 and 2000 and further in 2001. The performance for 2001 was, for most indicators, worse than in 19957. Table 2 provides the output indicator measurements for five key Mother and Child Health indicators for the beginning and the end of the series. There are also significant variations in the performance of provinces (Table 3). A well designed management information system(MIS) for the health sector put in place with ADB support made available good data on these reversals and helped in signalling policy makers to take corrective action. Several factors have contributed to the mixed performance of the health sector in PNG. Some indicators are particularly resistant to change, being influenced by factors outside of the health sector (such as women's education and nutritional status) and geographical barriers (such as supervised delivery rate which is influenced by accessibility to health centers). Difficult terrain, combined with poor infrastructure and lack of transport, reduces considerably the physical accessibility of health services. Other indicators have fluctuated according to availability of funds, quality of support and supervision, and personnel involved (such as immunization coverage). 7A part of the reason for unsatisfactory improvements as revealed by the indicators is due to the adoption of a well- developed management information system (MIS) which was put in place through ADB support in 1991. 10 44. 7 )PDGfotnecr 2001 11. 73. 73. 33. 42. tal To 2.3 9 1997 10. 2.3 7.3 6.3 4.2 10 (8) pe(serutidn 50. 91. 61. 81. 81. etavi 2001 .)8 Pr 4.0 8.9 6.1 6.1 0.2 8.1 1997 xpeEhtlaeH 93. 71. 81. 12. 51. 60. 2001 to6,4,3sn licbuP 8.2 1.1 6.1 1.2 6.1 6.0 1997 03. 51. 61. ... 70. 30. ONSI 1990 NT VERE tea 95 20 29 40 48 5r R 2000 124 umlo(cesabatadenilnoID NT 00)0 I (7) HTL denU lityatro 1,r(pe 1990 101 115 26 40 63 91 Wdna, M EA H P rep( 5)n U ntaf evil 79 95 17 25 30 35 2000 NGI (6) litya )shtrbi 79 80 22 34 45 60 CALS:AK sor In tro 0001, 1990 M umlo(csrotac e 60 44 .srotacid 11 ANL RI catdiIn (5) S lanret p(lityatro vil000100, re 1995 470 370** 437 172** IndiyeK InyeK Ma )shtrbi 930 900 140 200 280 650 2002 M 1990 BDA2 NDA ent ( le 57* 24 3 52 65 94 BDA 200 A 2000 89. NE UI (4) G maeFtludA teaR fot 51se )e 57. 90. 92. 81. 2), ag acyr cenrep ovba & 1990 48* 79 67 45 21 51 &1sn 48. 84. 89. 91. 72. EW opmleveD iteL maleef .srotacidnIyeK N AU APP anmuH 5 9 3 2 8 1 morferaaeniuG . atpiacr BDA 2056. 1746. 3083. 5932. 3671. 2768. 02 (3) 20 ndaht pe PDG 5991tnastnoc( )$lan 2001 umlo(ctr tioanrteni poeRtne weN . PPP ealH I kn 62 75 31 21 23 33 (2) HP 2002 Ra edcteleS I kn 133 130 89 70 77 110 pmoleveDnamuH 2003aidobm morferaat daserut (1) HD 2002 Ra Ca,gn 2002 nnialP puaaProfsetarycaretilelameft 2000-6991 morfelblaiavaraeyt ndi 1.e seirt un GNP aidobm aisen xpeecilbuP do ablT Co Ca aknaLirS dnaliahT senpipilihP In PDNU:sec uroS s:et ofyrtsini dularofata No M *D ecenrtsomethroF** ***1990 CASE STUDIES IN SCALING UP POVERTY REDUCTION Table 2: Output Indicators for five key MCH activities, 1995 and 2001. Indicator 1995 2001 percent of deliveries in health facilities 42 percent 38 percent percent of pregnant women receiving TT vaccination 62 percent 63 percent percent of women getting at least one antenatal visit 68 percent 58 percent percent of children, <1 year, receiving 3rd dose TA vaccination 61 percent 55 percent percent of children getting measles vaccination 42 percent 47 percent Source: Data provided by Monitoring and Research Branch, MoH. Table 3: Performance of best & worst performing Provinces for five key MCH activities, 2001. Indicator Best Worst percent of deliveries in health facilities 68 percent 11 percent percent of pregnant women receiving TT vaccination 77 percent 39 percent percent of women getting at least one antenatal visit 108 percent* 37 percent percent of children, <1 year, receiving 3rd dose TA vaccination 107 percent* 40 percent percent of children getting measles vaccination 79 percent 30 percent Source: Data provided by Monitoring and Research Branch, MoH. Note: * the figures of over 100 percent probably reflect problems with the population figures used as the denominator. Impact on the Poor A special study conducted by ADB8 on trends of health services delivery to the poor found that poor households in Papua New Guinea have the worst access to health facilities, lowest uptake of services and the worst health outcomes. The study confirmed that there are large disparities in the delivery of services between the poorer and better off households and that the poor had been affected by closure of health facilities although towards the end of the period some improvements in quality of services run by government facilities benefited the poor (see Box1). An ADB Health Sector Review examined the changes in health care delivery between 1991-2000 using the 1996 Demographic and Health Survey (DHS) to divide the population into five socio-economic groups according to the possession of selected household assets. The major findings were: Management Culture and Sustainability Although output indicators have not improved, HSDP has contributed to changes in management culture that may have longer-term effects. Lack of impact on output indicators could also be due to the fact that the efficacy of bringing about improvements in the health sector also depends on improvements in the overall administrative system of the government which did not take place. The reform process initiated under HSDP introduced a framework, which established a working relationship between the National Department of Health (NDOH) and the provinces, in line with the Organic Law on Provincial Governments and Local Level Governments, and to which other Departments, government agencies and non-government agencies (e.g. Church managed health services) are also party. Management innovations were instituted, creating an environment of greater transparency and accountability. Issues related to 8"Trends in the Distribution of Health and Health Services in PNG 1991-2000" conducted under ADB TA 3762: PNG Health Sector Review. 12 PAPUA NEW GUINEA AND SRILANKA: SCALING UP HEALTH INTERVENTIONS performance began receiving greater attention and public service officials have become increasingly responsive to governance issues. This is of fundamental importance in a country where public spending on health hardly translates into a larger supply of effective health services, due to inefficiency, poor accountability, and lack of transparency in public service provision. Box 1: Trends in the Distribution of Health Services in Papua New Guinea 1991­ 2000 Poverty: The Highlands and Momase regions have the highest incidence of poverty as measured by the asset index. The difference between regions is partly due to the percentage of the population living in urban settings but also due to differences in socio-economic status between rural populations. Availability and Access to Health Services: The percentage of the population living within 5km of a facility was 85 percent in 1995. Poorer households live further away from health facilities (aid posts or mission health centers). The number of doctors, HEOs and CHWs employed in rural facilities decreased during the 1990s. The number of nurses increased but not in line with population growth. Many aid posts were closed during the course of the 1990's. Consequently the percentage of the population living within 5km of a facility decreased to 80 percent in 2000. The change initially affected the very poorest quintiles. Quality of Services: The services offered to poorer households are of lower quality than those available to more prosperous households in terms of qualifications of staff, equipment available at health facilities, availability of medical supplies and the extent of clinical supervision. There was some improvement between 1998 and 2000. Such improvements directly benefited the poor, but improvements in clinical supervision and medical supplies were largely confined to government run facilities. Family Planning Services: There are large disparities in the uptake of family planning between quintiles. Utilization of contraception is lowest amongst the poorest two quintiles. The disparities in the use of family planning between quintiles were probably exacerbated by the closures of aid posts. Maternal Health: Poor women have more complications during pregnancy and are less likely to receive antenatal care or a supervised delivery. Distance appears to have a strong influence on the rate of supervised delivery, but is not the only barrier to the uptake of services since many poor women living within 5km of a facility do not use it to deliver. There was a fall in the proportion of deliveries undertaken in all types of facilities between 1991 and 2000 with a sharp drop in mid-1995; most pronounced in rural facilities run by missions. Child Health ­ Immunization: Immunization services have some success in reaching poorer households as judged by BCG coverage and 1st dose triple antigen. However, fewer children in poorer households complete the immunization schedule. Overall rates for completion of triple antigen immunization fell by approximately ten percentage points in the 1990s. It appears that immunization rates were maintained in urban areas but were reduced in rural areas. Child Health ­ Use of Curative Services: Poorer quintiles have a higher incidence of self-reported illness, particularly fever, but less likely to seek treatment. Poorer quintiles are more likely to use government health centers followed by aid posts then mission health centers for treating children. Child Health ­ Mortality: There are large differences in childhood mortality rates between quintiles. The difference can be ascribed primarily to differences in post neonatal mortality. The leading causes of death from 1 to 11 months are pneumonia and other respiratory diseases (50 percent) followed by meningitis (12 percent), malaria (6 percent), septicaemia (5 percent) and diarrhea (4 percent). Early recognition of symptoms and appropriate treatment for pneumonia are essential but do not address the route causes of severe respiratory illness in poor children which relate to nutritional status, housing and environment 13 CASE STUDIES IN SCALING UP POVERTY REDUCTION Cambodia In contrast to PNG where health conditions failed to improve despite support, there was general improvement in Cambodia both in areas where the ADB projects was active and in other parts of the country, where other development partners including the World Bank were active. However, in comparison to PNG and Sri Lanka, poverty seemed to be extremely severe. This makes it difficult to analyze the overall impact of the ADB project or to isolate many of the positive aspects. However the contracting out component did have a separate and rigorous analysis showing how contracting of health services to NGOs affected the welfare of the poor. Initially only five districts were selected for contracting and the selection of districts was done randomly to ensure both treatment and control areas. In general, the results of the mid-term surveys, taken only after 2.5 years of the experiment, indicate that all contracted districts had already achieved their contractual obligations for most of their service coverage evaluation indicators. The overall results suggest that measurement of objectively verifiable baseline service coverage indicator values, combined with close monitoring, well defined contractual goals, and independent follow-up measurement provides a level of accountability which encourages increased performance. Figure 1 outlines the impact of the project in both types of contracting districts and in control districts9. Coverage of health services in contracted districts is seen to achieve significant improvements in a short time. The results of the mid-term survey suggest that contracted districts outperformed the control districts with respect to the contractually obligated coverage indicators. In addition, contracted-out districts appear to have out-performed contracted-in districts, on average.10 This is particularly the case for the use of reproductive health services, child health services and curative health services. For example, while all of districts achieved increases in antenatal care, coverage of antenatal care in contracted-out districts, however, increased by more than 43 percentage points, compared to 27 percentage points contracted-in districts, and 14 percentage points in control districts. A similar pattern is observed for maternal tetanus toxoid immunization. Contracted services, in general, appear to be more effective in reaching the poor, both in absolute and relative terms. Contractual targets established at the baseline defined "the poor" as the lowest 50 percent socioeconomic group in each district, which recognizes that being poor in these districts is not uncommon compared to the rest of the population of Cambodia11. The data suggest that utilization of primary and district increased dramatically. Contracted-out districts appear to have highest increase in utilization by the poor, with utilization rates increasing from 3 percent to 33 percent, compared to an increase of 4 percent to 8 percent. lower half of the socioeconomic group by nearly 30 percentage points. 9 In the Basic Health Services Project, two different types of contracts were used. Contracting out involved having an NGO administer all aspects of the government system. In contracting in districts, NGOs provided management support to the government system. 10 The differences in performance are statistically significant at the 5 percent level or better. 11 An index of socioeconomic status was developed based on ownership of household assets which serves as a proxy variable for household wealth. These assets include i) whether there was a permanent type of roof on the house (brick, cement, metal, or a combination of these materials), and whether anyone in the household owned a ii) bicycle; iii) radio; iv) motorcycle; v) television; vi) oxcart; vii) motor boat; and viii) at least one cow. 14 PAPUA NEW GUINEA AND SRILANKA: SCALING UP HEALTH INTERVENTIONS One possible explanation is that contracted-out districts did not implement official user fees and discouraged "unofficial" user fees by paying significantly higher salaries to lower level health care providers than the other types of districts. Figure 1. Absolute Percent Change in Service Coverage 60 50 40 30 20 10 0 ANC TT2+ HF del. Birth Spacing EPI, FIC Vit. A, Utilization -10 Control CI CO Contracting also reduced the out-of-pocket expenditure on health services by the poor. Contracted-in districts significantly decreased per capita out-of-pocket expenses for the poorest half of the population while contracted-out and control districts appear to have slightly increased these expenditures. There was a significant decline in the per capita private out-of-pocket expenditure in the contracted districts, especially for the poor. Private out-of-pocket health care expenditures by the bottom half socio- economic group in contracted out districts fell by 70 percent. The reduction in out-of-pocket costs was greater among this population than among the overall population, indicating successful targeting of the desired beneficiaries and efficient transfer of subsidies. Out-of-pocket health care expenditures for households in the bottom half of the socioeconomic scale decreased by US$35 per capita per year, which is a very attractive return on about US$5 per capita per year public investment through contracting of services. This is particularly important in Cambodia, where most health care expenditures are private (estimates are around 75 percent of health care is paid out of pocket) and the high cost of health care is a leading cause of poverty and vulnerability. Sri Lanka Health Expenditure ADB's policy dialogue has always stressed adequacy of health expenditures. The evaluation report of the Second Health and Population Project had indicated that public expenditures were inadequate for proper maintenance of facilities. In 2001, the total Government health expenditures in Sri Lanka amounted to around 1.8 percent of GDP, which shows a small increase over the 1.6 percent recorded in 1990. The report found that the share of health in the budget was only 4.2 percent in 2000, declining from the 15 CASE STUDIES IN SCALING UP POVERTY REDUCTION average of 4.6 percent in the first half of the 1990s. The reduction in public health expenditures was influenced by general market based policies adopted since 1977 which encouraged a greater role for private provision and recovery of costs from users. The main effort of public health expenditures was to sustainably maintain public health infrastructure that was already in place and increase coverage of referral centers in rural areas to address the needs of the poor. Targeting government primary referral centers to rural areas in Sri Lanka appears to have had a beneficial effect on the rural poor who began making increasing use of them. Health Outcomes Data from post evaluation report of the Second Health and Population Project indicates that there was a strong inverse relationship between Government health expenditure and the average duration of stay for inpatients in the two main categories of PHC hospitals- district and rural. From 1993 to 2000, duration of stay fell by 26 percent and 30 percent in the district and rural hospitals respectively. This suggests that the quality of services in the lower levels of the referral system improved in the target areas during this period. Overall, the patients benefited socially and economically from having to spend fewer days in hospital, while health providers delivered their services more efficiently. As most of the PHC hospitals are located in remote areas, the main beneficiaries were poor families. The report also concluded that significant improvements have been made in the referral system in geographically disadvantaged areas and in upgrading the quality of services provided by key PHC training institutes in Sri Lanka. Improved trends in family planning practices were also observed. During the latter half of the 1990s, there was a steady increase in new contraceptive acceptors and in the contraceptive prevalence rate under the national program. The 2000 Demographic and Health Survey indicated that around 99 percent of married women had knowledge of at least one contraceptive method. The total fertility rate was 2.1 in 2000 compared with 2.6 per woman in 1990. One of the contributors to success was a highly effective in-service and on-the job- training was provided to the entire range of health service personnel. Training on counseling was provided, especially for midwives, and the rural poor went regularly to midwives and rural hospitals for family planning counseling. One key outcome pursued successfully by the Project in a national context was to reduce the problem of bypassing of district, peripheral, and rural hospitals - prior to the project they were very much underutilized. The proportion of registered live births in public hospitals increased from 85 percent in 1993 to 93 percent in 1999 which speaks well for the referral system. Between 1993 and 2000, malnutrition among preschool children declined sharply from 23.8 percent to 13.5 percent, while the prevalence of wasting declined slightly from 15.5 percent to 14 percent. The infant mortality rate decreased marginally from 17 to 16 per 1,000 live births, and the maternal mortality rate fell from 25 to 23 per 100,000 live births. While it is difficult to indicate in quantitative terms the Project's contribution, it seems to have had a significant impact on the health sector in qualitative terms. The performance appraisal report reported that by "training around 1,600 health workers, upgrading PHC hospitals and training facilities across the country, strengthening the referral system, and targeting support for antenatal care, child welfare, preventive medicine and family planning", the Project helped improve the delivery services of 16 PAPUA NEW GUINEA AND SRILANKA: SCALING UP HEALTH INTERVENTIONS the health and population sector to the poor. The Project also had significant institutional impacts. The strengthening of the planning and management capabilities of the health administration system at central, provincial, and district levels and the training of around 1,600 health workers had positive effects on the management and delivery of PHC services. Impact on Poverty The impact on poverty reduction was significant as the PHC services were directly targeted to reach the rural poor in isolated areas. Closer access to health services reduced travel time and cost, while improvement of health enhanced employment opportunities and reduced work absenteeism, thus having a direct impact on income generation and poverty reduction. It helped in stemming the decline in basic services for the poor. The Project reinforced Government's concerted efforts over several decades to achieve better health for all. In many ways, it contributed to achieving some of the Millennium Development Goals (MDGs) for health. Sri Lanka is one of the few countries in the South Asian region to have achieved MDG goals for health and social development ahead of time. Factors that Influenced Scaling-up in PNG, Cambodia and Sri Lanka Commitment, Leadership and Motivation: The PNG case suggests that many favorable factors were present which may have implications in the long-term. The full range of interventions has been brought to bear in the health sector ranging from extension of infrastructure to systems improvements and finally policy change. In all this, there was strong leadership at the central ministry level. This was supported by the donor community and resulted in testing new ways of partnership between the central and provincial level, the development Table 4. Sri Lanka Health Output Indicators (percent, except as noted) Year 1993 2000 Registered live births in public hospital 85.0 93 (1999) Malnutrition among preschool children 23.8 13.5 Prevalence of wasting (low weight for height) 15.5 14 Infant mortality rate, per 1,000 live births 17 16 Maternal mortality rate, per 100,000 live births 25 23 Total fertility rate (births per woman) 2.6 (1990) 2.1 Contraceptive prevalence rate (any method) 66.1 70.8 Inpatients treated in Rural PHC Hospitals (upgraded under Proj.) 39,280 72,432 Family planning new acceptors 169,689 214,627 One-year olds fully immunized against: tuberculosis 90 (ave. 1995-98) 97 (1999) measles 91 (ave. 1995-98) 95 (1999) Births attended by skilled health personnel 85 (1990) 95 (1996) Pregnant women immunized against tetanus 83.7 (1997) 89 (avg. 1995-97) Note: Years are 1993 and 2000, other than those specified. 17 CASE STUDIES IN SCALING UP POVERTY REDUCTION Source: ADB Project Performance Audit Report; 2002 Human Development Report; and 2003 ADB Key Indicators of an exemplary health policy and national health strategic plan and initial brainstorming about how best to structure a sustainable partnership between government and development partners. Necessary mechanisms for project monitoring and coordination were drawn up which also included provincial representatives and all interested stakeholders in the public health sector. Strong leadership was provided by the Secretary for Health, on all issues of importance covering health services delivery. The weakness, however, was that there were inadequate incentives to motivate service delivery staff at lower levels, especially provincial. The fact that the performance of other key government agencies was also crucial, including supportive interventions in education, transport etc., was borne out by the implementation experience of this project. Improved functioning and sustainable change in public service culture of the health sector cannot occur without the active involvement of the relevant government support agencies in addition to the health agencies directly concerned. Cambodia, on the other hand, was going through a period of reconstruction and recovery from a protracted and very violent conflict. Indeed, when the Basic Skills and the Basic Health Services Projects were approved, there was still a situation of conflict in some parts of the country. There was, however, a strong commitment on the part of all stakeholders to work together to restore peace and stability. This was enunciated clearly by the new government after the elections of 1998. At the same time, there was an understanding that while the level of health services for the poor was extremely low or non-existent, hard and realistic choices needed to be made. The Ministry of Health, working with WHO, built on the experience of other countries and an assessment of Cambodia's needs and developed a minimum package of services. This helped keep expectations realistic. Also, with the very low level of service, it was easy for the Ministry and development partners to show early success, which is an important incentive to continue investing in the system and reconstructing damaged facilities. The Project combined tried and tested elements such as training and infrastructure with new components such as contracting implemented on a pilot basis. This slow approach was realistic given the limited capacity in Cambodia, a country literally being rebuilt, but was also necessary to ensure political buy in. Some policy makers also saw contracting as an erosion of their area of influence and authority and policy makers often found it uncomfortable to deal with NGOs. One important factor for success in Cambodia is a commitment to invest more in primary health care. In Cambodia since the program was new, it was possible to start with a primary health care focus in poorer areas. It also required political will to prioritize in favor of the poor. In the Sri Lanka case, the Government's strong commitment to improve health care delivery was clearly articulated in the new National Health Policy adopted in 1992. It reinforced again Sri Lanka's traditional commitments to basic services provision, some of which had been affected by the shift to market based policies. The second ADB project was one of the main instruments through which the new policy was implemented successfully in targeted areas. As a continuing endeavour to maintain the gains of the last few decades in human development in general and health and population in particular, the Government ensured that all loan covenants were generally complied with and adequate counterpart funds were provided and sustained. In its poverty reduction strategy (PRS) the Government declared its intention to modernize its health care services and has committed itself to continuing the work undertaken under the Project to enhance institutional capability at all levels and promote private sector participation in both preventive and curative health care. 18 PAPUA NEW GUINEA AND SRILANKA: SCALING UP HEALTH INTERVENTIONS Innovations in Lending and Service Delivery In PNG, the sector development loan modality was an innovation in the PNG context. Its non-prescriptive nature and the diversity of objectives made it poorly understood, especially outside of the DOH. Since HSDP was widely seen as ADB's participation in the Economic Recovery Program (ERP), there was not uncommon belief that the sector development program loans were intended to be budgetary support, albeit to the health sector, similar to the assistance from the other contributors to the ERP. At different times, Department of Finance's (DOF) top management felt that the loan proceeds could be used to cover DOH's recurrent budget without the increase in funding necessary to implement the National Health Plan. Protecting funds under it became difficult under the general budgetary scarcities. While the sector development program modality has many advantages when driven by a well-coordinated and determined program of action, giving flexibility to the government to use available funds for high priority areas of intervention, it can break down when the general governmental system particularly at local levels is regressing. Given the obvious stresses and dislocations of fiscal decentralization and lack of coordination that was quite clearly evident, the situation called perhaps for continuation of pre-appraised investment projects that left no ambiguity about manner of use of funds, as in the case of Sri Lanka and Cambodia where all projects supported by ADB were investment projects with simple and clearly specified components. In the case of Cambodia, ADB's health sector investments combined established project components with innovative experiments in contracting for health system and with other proposed innovations in health finance. This innovation and the willingness of the Government to try pilots clearly contributed to the success of the contracting pilot. The design of the contracts themselves contributed to the success of the pilot. At the center of contracting was the assurance that providers are adequately compensated for their services and effectively supervised and supported. NGOs working in contracted-out districts revised the salaries of health care providers and brought them in line with average salaries in the private sector. In return, the NGOs required the providers to work full time in health facilities and have no private practice. In contracted-in districts, the NGOs supplemented providers' salary through their own funds and by allocating a larger share of user fee incomes. This ensured that providers in poor areas were not at a disadvantage only because they were working with people with low paying capacity. Market- based salaries motivated the staff to work in remote and poor areas and made it easier for the contractors to enforce full commitment from them. Regular supervision and support from the District Health Management Teams kept their motivation high and ensured that they performed as per the agreement. Decentralization: The design of HSDP in PNG was overtaken by decentralization embodied in the New Organic Law, which affected the health sector also. Responsibilities were devolved first to provincial governments then to districts. As a result of this process, the Department of Health lost control over the delivery of health services at sub-national levels. At the same time, capacities in districts were extremely limited, because they lacked (i) technical knowledge, (ii) staff (who were to be transferred from provinces but actually remained in their positions for the lack of physical facilities at district level), and (iii) financial resources (funds were not transferred from provinces to districts). In summary, the implementation of the New Organic Law was not adequately prepared and lacked support from the central level. The funding levels to local governments initially envisioned under the New Organic Law were unrealistic and unattainable. 19 CASE STUDIES IN SCALING UP POVERTY REDUCTION The framework of the Grants to local governments sought to establish both equity and respect for Central government priorities but the low funding upset these objectives. As a result, performance was varied and, in the period 1999­2001, one third of the provincial governments were suspended for poor financial management and failure to deliver services. Other provincial governments experienced similar problems and, although not suspended, they were largely inoperative. Funding responsibility was given to the provincial administrations but most failed to allocate funds adequately to health care. Authorities in Sri Lanka appeared to have managed the decentralization well and retained good supervision over the process. A constitutional amendment in 1988 established provincial councils responsible for the formulation, implementation and monitoring of provincial health plans based on national policies and strategies. The second Sri Lanka project was therefore implemented in an already decentralized environment and did not suffer as much from the trauma of decentralization. The post evaluation report for the Second Health and Population Project in fact felt that the Project had helped the fiscal decentralization process, which called, inter alia, for improved governance at the provincial level through its HRD components, which have benefited mainly provincial institutions. The quality of policy planning and program budgeting has improved, and so has the capacity of planning units under the provincial directors of health to undertake performance monitoring. This is shown by the quality of annual planning documents, which have become more comprehensive, analytical, and result-oriented, thus allowing greater accountability, and supports the Government's move toward performance-based budgeting. Governance Growing influences hindered initiatives to improve upon governance and management in PNG. Opposing pressures included the rising consumer price index and weakening purchasing power of public servants' salaries, coupled with the poor management established by weak government leadership over the years. Cultural parameters such as clan affiliation and the pervasive wantok12 system contributed to limiting good governance. An independent provincial level performance audit highlighted governance issues. This has served to underscore the need for a similar review process in DOH. The report established that controls over procurement and expenditure within DOH were ineffective and were not conducive to efficient operation. Only incremental progress has been achieved thus far, but with the potential to positively influence public service culture. Again, this contrasts sharply with the Sri Lankan case. An ADB impact evaluation study for the health sector13 has noted that "the contrasts in impacts between PNG and Sri Lanka could be attributed to governance. Sri Lanka's stronger social services served as a foundation that the ADB projects built on to succeed, whereas PNG's weak governance cut into the projects' effectiveness." Sri Lanka was able to devote its scarce budgetary resources in a cost effective manner. This was also reflected, for instance, in the Second Health and Population Project's own implementation record. Despite civil unrest, construction works were completed with only minor delay. The actual project cost was 16 percent lower than the 12 In most traditional settings, the wantok system provides for an egalitarian sharing of wealth and responsibilities. It provides a kind of safety net with strong community obligations. While the wantok system provides advantages for the clan, it often does so at the expense of modern and social obligations, which go beyond the boundaries of one's extended family or clan. In the public service, the wantok system leads to much conflict of interest and nepotism. 13ADI Series No.2: Fighting Poverty Through Better Health Care. 20 PAPUA NEW GUINEA AND SRILANKA: SCALING UP HEALTH INTERVENTIONS appraisal estimate, part of which is explained by lower savings in procurement. The coordination between the PMO and provincial authorities was good and no undue bureaucratic delays took place in the large subprojects. External Catalysts: Interventions in the health sector do not happen in isolation. External factors always play a major role in determining policy success. In addition, as previously argued the health sector is only one of many factors contributing to health and is only effective if households take advantage of the services and information offered by the health system. In the case of Papua New Guinea, many external factors intervened to complicate the reforms proposed under HSDP. In particular, the Project was introduced in a time of serious economic and fiscal problems. While development partners were active in the country to try to limit the social consequences of these reversals, their support is often not sufficient to compensate for the lost income and development opportunities. The general trend of health indicators was negative partly due to the general economic problems that the country faced. Cambodia, in contrast, benefited from a number of positive trends. This included economic growth, as a result of stability and the "peace dividend", and a general improvement in the health status of the population as a result. Although there are serious concerns about governance, development partners have generally been supportive of the country and have provided generous financing for development programs including in the health sector. The capacity of the Government and the Ministry of Health has increased noticeably in the past decade, allowing them to develop realistic strategies and implement them given the country's real resource constraints. In Sri Lanka, the projects were isolated from the internal civil strife hence remained unaffected. External development partners have played an important role in all three countries. ADB and other development partners have provided critical assistance in formulating policies and shaping investments. In Cambodia, by the time, HSSP was approved in 2002, there was a growing consensus about the need for partners to work together under the Government-owned strategy. The health sector strategy was built on existing policies and was an attempt by the Ministry of Health to consolidate and up-scale these policies, rather than to start from scratch. The partnership of ADB, DFID, and the World Bank also builds on existing efforts and was not an attempt to develop a "textbook" SWAP but rather to improve coordination and cooperation in the short term. In Cambodia, the Government and partners were able to channel the real commitment to rebuild the country and improve access of health service for the poor into a set of realistic policies and actions designed to continue the positive momentum. Lessons Learned Targeting the Poor Establish a process whereby the needs of the poor are incorporated into development planning and implementation, through monitoring and granting specific incentives. The PNG study had highlighted the wide differences in delivery of health services between the poor and better off households. This calls for particular emphasis to reach the poor particularly poor women and children in rural areas. Their needs should be explicitly considered in government and donor policies. Health interventions should be so 21 CASE STUDIES IN SCALING UP POVERTY REDUCTION targeted that they would benefit the poor in terms of diseases that mainly affect them, facilities that they are more likely to use and geographic areas where there are concentrations of the poor. For example in the PNG case targeting should involve focusing on acute respiratory infections which the affect the poor; emphasizing highland areas with concentration of the poor; and strengthening of aid-posts and missions on which the poor rely heavily. But more important than specific interventions is the establishment of a process whereby the needs of the poor are incorporated into development planning and implementation. The projects in Cambodia specifically targeted the poor both in their design (focusing on poor and underserved districts) and through the monitoring framework that gave incentives to contractors that had impact on the poor. The evaluation suggests that this approach was successful. Similarly, the Sri Lanka projects specifically focused on primary health care in rural areas which benefited the poor. Performance indicators and an evaluation system Projects should develop a set of monitoring indicators and a framework to evaluate the results. This can be used to increase incentives for providers and to build political support for scaling up successful interventions. One critical element in the success of Cambodia was the use of a set of performance indicators that could be used to monitor the progress of the providers (in this case, NGOs) and also to make the case to Government and other development partners on successes in the health sector. Contracting requires clear agreements on deliverables and an enforceable contract. It also requires an independent and mutually agreeable performance verification system. In the Cambodia pilot, targets for thirteen key health indicators were agreed upon. These indicators specifically included measures for improving access to health services by the poor. The progress towards achieving these targets was measured through independent household surveys and spot checks by Government staff. Payment for contracts was linked to achievement of targets. When performance exceeded the minimum level, the contractor was entitled to a bonus payment. The same data were also used to carry out rigorous evaluations that were widely disseminated. This played a central role in building political support for scaling up the contracting and in attracting new sources of finance. Performance monitoring was also emphasized in the capacity building efforts under the second Sri Lanka project. Decentralization and Basic Services Delivery Decentralization can be a powerful instrument to increase local ownership and participation. It should be realistic in its design and phased in carefully so as to not create too many disruptions in transferring services and responsibilities, but also allow for strong local involvement and participation. Decentralization brings opportunities but also risks. Apart from the issue of developing a proper fiduciary environment to prevent malfeasance, there is the general issue of capacity building for basic services delivery. This also has implications for attainment of MDGs in general as many countries in Asia have decentralized basic services delivery functions to decentralized levels without adequate capacity building. The PNG and Sri Lanka examples have both good lessons to offer. An important cause of the lackluster performance of the health sector in PNG is the breakdown of the vertical integration of health services brought by the New Organic Law. Decentralization can alleviate overloading of central government and improve access to decision-making and participation by more people. However, decentralization can also lead to deterioration in the use and control of resources if the administrative capacity is lacking, particularly at local levels if it leads to loss of previously available central expertise, or if there is 22 PAPUA NEW GUINEA AND SRILANKA: SCALING UP HEALTH INTERVENTIONS insufficient prioritization for national level goals, which is the case in PNG. Many provincial authorities did not prioritize health services through budget support. Provincial governors had all agreed to allocate 15 percent of provincial revenue to health; yet none did so. The Organic Law did not allow for prioritizing funds to poorer provinces. There was inadequate supervision from the central level. While clearly the PNG case suggests lack of sufficient preparation for decentralization, the Sri Lanka case illustrates a different issue connected with decentralization. It was noted at the evaluation of the Second Health and Population Project that the central government project managers were hesitant to part with authority and maintained strict supervision. While the authorities' response could be understood in the context of perceived weaknesses in capacities at decentralized levels and need to maintain strict monitoring over a matter of high national priority, these considerations need to be balanced with local needs and priorities, calling for better participation by all relevant parties. Cambodia is introducing decentralization gradually and most financial responsibility remains at the central level. Health Financing The distribution of subsidies and the organization of the health finance system can have a major impact on poverty and projects should take explicitly take this into account when designing interventions in the health sector. How health is financed has a major impact on the poor. Although the private sector can play an important role in providing health care, the high cost of health care can also be a major burden on the poor and is often a leading cause of poverty. In Cambodia, better management of the health care, including reduction of informal fees and the establishment of a transparent fee schedule have aided this process. In Sri Lanka, the state's comparatively low spending on health was at the cost of high private expenditures and by cost-sharing by the users of services. However, the Government's emphasis on extension of public primary health services in rural areas encouraged a large increase of the use of these services by the poor. In Papua New Guinea, however, it appears that despite increase in funding, insufficient health outcomes occurred particularly outcomes for the poor. Capacity and Human Resource Development and Management Capacity constraints should be specifically addressed. Project designs should be realistic and should take into account the capacity of governments. Capacity development is crucial for delivery of health services. In PNG, the health sector suffered from the general malaise of weak governance, and HR management in the PNG Government. This included inadequacy in data collection, planning, supervision, enforcing transfers, defining roles and responsibilities, control of absenteeism and failure to implement retrenchment programs. This lack of capacity was sought to be bypassed by engaging contractual staff paid for by external technical assistance and project funds. Public service reforms addressing personnel management issues should reach across all government agencies and all levels of government in order to have a permanent impact on public service culture. In Cambodia, the delivery capacities in the Government were also weak. However, the design of the projects took the inadequate Government capacities into account. The initial design of the Project was quite simple and focused on increasing capacity and generating the right set of incentives. In Sri Lanka, emphasis on human resource development was part of a long learning process. Initial health sector interventions had neglected HR issues: this highlighted in the evaluation of the first ADB project. However, the second project successfully addressed this issue. The Government learnt that improving the 23 CASE STUDIES IN SCALING UP POVERTY REDUCTION technical and managerial skills of human resources contributed to improving the delivery of PHC services, including the referral system. This was achieved through domestic and overseas training programs. Scaling-up to a Sector-wide Approach Partnerships can play an important role in coordination and lowering transaction costs for the Government and development partners, but their design should be realistic. Rather than simply trying to apply the best practice examples used in other countries to develop a SWAP, the scaling-up process should be gradual and be built around the capacity of local stakeholders. In all three countries, delivery of health services (supported by ADB and other donors) have scaled up health interventions beginning from improving health infrastructure to quality improvements and human resources development and gone further into policy reform in the case of PNG with limited success. Both Cambodia and Papua New Guinea, there is interest in utilizing sector wide approaches (SWAPs) for the health sector. Much still is required to be done in Papua New Guinea, particularly in coordinating the multiplicity of interventions and the different agencies in the Government involved with health related issues and also different approaches of partners. Although there is an on-going SWAP, the conditions in Papua New Guinea may not be right. Managing a SWAP requires a high degree of capacity in coordination and close cooperation amongst all agencies. In Cambodia, the goals of the partnerships have been more modest, focusing on improving coordination and working to implement the Government health sector strategy. 24 PAPUA NEW GUINEA AND SRILANKA: SCALING UP HEALTH INTERVENTIONS References Abel-Smith B., Rawal P. 1992. "Can the poor afford `free' health services? A case study of Tanzania." Health Policy and Planning. 7: 329-41. ADB (Asian Development Bank). 1991a. Report and Recommendation of the President to the Board of Directors on a Proposed Loan and a Technical Assistance Grant to the Independent State of Papua New Guinea for the Third Rural Health Services Project, 7 August 1991. Manila. _____. 1992b. Report and Recommendation of the President to the Board of Directors on a Proposed Loan and a Technical Assistance Grant to the Democratic socialist Republic of Sri Lanka for the Second Health and Population Project, 26 October 1992. Manila. _____. 1993c. Project Performance Audit Report. Health and Population Project in Sri Lanka, August 1993. Manila. _____. 1997d. Report and Recommendation of the President to the Board of Directors on a Proposed Loan and a Technical Assistance Grant to Papua New Guinea for the Health Sector Development Program, February 1997. Manila. _____. 1999e. Progress Report on the Health Sector Development Program (Loan No. 1516-PNG; Loan No. 1517-PNG), in Papua New Guinea. November 1999. Manila. _____. 1999f. Poverty Reduction Strategy. Manila. _____. 1999g. Policy for the Health Sector. Manila. _____. 2001h. Assessing Development Impact (ADI), 2001. Manila. _____. 2002i. "Trends in the Distribution of Health and Health Services in Papua New Guinea: 1991- 2000" conducted under ADB TA 3762 - PNG: Health Sector Review. Manila. _____. 2003j. Project Performance Audit Report on the Second Health and Population Project (Loan 1189 ­ SRI) in Sri Lanka. September 2003. Manila. _____. 2003k. Millenium Development Goals in the Pacific, Relevance and Progress. Manila. _____. 2003l. Country Assistance Programme Evaluation for Papua New Guinea. Manila. _____. 2003m. Moving Toward a Sector-Wide Approach. Papua New Guinea. The Health Sector. Manila. AusAID/Independent Review Team, 2002, Papua New Guinea Health Sector, Annual Sector Review, Final Report Besley T, Kanbur R. 1993. "The Principles of Targeting." in Lipton, M. and J. Van Der Gaag (eds) Including the Poor. Washington, D.C.: The World Bank. Bhushan, I., E. Bloom, Nguyen Hai Huu, and Nguyen Minh Thang. 2001 Human Capital of the Poor in Viet Nam. Manila: Asian Development Bank. 25 CASE STUDIES IN SCALING UP POVERTY REDUCTION Bhushan I, Keller S. Schwartz J.B. 2002. Achieving the Twin Objectives of Efficiency and Equity: Contracting Health Services in Cambodia. ERD Policy Brief, Number 6. Manila: Asian Development Bank. Bloom, D, D. Canning, and D. Jamison. 2004. "Health, Wealth, and Welfare." Finance & Development Vol. 41, No. 1: 10-15. Carr, D. 2004. "Improving the Health of the World's Poorest People" Health Bulletin. Number 1. DoH. 2000. PNG-Australia Study Mission to South Africa and Ghana: Health Sector-wide Approach _____. 2001a. Minimum Standards for District Health Services in Papua New Guinea. _____. 2001b. National Health Plan 2001-2010; Health Vision 2010: Volume I, Policy Directions and Priorities. _____. 2001c. National Health Plan 2001-2010; Health Vision 2010: Volume II, Program Policies and Strategies . _____. 2001d. National Health Plan 2001-2010; Health Vision 2010: Volume III, Provincial and District Health Profiles. _____. 2001e. National Health Plan 2001-2010; Health Vision 2010: Volume IV, A Guide to Implementing the Plan. _____. 2002f, Annual Health Sector Review National Report: 2001 Performance Review. _____. 2002g. Health Medium Term Expenditure Program, summary paper. _____. 2003a, Annual Health Sector Review, Preliminary national report, 2002 Performance Review. DoH/SMRG. 2002a, Annual Sector Review. _____. 2002b. National 2001 Performance Review. Dwyer, G. 2004. "PNG's Quiet Killer," ADB Review. April, 2004. Global HIV Prevention Working Group. Access to HIV Prevention: Closing the Gap. Supplement to Foreign Affairs. May, 2003. GoPNG. 25 Sept 2003, Health Sector Medium Term Expenditure Framework, 2004-2006. _____. 1996. Organic Law on Provincial Governments and Local-level Governments. GoPNG/ADB. 1997a, Loan Agreement for Loan Number 1516 (HSDP). _____. 1997b, Loan Agreement for Loan Number 1517 (HSDP). GoPNG/AusAID, 2002, HSIP Trust Account Management Review. GoPNG/HSSP, 2001, Provincial Health Finance Review. Gwatkin D. 2000. "Health inequalities and the health of the poor: What do we know? What can we do?" Bulletin of the World Health Organization. 78: 3-18. 26 PAPUA NEW GUINEA AND SRILANKA: SCALING UP HEALTH INTERVENTIONS HERA/ADB. 2002a, Health Sector and SWAp Review, PNG Volume I. _____. 2002b, Health Sector and SWAp Review, PNG Volume II. _____. 2003c, Development of a Sector-wide Approach in the Health Sector of Papua New Guinea. Hiawalyer, G. and M. Spohr. 2002. Burden of Disease. Port Moresby, Papua New Guinea. National Department of Health. Hsaio, W. 2004. Health Care for the two billion poor. Is Community Financing a Viable Option? World Bank-Health System Development Newsletter. April 22, 2004. Jha P, Mills A, et al. 2002. "Improving the health of the global poor." Science 295 (5562): 2036-9. Ministry of Planning. 2003. Cambodia Millennium Development Goals. Phnom Penh: Ministry of Planning. National Institute of Statistics, Directorate General for Health (Cambodia), and ORC Macro. 2000. National Health Survey. Phnom Penh: National Institute of Statistics. NEFC. 2003. Brief on 2004 Proposals, September 2003, National report, National Economic and Fiscal Commission. PSRMU. 2001a. Functional and Expenditure Review of Health Services: Interim Report on Rural Health Services, draft version 7 ­ 20th November 2001. _____. 2001b. Making the Public Sector Work: a Medium Term Plan for Public Sector Reform in Papua New Guinea 2000-2003, Part A: A Strategy for Public Sector Reform. _____. 2001c. Making the Public Sector Work: A Medium Term Plan of Action for Public Sector Reform in Papua New Guinea, 2000-2003: Part A: A Strategy for Public Sector Reform, and Part B: Programmes and Activities. Sen A., Keynote Address, World Health Assembly, 18 May 1999" Health in Development. Bulletin of the World Health Organization, Vol 77, No. 8, 619-623. SMRG. 2002a. Annual Sector Review; Analysis of Process. _____. 2002b. Annual Sector Review; Independent Review Team Final Report. Wagstaff A. 2000. "Socioeconomic inequalities in child mortality: comparisons across nine developing countries," Bulletin of the World Health Organization. 78:19-29. Weiss, C. 1998. Evaluation. Upper Saddle River: Prentice Hall. World Bank. 1993a. World Development Report, 1993. Washington, D.C. _____. 2003b. World Development Report, 2004: Making Services Work for the Poor. Washington, D.C. 27