Why performance-based funding of health services in developing countries is getting more attention
DEVELOPING countries and their international partners are increasingly adopting methods of financing health care activities in developing countries that link the availability of funding to concrete, measurable results on the ground. Such “performance-based” financing was advocated a decade ago in the World Bank’s 1993World Development Report—Investing in Health and other policy documents in the early 1990s, although relatively little practical experience with this type of financing was available. Since then, much experimentation has taken place, and we are seeing with growing clarity the important potential—as well as the challenges—of performance-based financing for achieving national and global health goals.
Governments and partner agencies are interested in performance-based financing for health for a number of reasons. First, there is a growing focus worldwide on achieving measurable results with development assistance, and performance-based financing spotlights such results. In terms of health care, these results are being closely tracked as governments and their partners strive to achieve the Millennium Development Goals (MDGs). The goals include reductions in child and maternal deaths; reductions in rates of infection from HIV, malaria, and tuberculosis; and improvements in the nutritional status of children. Governments and their partners are thus naturally attracted to the idea of providing funds for programs that achieve or make progress toward the MDGs in health or that at least show increases in some of the key services needed to reach the goals. For example, where immunization and prompt treatment of pneumonia are crucial for halting child deaths, funding for health care might be tied to advances in the coverage of these services.
Second, even though external funding for health care in developing countries is currently in excess of $8 billion a (Michaud, 2003), substantially greater development assistance will be needed to reach the health MDGs. Politicians and legislators in donor countries are under growing pressure from their constituencies to show that development assistance budgets, in health as in other areas, are having measurable results. Partner agencies are thus seeking to increase the effectiveness of these resources by allocating them to countries and programs that demonstrate progress as measured by performance indicators.
Third, linking the availability of financing to measurable results—whether in terms of changes in health status or in the coverage and quality of health services—is consistent with the objective of making service providers more accountable. Increasing accountability of service providers to clients in low-income communities and to government policymakers is the theme of the 2004 World Development Report—Making Services Work for Poor People. Linking financial payments to getting the job done—immunizing infants, treating tuberculosis patients, or testing more young men and women for HIV and counseling them on their status—can be a tremendous incentive for those providing the services, not least because it exposes their performance to their clients and others footing the bill. The 1993 World Development Report advocated the expanded use of public monies to pay private nongovernmental and for-profit doctors and clinics to deliver basic health services to the poor. Performance-based contracts between the government and these private providers are the principal instrument for putting this recommendation into practice.
Performance-based financing in health is now being widely and actively tested at several levels of the health care system. Here are some examples: (1) developing country governments pay health care providers in nongovernmental organizations (NGOs) and the private sector for delivering essential health services to poor households; (2) central governments determine the transfer of funds to local governments on the basis of their performance in strengthening health services; and (3) donors release funding (disbursements) to recipients in developing countries as and when certain key health targets are achieved.
Performance-based contracts with NGOs. A number of governments in low-income countries are funding NGOs to deliver basic health services on a performance basis. Many of the earliest experiments are from Latin America and the Caribbean. In Haiti, for example, NGOs were contracted to provide child health and family planning services (World Bank, 2001). They were given an advance each year and then a quarterly sum, based on a negotiated budget. At the end of the year, performance was measured against various indicators, including immunization coverage, percentage of families using oral rehydration to treat acute diarrhea, number of pregnant women attending prenatal care, and average waiting times in clinics. The NGOs’ performance determined the bonus they received, which could be up to 10 percent of the original prenegotiated budget. As a result, the Haitian NGOs made changes in their service delivery schemes and improved their performance, especially in immunization and oral rehydration. In Guatemala, the government is implementing a large performance-based program with NGOs that currently covers nearly four million persons, mostly among the country’s indigenous population (see Box 1). Other schemes have been implemented in Argentina, El Salvador, and Nicaragua.
Countries in South Asia are also moving into performance-based health programs with NGOs. In the Islamic State of Afghanistan under a recently approved project for health service rehabilitation financed by the World Bank, NGOs are being contracted by the government to run health centers. NGOs that achieve specific targets will be eligible to receive additional payments of up to 10 percent of their baseline subsidies from the government.
Box 1Guatemala leads the way
Guatemala has successfully implemented, on a large scale, the contracting of nongovernmental organizations to deliver health services. The government started the Program to Extend Coverage of Basic Health Services (PESCB in Spanish) in 1997, soon after the end of a long civil war. The program has continued under successive administrations. By 2000, 89 NGOs under 137 separate contracts provided health care to about 3.7 million of Guatemala’s population of 14 million.
The contracts specify a range of maternal and child health services and prevention and treatment of a number of diseases, including malaria. The NGOs are paid about $8 for each person served, mostly in cash but also in kind, in the form of such items as vaccines and medicines. Payments are released quarterly, once performance has been checked and verified.
Performance is measured by a series of indicators, including coverage of immunization and prenatal care, distribution of iron sulfate tablets to pregnant women and children, and frequency of home visits by the NGO outreach staff. Private firms have been hired to develop the monitoring system, which also looks at the accounting practices of the NGOs.
The contracting system under the PESCB appears to have produced important gains in health service delivery. Immunization rates in Guatemala rose from 69 percent to 87 percent between 1997 and 2001. Household surveys now under way will be able to assess the impact of the program on mother and child health outcomes.
During the early years of PESCB, a number of obstacles had to be overcome. Government health workers resisted the scheme because they feared that contracting with NGOs was a hidden form of privatization of the government health services. The NGOs were initially reluctant to get involved, too, because they felt that the government was demanding too much in the way of improved performance, and they doubted that the government would pay them in a timely manner. Given the financial fragility of many local NGOs in Guatemala, it was critical for the government to make advance payments and release quarterly payments without delay to build confidence in the relationship between the public and private sectors in the country.
In a similar vein, the central and state governments in India have started to reimburse NGOs and private providers according to their performance. The national tuberculosis program reimburses private laboratories for testing sputum samples to detect tuberculosis and also pays NGOs and private doctors a fixed sum for each infected patient who is cured using the directly observed short course therapy (DOTS) approach. In Kerala state’s Kannur district, where this scheme is well advanced, NGOs and private providers have helped boost the case detection rate (the share of those with active tuberculosis who are diagnosed and treated) by about 20 percent (Granich, 2003). In the state of Tamil Nadu, the government is testing another program under which private doctors are paid for providing family planning services and for performing cesarean sections to deliver babies.
Box 2GAVI pioneers performance-based grants
Eight developing countries will receive $15 million in performance-based payments for their achievements in increasing immunization rates as part of a groundbreaking new grant program approved in December 2003 by the Global Alliance for Vaccines and Immunization (GAVI) Board. The eight are Azerbaijan, Ghana, Mali, Pakistan, Rwanda, Tajikistan, Tanzania, and Uganda. These countries’ externally audited health data show that they have succeeded in reaching more of their children with life-saving vaccines in the past three years.
Under the program, countries applied for grants by submitting to GAVI their long-term strategies for reaching more children. The GAVI Board approved successful applications for three years of investments in the countries’ immunization systems. These funds could be used in any way countries deemed most appropriate; the only requirement was results. In the fourth year, additional funding is available only to countries that have actually reached more children.
For example, when Tanzania first applied to GAVI in 2000, the country was immunizing 950,000 infants annually, or 74 percent of those born, with three doses of the diphtheria-tetanus-pertussis vaccine (DTP3) used as an indicator for basic immunization coverage. By 2002, 1.2 million infants, or 89 percent of those born, had access to DTP3. This success means that, in addition to the $2.4 million investment provided between 2001 and 2003, GAVI will provide an extra $3 million to Tanzania in 2004.
To date, 16 countries have each received three years of immunization system investments from GAVI, Ten of them—Armenia, Burkina Faso, Cameroon, Côte d’Ivoire, Haiti, Kenya, Liberia, Madagascar, Mozambique, and Sâo Tomé and Príncipe—will not qualify for performance-based payments. Either these countries were not able to increase their immunization rates in the past three years, or their reported coverage data could not be externally verified. As soon as the countries are able to turn around their performance, however, the funding will start to flow again.
Central government transfers to local authorities. In the World Bank—supported Family Health Project in Brazil, the central government is making per capita transfers to the local municipalities on the basis of planned increases in certain services, such as safe delivery of babies for low-income women, monitoring of infants’ nutritional status and growth, and treatment of poor children for various illnesses. For example, at least 40 percent of babies should be delivered in maternity facilities managed under the government’s family health program; participating outreach workers should provide an average of at least nine home visits to targeted low-income families each year; and all doctors enrolled in the program should undergo special training. If the municipalities reach these targets and several others, they will continue to be eligible for future financial transfers; otherwise, the level of central government support will be reduced and other remedial measures put in place in an effort to improve the targeting and effectiveness of the activities of those underperforming municipalities.
Donor disbursements to national governments and other recipients. A number of innovative approaches are in place that make donor financing of health programs conditional on successful performance on the ground. One example is the World Bank’s credit “buy down” program for polio eradication. Under the program, countries receive low-interest loans to purchase polio vaccines in an effort to eliminate the last remaining pockets of the disease that persist in Africa and South Asia. If the vaccine is judged to be purchased, delivered, and administered in a timely and effective manner, additional resources in a trust fund financed by the Bill and Melinda Gates Foundation, the United Nations Foundation, and Rotary International are used to buy down the interest and principal repayment on the loan, thus converting it to a grant. So far, Nigeria and Pakistan have initiated polio eradication projects for about $50 million. If they are successful, about $20 million from the Gates-Rotary trust fund will be used to transform the World Bank loans to pure grants. In this way, every dollar from the foundations will leverage $2.00–$2.50 of external assistance for the polio program. This leveraging has tremendous potential and gives an important incentive for donors to get involved.
The Global Alliance for Vaccines and Immunizations (GAVI) has been a pioneer in the performance-based approach to grant assistance (see Box 2). Through its sister organization, the Vaccine Fund, which raises and disburses funds for the alliance, GAVI provides commodity assistance to countries in the form of new and underused vaccines (hepatitis B, hemophilus influenza type b, and yellow fever, with new products for rotavirus and pneumococcus to follow) and safe injection supplies. In addition, GAVI allocates grant funds to countries that increase coverage rates for diptheria/pertussis/tetanus (DPT3). Countries’ applications to GAVI specify current coverage levels. On the basis of these data, performance is assessed annually, and countries receive $20 for each additional child immunized with DPT3. This year, GAVI will make its first payment for performance verified through externally audited health data. Eight countries will receive $15 million in performance-based payments for increasing immunization rates. The Global Fund to Fight AIDS, Tuberculosis, and Malaria is also planning to disburse its financing to dozens of countries for disease control activities on the basis of measured changes in program performance. The Global Fund is currently refining its monitoring arrangements to do this effectively.
Learning what works
The recent experience with performance-based financing in health has been encouraging. When properly designed, performance-oriented “contracts” can help to stimulate individual providers—doctors, nurses, midwives, village health workers—to expand their coverage, reach poor people, and enhance the quality of what they do. When the contract is between a central government and local authorities or between an international development assistance agency and a government, improvements in program performance can also be stimulated. Performance-based financing is helpful in focusing all parties on the services produced and their impact on the health and nutritional status of the intended population, rather than simply on counting inputs such as drugs, doctors, ambulances, hospital buildings, and equipment.
But performance-based financing for health must also overcome a number of serious hurdles to work well. One is the difficulty of measuring performance quickly and accurately. Data on such key outcomes as maternal mortality are rare in many countries, and even intermediate indicators, such as the number of women who have their babies under adequate medical care, can be hard to monitor in the poorest regions and countries. Much work is required to raise the quality and comprehensiveness of national monitoring systems to track health performance, but countries like Albania and Tanzania (Settles, 2002) show that this is possible. When a government’s own performance is being assessed, and the results are tied to a financial reward, it is perhaps unrealistic to expect that a government-run monitoring system will be wholly objective. For this reason, it sometimes makes sense to commission an independent institution to do the monitoring, as in the cases of Haiti and the polio buy-down schemes in Nigeria and Pakistan.
Another related problem is the widespread lack of capacity in ministries of health to design, negotiate, and enforce performance contracts with NGOs and private health care providers. In most countries, the ministry of health has traditionally seen its role as one of owning and operating its own hospitals and clinics. The task of managing thousands of contracts with private health care providers and paying for services, such as good prenatal care, or treating children for acute respiratory infections is daunting for most ministries. A major shift in the ministry’s fundamental mission and operating mode is required to implement large-scale performance-based systems.
A final hurdle is the risk that performance-based financing might be perceived as a harsh or an unfair imposition of conditions by the financing source on the health service providers. Since relations between ministries of health and NGOs are already strained in many developing countries, successful performance-based contracts tend to be flexible and respectful, allowing NGOs to operate with some freedom and ensuring that government payments for good performance are timely. Similarly, performance-based financing of governments by donor agencies should be structured to benefit both parties and so avoid “conditionality” that is not embraced by the government.
Performance-based financing for health is likely not only to continue but to expand. This trend is being spurred on by several factors. They include government and donor concern for health outcomes; interest in improved measurement of results; the push for greater accountability of health care providers to their clients and to governments and for stronger accountability of governments to donor agencies; and a recognition that NGOs and the private sector can, in some cases, deliver essential health services to poor people more efficiently than the public sector. It is vital for the development community to continue to monitor closely these promising experiments in performance-based financing and to disseminate and apply the lessons of success and failure as rapidly as possible to maximize the benefits of development assistance in pursuit of the health Millennium Development Goals.
Robert Hecht is Sector Manager and Amie Batson and Logan Brenzel are Senior Health Specialists in the World Bank’s Health, Nutrition, and Population Department.
GranichRuben2003 “The Private Sector and the Revised National Tuberculosis Control Program” (Delhi: World Health Organization).
La ForgiaGerardPatriciaMintz and CarmenCerezo (forthcoming) “Contracting for Basic Health Services: A Case Study of the Guatemalan Experience” (Washington: World Bank).
MichaudCatherine2003 “Development Assistance for Health: Recent Trends and Resource Allocations” (Cambridge, Massachusetts: Harvard Center for Population and Development).
SettlesPhilip2002 “Improving Monitoring and Evaluation in the Tanzania Essential Health Investment Program” (Dar es Salaam: Government of Tanzania).
World Bank2001Output-based Health Care: Paying for Performance in Haiti (Washington).