35662 The World Bank PREMnotes F E B R U A R Y 2 0 0 6 N U M B E R 1 0 6 PUBLIC SECTOR Contracting for Delivering Health Current evidence indicates that contracting for health services is effective in improving ac- cess to and quality of services. Background · Using competition to increase effective- Substantial improvement in the delivery of ness and efficiency health services will be necessary to achieve · Allowing governments to focus more on the health-related Millennium Development other roles that they are uniquely situated Goals (MDGs). For example, it has recently to carry out, such as planning, standard Contracting has been estimated that 63 percent of child deaths setting, financing, regulation, and various in developing countries could be prevented public health functions. many attractive through the full implementation of a few effec- Critics of contracting have raised a num- features, but tive and low-cost interventions. Unfortunately, ber of concerns, including: the performance of the public health systems · Contracts will not be feasible at a suf- success requires that could deliver these interventions remains ficiently large scale to make a difference addressing some weak in many countries. While there is likely at country level practical issues a need for additional resources to strengthen · NGOs and other nongovernment enti- health systems, previous experience suggests ties will not want to work in remote or that simply throwing money at the problem of difficult areas and are less capable of service delivery is unlikely to have much of an providing services to the very poor, thus impact. Another response to the challenge increasing inequities in health service of improving service delivery has been to use delivery public funds to contract with nonstate provid- · Governments will have limited capacity ers (NSPs), including nongovernmental orga- to manage contracts effectively nizations (NGOs), individual practitioners, or · They will be more expensive than govern- for-profit firms. ment provision of the same services, part- Proponents of contracting see primary ly reflecting greater transaction costs health care as relatively easy to buy because it · Even if successful, contracting will not be is measurable, the market is easily contested, sustainable. and it is widespread in high- and middle-in- A review of developing country experi- come countries. They also argue that contract- ence with contracting for health service ing has some potentially attractive features, delivery was undertaken to examine its effec- including the possibility of: tiveness, determine the extent to which the · Ensuring a greater focus on the achieve- posited difficulties actually occur in practice, ment of measurable results and make recommendations regarding future · Utilizing the private sector's greater efforts in contracting. flexibility and generally better morale to improve services Methodology · Overcoming "absorptive capacity" con- The focus of the review was instances in de- straints and improving the distribution veloping countries of governments contract- of trained health workers ing with NSPs to deliver primary health care · Increasing managerial autonomy and de- services including nutrition (but excluding centralizing decision making to managers hospital care or ancillary services such as on the ground FROM THE DEVELOPMENT ECONOMICS VICE PRESIDENCY AND POVERTY REDUCTION AND ECONOMIC MANAGEMENT NETWORK drug procurement and distribution). To be Guatemala. Improvements were also achieved included in the review, the example had to fairly rapidly, usually within two to three years. have been explicitly evaluated using mea- The most rigorously evaluated cases tended to sures of quality of care, or outputs such as demonstrate the largest impact. There were All the studies increase in the amount of services provided. four cases with controlled, before, and after that compared The evaluations also had to, at a minimum, evaluations that allowed the calculation of involve before and after or controlled designs. double differences (follow-up minus baseline government and The review excluded grants to NGOs in which in the experimental group, minus follow-up, contractor the latter determined the locations or type minus baseline in the control group). In these performance of services to be delivered. Also excluded four studies, the median double difference were contracts between different levels of of the various parameters studied ranged found that governments. Examples of contracting were from 2 to 26 percentage points (Figure 1). contractors were found through discussions with experts from The parameters varied between examples, a variety of institutions, previous reviews but most included coverage of services such more effective of contracting, a computerized search of as immunization, outpatient visits, and pre- the published literature using a variety of natal care. databases, and manual review of journals Six of the ten studies compared contrac- that often publish articles related to health tor performance to government provision of systems in developing countries. Structured the same services. All six studies found that interviews were conducted with people who the contractors were more effective than the had intimate knowledge of the examples that government, based on parameters related to met the inclusion criteria. both quality of care and coverage of services. The current weight of evidence indicates Results that contracting with NSPs will provide bet- Ten examples of contracting were found that ter results than government provision of the met the inclusion criteria and all ten con- same services. cluded that contracting with NSPs improved service delivery. Good results were achieved Specific Examples of Contracting in a variety of settings and for a variety of Cambodia: Many years of war and politi- different services ranging from nutrition cal upheaval left Cambodia with a limited services in Africa to primary health care in health infrastructure, particularly in rural Figure 1. Double Differences in Performance Parameters in Four areas. Health worker morale was poor and Examples with Controlled, Before, and After Evaluations management capacity at the district level was very weak. A 1998 Demographic and Health Survey found that nationwide only 39 percent of children were fully immunized. To address these serious issues, the Ministry of Health (MOH), with financing provided by the Asian Development Bank, contracted with NGOs in two different ways: (i) contracting out (CO), in which the contractors had complete re- sponsibility for service delivery, including hir- ing, firing, setting wages, and procuring drugs and supplies; and (ii) contracting in (CI), in which the contractors worked within the MOH system and could not hire or fire staff, although they could request staff transfers. Drugs and supplies were provided to the dis- trict through the normal MOH channels. In control districts, the management of services remained in the hands of the district health Note: CO = Contracting Out, CI = Contracting In, BINP= Bangladesh Inte- management team, which received technical grated Nutrition Project, UPHCP = Urban Primary Health Care Project. assistance and management training. PREMNOTE FEBRUARY 2006 Twelve districts with a combined popula- Figure 2. Change in Key Indicators in Cambodia, Follow on--Baseline in tion of 1.5 million were randomly assigned to Percentage Points the three different approaches, and baseline household and health facility surveys were car- ried out. Follow-on surveys were conducted about 2.5 years after implementation began. As shown in Figure 2, there were much larger improvements in immunization coverage, the use of antenatal care, and other indicators in the CO and CI districts than in the control districts, although they were quite similar at baseline. The poor appear to have benefited disproportionately from contracting, with concentration indices showing that services became more pro-poor in the contracted Note: FIC = percent of children 12­23 months of age who are fully immunized; Vit. districts and less pro-poor in the control A = proportion of children 6­59 months of age who received vitamin A supplements districts. in the last 6 months; ANC = coverage of antenatal care, one or more visit; HF Del Pakistan: There is a widespread feeling = proportions of infants delivered in a health facility; MBS = proportion of couples in Pakistan that first level facilities known who have children 12­23 months of age who are using a modern form of birth spac- as basic health units (BHUs) are providing ing/contraception; Use = proportion of people sick in the last month who used a only a limited amount of services to the rural government health facility; CC = Control/Comparison; CI = Contracting In; CO = population despite the investment of a large Contracting Out. amount of resources in their construction, staff, equipment, and supplies. In a poorly the large scale needed to make a difference performing district of Punjab, an NGO, the at country level appears to be unwarranted. Punjab Rural Support Program (PRSP), was Half of the examples studied involved popula- given a contract to manage all the BHUs tions of millions of beneficiaries, and in one and considerable autonomy to implement example, contracts now cover one-third of ru- changes in organization and management. ral Bangladesh, more than 30 million people. The NGO was given the same budget as had The apprehension that nonstate providers will previously been allocated for the 104 BHUs not want to work in remote or difficult areas in the district that served 3.3 million people. and are less capable of providing services to The NGO quickly introduced a number of in- the very poor also appears to be unwarranted. novations, including: (i) bringing in talented Given the resources and the explicit responsi- managers who were paid at market rates; (ii) bility, many contractors were willing and able The cost of increasing the salaries of doctors 150 percent to work in difficult areas that had previously provision of a and having them cover three different BHUs been underserved. However, only the evalu- instead of one; and (iii) improving the supply ation in Cambodia explicitly addressed the basic package of of drugs and renovating the BHUs. issue of whether contracting could improve primary health Information from the routine reporting equity, and it found contractors were able to system on the number of outpatient visits in significantly improve health services for the care in low the district was tracked over time. As can be most marginalized groups. Concerns about income settings seen in Figure 3, there was a dramatic increase governments' ability to manage contracts varied between (threefold) in the number of outpatient visits appear to be justified. Contract management to the BHUs after the government gave the was seen as a significant issue in at least three $3 AND $6 per NGO the authority and budget to run the of the examples studied; however, it did not capita per year system. prevent contractors from achieving signifi- cant improvements in health service delivery Insights from the Evidence even under those circumstances. There were instances where contracting The global experience with contracting pro- was more expensive than government provi- vides some insights on the concerns that have sion. However, in most of the cases, NSPs been raised about the approach. The fear that performed better even when public institu- contracts are unlikely to provide services on tions had the same amount of resources. PREMNOTE FEBRUARY 2006 Figure 3. Outpatient Visits per Month, before and after NGO Management that can achieve good results under different circumstances. Nonetheless, there are practi- cal considerations that can likely increase effectiveness 200,000 and efficiency, such as assuring managerial autonomy, mak- ing the contracts large scale, and focusing on performance 150,000 measurement. PRSP This note was written by Begins PR Benjamin Loevinsohn (Lead 100,000 SP begins Health Specialist, South Asia Human Development) and April Harding (Senior Econo- mist, Latin America and the 50,000 Caribbean Human Develop- ment) and is based on their pa- per, Buying Results: A Review 0 of Developing Country Experi- May- Aug- Nov- Feb- May- Aug- Nov- Feb- May- Aug- ence with Contracting. 02 02 02 03 03 03 03 04 04 04 The actual prices paid for provision of a basic Further Reading package of primary health care in the low Loevinsohn, B., and A. Harding. 2005. Buy- income settings varied between US$3 and ing Results: A Review of Developing US$6 per capita per year, which represents Country Experience with Contracting. a fairly small percentage of gross domestic The Lancet 366(9486): 676­681. It is product. Sustainability has also not been an also available as a Health, Nutrition, and issue. Of the seven examples with more than Population Discussion Paper at: http:// three year of elapsed experience, all seven siteresources.worldbank.org/HEALTH- Achieving the have been continued and expanded. NUTRITIONANDPOPULATION/Re- health-related sources/281627-1095698140167/Chap- 11LoevinsohnHarding.pdf. MDGs requires Recommendations Jones, G., Steketee, R.W., and Black, R.E., et Based on the success thus far, health services' improvements al. 2003. How many child deaths can we contracting should be significantly expanded prevent this year? Lancet 362: 65­71. in health service in developing countries as a way of helping World Bank. 2004. Making Services Work delivery for which achieve the MDGs and increasing account- for Poor People, World Development ability. While the evidence for contracting is contracting can Report, International Bank for Recon- reasonably good, future efforts should still struction and Development, World Bank, play a useful role include rigorous evaluations to obtain more Washington, DC. robust estimates of the effects under various Harding, A., and A. Preker, Eds. (2003). conditions and address unresolved issues, Private Participation in Health Services. such as the effectiveness of performance Washington, DC: World Bank (Chapter 3 bonuses and means for improving contract is on Contracting for Health Services). management. Based on global experience, it appears that contracting is a robust approach This note series is intended to summarize good practices and key policy findings on PREM-related topics. The views expressed in the notes are those of the authors and do not necessarily reflect those of the World Bank. PREMnotes are widely distributed to Bank staff and are also available on the PREM Web site (http://prem). If you are interested in writing a PREMnote, email your idea to Madjiguene Seck. For additional copies of this PREMnote please contact the PREM Advisory Service at x87736. This PREMnote was edited and laid out by Grammarians, Inc. Prepared for World Bank staff