POLICY AND PRACTICE
Externaliser, mais avec prudence : expérience acquise dans le domaine des prestations de santé dans certains pays de la Région Méditerranée orientale
Contratar, pero con prudencia: experiencia de externalización de servicios de salud en países de la Región del Mediterráneo Oriental
Sameen SiddiqiI,1; Tayyeb Imran MasudII; Belgacem SabriI
IEastern Mediterranean Regional Office, World Health Organization. Abdul Razzak Al Sanhouri Street, P.O.Box 7608, Nasr City, Cairo 11371, Egypt
IIDivision of Health System and Services Development, Johns Hopkins School of Public Health, Baltimore, MD, USA
The public sector in developing countries is increasingly contracting with the non-state sector to improve access, efficiency and quality of health services. We conducted a multicountry study to assess the range of health services contracted out, the process of contracting and its influencing factors in ten countries of the Eastern Mediterranean Region: Afghanistan, Bahrain, Egypt, Islamic Republic of Iran, Jordan, Lebanon, Morocco, Pakistan, the Syrian Arab Republic and Tunisia. Our results showed that Afghanistan, Egypt, Islamic Republic of Iran and Pakistan had experience with outsourcing of primary care services; Jordan, Lebanon and Tunisia extensively contracted out hospital and ambulatory care services; while Bahrain, Morocco and the Syrian Arab Republic outsourced mainly non-clinical services. The interest of the non-state sector in contracting was to secure a regular source of revenue and gain enhanced recognition and credibility. While most countries promoted contracting with the private sector, the legal and bureaucratic support in countries varied with the duration of experience with contracting. The inherent risks evident in the contracting process were reliance on donor funds, limited number of providers in rural areas, parties with vested interests gaining control over the contracting process, as well as poor monitoring and evaluation mechanisms. Contracting provides the opportunity to have greater control over private providers in countries with poor regulatory capacity, and if used judiciously can improve health system performance.
Le secteur public des pays en développement passe un nombre croissant de contrats avec le secteur non public afin d'améliorer l'accès, l'efficacité et la qualité des services de santé. Nous avons mené une étude multi-pays pour évaluer la gamme des services de santé externalisés, le processus de passation des contrats et les facteurs influant sur ce processus dans dix pays appartenant à la Région Méditerranée orientale : Afghanistan, Bahreïn, Égypte, République Islamique d'Iran, Jordanie, Liban, Maroc, Pakistan, République arabe syrienne et Tunisie. D'après les résultats de cette étude, l'Afghanistan, l'Égypte, la République Islamique d'Iran et le Pakistan ont fait l'expérience d'une externalisation des services de santé primaire; la Jordanie, le Liban et la Tunisie ont externalisé très largement leurs services de soins hospitaliers et ambulatoires; tandis que le Bahreïn, le Maroc et la République arabe syrienne n'avaient externalisé pour l'essentiel que des services non cliniques. L'objectif de la passation de contrats avec le secteur non public était de garantir une source régulière de revenus et d'obtenir davantage de reconnaissance et de crédibilité. Si la plupart des pays concernés favorisaient la passation de contrats avec le secteur privé, la présence du dispositif juridique et administratif nécessaire dépendait de la durée de l'expérience acquise en matière d'externalisation. Les risques inhérents à l'évidence au processus de passation de contrats étaient la fiabilité des bailleurs de fonds, le nombre limité des prestateurs en zone rurale, l'intervention de parties ayant des intérêts dans la mainmise sur le processus de passation de marchés, ainsi que l'insuffisance des mécanismes de surveillance et d'évaluation. L'externalisation offre la possibilité d'exercer un plus grand contrôle sur les prestateurs privés des pays dont le dispositif réglementaire est limité et, si elle est utilisée à bon escient, d'améliorer les performances du système de santé.
El sector público de los países en desarrollo está contratando cada vez más servicios con el sector no público para mejorar la accesibilidad, la eficiencia y la calidad de los servicios de salud. Realizamos un estudio multipaíses para evaluar la variedad de servicios de salud subcontratados, el proceso de contratación y los factores que influyen en ellos en diez países de la Región del Mediterráneo Oriental: Afganistán, Bahrein, Egipto, República Islámica del Irán, Jordania, Líbano, Marruecos, Pakistán, República Árabe Siria y Túnez. Nuestros resultados muestran que el Afganistán, Egipto, la República Islámica del Irán y el Pakistán tenían experiencia de contratación externa de servicios de atención primaria; Jordania, Líbano y Túnez subcontrataban gran parte de sus servicios de atención hospitalaria y ambulatoria; y Bahrein, Marruecos y la República Árabe Siria externalizaban sobre todo servicios no clínicos. El interés del sector no estatal por conseguir contratos se debía a su deseo de asegurarse una fuente regular de ingresos y lograr un mayor reconocimiento y credibilidad. Aunque la mayoría de los países promovían la contratación con el sector privado, el apoyo jurídico y burocrático en cada país dependía de la duración de la experiencia de contratación. Los riesgos inherentes al proceso de contratación eran la dependencia de los fondos de donantes, el número limitado de proveedores en las zonas rurales, la posibilidad de que partes con intereses creados llegaran a controlar el proceso de contratación, y unos mecanismos de seguimiento y evaluación deficientes. La contratación brinda la oportunidad de ejercer un mayor control sobre los proveedores privados en los países con escasa capacidad normativa, y si se usa de forma juiciosa puede mejorar el desempeño del sistema de salud.
Contracting is being increasingly used by the public sector in developing countries for the purchase of specified services from the private sector,1-5 and in some developed countries through internal contracts with autonomous public providers.6-9 Contracting is an important element of health sector reform programmes in many countries because it provides governments with a management and regulatory tool that creates incentives for improved performance and increased accountability.1,10 Recent reviews have suggested that contracting can influence access, equity, quality and efficiency of health services; promote public health goals; and create an environment conducive to public-private collaboration.11-13 The process is challenging and requires well designed contracts, transparent bidding, clear performance obligations and credible funding mechanisms. In addition, governments need to be able to monitor contracts and have the reputation of being a trustworthy partner.
We conducted a study on contracting out of publicly financed health services in ten countries of the Eastern Mediterranean Region (EMR) with the aim to (1) assess the rationale and capacity of the ministries of health (MOHs) and the private providers to enter into contractual arrangements; (2) review health interventions in which an outsourcing arrangement was implemented; and (3) identify factors that influence contracting.
We performed the study in 2004 in Afghanistan, Bahrain, Egypt, Islamic Republic of Iran, Jordan, Lebanon, Morocco, Pakistan, the Syrian Arab Republic and Tunisia. These countries accounted for two-thirds of the 500 million population of the EMR and represented its diversity in terms of variation in income, population, size of the private sector and experience with contractual arrangements.
We developed an open-ended checklist (see Annex 1) to guide country investigators in data collection. Although researchers were given some flexibility to adapt the checklist to their countries' needs, none of them modified it. This common checklist was the basis for ensuring comparability of data. Our study had two sections: (1) assessment of the overall status of contracting out of health services; and (2) review of interventions in which a contractual arrangement was the principal implementation strategy. We collected both quantitative and qualitative information by reviewing existing documents and studies and interviewing stakeholders including MOH staff and private providers. Our analysis, which was predominantly qualitative, was based on information provided by each country on different aspects of contracting as identified in the checklist. We asked the country investigators to review the studies for any major gaps in information. We monitored progress by establishing an electronic network with country investigators and occasionally through in-country visits. We shared the results in a regional consultative meeting with national policy-makers in April 2005.
For the purpose of this study we defined "contracting" as a purchasing mechanism used to acquire a specified service, of a defined quantity and quality, at an agreed-on price, from a specific provider, for a specified period.14 "Contracting out" was defined as the development and implementation of a documented agreement by which one party (purchaser) provides compensation to another party (provider) in exchange for a definite set of services for a specific target population.15 We have used the terms contracting out and outsourcing interchangeably in this paper. "Internal contract" was defined as contracting among different tiers within the public sector.14
Rationale and institutional capacity for outsourcing
Rationale and interest
Our study revealed that most countries undertook contracting to improve access, efficiency and quality of health services (Table 1). In Bahrain and Lebanon, national policy to engage the private sector influenced the public health sector to outsource health services. In the Islamic Republic of Iran and Morocco the policy of decentralization was the underlying reason for contracting out. Afghanistan contracted out health services to rapidly expand basic health services disrupted due to years of conflict. In Tunisia contracting out to private providers was to help decrease the cost of treatment incurred on patients sent to foreign countries for treatment. While Pakistan contracted out to access the population at risk for HIV infection through nongovernmental organizations (NGOs), Jordan wanted to improve the use of private hospitals and save capital investment on public facilities. Egypt contracted out to improve coverage and quality of services and increase access to advanced medical technology available in private hospitals.
We found that the interests of the non-state health sector in contractual arrangements were the assurance of a regular source of revenue, enhanced recognition and credibility, increased volume of work and utilization of spare capacity. In Afghanistan, contracting out was a way of continuing the essential services being provided by NGOs during the post-conflict period.
Institutional capacity for outsourcing
We have summarized the overall capacity of the public and the non-state sector to engage in contracting (Table 2). Our study showed that all countries had some experience with contracting out of clinical services with the exception of Bahrain and the Syrian Arab Republic, where only non-clinical services were contracted out.
All countries, except the Syrian Arab Republic, had a policy that promoted contracting; Afghanistan had the most explicit policy. Bureaucratic support varied depending on the duration of contracting experience. Jordan, Lebanon and Tunisia had extensive experience, while the experience in Afghanistan, Egypt and Pakistan was relatively recent. We found that the legal framework and the necessary rules and procedures for outsourcing of health services required updating. Most ministries had limited capacity for cost and price analysis and transaction cost estimations were usually not done.
Most countries had some type of competitive bidding process for the award of contracts. However, many MOHs did not have independent contracting units, such as Pakistan, or lacked the institutional capacity to award contracts, such as Lebanon. In the Islamic Republic of Iran, although bidding procedures were in place, often, good local reputation and recommendations by experts and colleagues were the criteria for provider selection.
We observed that while NGOs were actively involved in the delivery of primary care services, especially in Afghanistan and Pakistan, the quality of services they provided, their lack of managerial capacity and the absence of a system for their accreditation were some of the associated key problems. In Egypt, Jordan, Lebanon, Morocco and Tunisia private providers had access to advanced medical technologies that made them attractive to contract with for secondary and tertiary care services.
With the exception of Egypt and Afghanistan, we did not find performance indicators included in the design of contracts. Moreover, since management information systems were inadequate to monitor the performance of private providers, such as in Afghanistan, performance was measured through third-party evaluations.
We observed that payment methods for most contractual arrangements were either fee-for-service or block grants. Afghanistan had experience with payments based on capitation along with incentives for good performance. We observed that experience with capitation-based payment is accumulating in Egypt and the Islamic Republic of Iran, and that delayed payments were a problem in Jordan and Pakistan. We noticed risks with contracting out, such as differing interpretations of loosely worded contracts, limited number of providers in rural areas and parties with vested interests gaining control over the contracting process. Skeptics were also concerned about the lack of commitment among governments in low-income countries to contract once donor funds dry out.
We found that decentralization provided impetus for contracting out of health services in the Islamic Republic of Iran, Morocco and Pakistan so as to benefit from the synergy between decentralization and contracting that encourages institutions to improve management, optimize use of resources, be more accountable and improve overall performance.
Interventions with contractual arrangements
Primary care services
In Egypt, the Family Health Fund (FHF) was the main contracting agency to split financing from service provision, ensure competition among providers and act as a forerunner of the National Health Insurance Fund.16 The FHF is piloting the purchase of a package of primary care services for registered families through contracts with accredited private providers and NGOs, as well as internal contracts with the reformed public sector facilities. It has identified a set of 30 coverage, utilization and quality indicators for monitoring performance.
Our evaluation of contracting out of 103 primary health care (PHC) facilities in Rahim Yar Khan district in Pakistan showed that utilization, physical condition of facilities and patient satisfaction had improved and out-of-pocket expenditure had decreased. The quality of care, drug availability and accessibility to remote communities did not improve and there was little effect on the coverage of preventive health services.17 The Ministry of Health and Medical Education is piloting contracting out of primary care services in several provinces of the Islamic Republic of Iran. While it is difficult to give a verdict in the absence of a comprehensive evaluation, early evidence suggests that contracting out has helped improve access and quality, and decreased the cost of services.
We found that up to 80% of the health facilities in Afghanistan were operated by NGOs, which were directly contracted by the donors when peace was restored.18 Subsequently, the Ministry of Public Health espoused contracting out of a Basic Package of Health Services (BPHS) as the cornerstone of its National Health Policy. By early 2005, though the population coverage with the BPHS had extended to 16.5 million (70%), the actual extent of the service coverage remains unclear. The BPHS, which is funded by several donors, forms the core of service delivery with an estimated cost of US$ 4.5 per capita.19 A recent balanced score card assessment of health services demonstrated some improvement in health services.20 We hypothesize that the major risk inherent in contracting out in Afghanistan is the decrease in external donor funds and their long-term sustainability.
Jordan, Lebanon and Tunisia have had extended experience with contracting out of hospital services. In Lebanon, the public sector outsourced a wide range of services in over 100 hospitals to cover its uninsured population. Some limitations were fragmentation of the contracting process between different agencies, the limited leverage of public over private sector, the inability to contain escalating health care costs and lack of public sector capacity to monitor performance. We found that the Jordanian MOH had internal contracts with the university hospital and military services for the provision of emergency and referral care with evidence that contracting improved efficiency. Tunisia outsourced haemodialysis services to private institutions, which were obliged to accept a flat rate determined by the MOH and not to charge co-payments. In these three countries, we found that payments were being made on a-fee-for service basis and performance was not adequately monitored.
In Bahrain, Morocco and the Syrian Arab Republic, the MOHs contracted out services for the maintenance of medical equipment and hospital building, as well as support services such as cleaning, catering, gardening and security. Contracts were being awarded through competitive bidding, with monitoring mechanisms varying among countries and payments being made as block grants on a quarterly or annual basis. In Bahrain, a third party the Tender Board in the Ministry of Finance was ensuring transparency in the selection process.
What compels contracting out of health services in the Eastern Mediterranean Region?
We found that disillusionment of the general public with directly provided services; the concern of the public sector to improve access, efficiency and quality of services; optimization of hospital bed occupancy; and better targeting of vulnerable populations were the main reasons for contracting out in the EMR. The increasing trend in engaging the private sector in service delivery led many governments to promote health services outsourcing policies, which countries such as Bahrain are now actively following. These policies have influenced social sector ministries to outsource health services despite good public sector health services. While in Afghanistan the principal reason for contracting out has been the multilateral and bilateral donor agencies, contracting in Pakistan and the Islamic Republic of Iran was an indigenous initiative.
Contracting out of health services is receiving increasing attention among low- and middle-income countries but, while evidence relating to the benefits and risks of this approach is accumulating,11,12 it is far from conclusive. Careful consideration and a thorough analysis of the local context are essential before deciding on outsourcing versus direct provision.2 Contracting takes place more frequently in the EMR than presumed and there is a wide variation in the kinds of contractual arrangements and the range of services outsourced.
We hypothesize that contracting out of health services is an evolutionary process and that all countries, developed and developing, pass through a learning phase before contracting is recognized as an effective management and regulatory tool. We would like to mention two important cautionary notes: (1) contracting is complex and cannot be a solution for all problems of the health-care system; and (2) not everything can or should be contracted out. Jordan, Lebanon and Tunisia, despite decades of experience with contracting, need time before the capacities among purchasers and providers are adequately developed and procedures streamlined. In Afghanistan, Egypt, Islamic Republic of Iran and Pakistan, experience with outsourcing is relatively recent and the long-term sustainability of this approach has yet to be ascertained.
Our study had a few limitations: (1) coordinating collaborative research across ten countries given the diverse environments, though our checklist as a guide to undertaking studies, the e-network of country researchers and on-site monitoring allowed a fair comparison; (2) comparing the wide range of services contracted out from clinical to non-clinical, although these had been considered together in an earlier study,2 but the existence of several types of contracting mechanisms within each country made comparisons easier; and (3) the focus of the study being the process of contracting and not the outcome.
We believe that several issues regarding assessment of the contracting process in countries of the EMR need to be considered before contracting out becomes accepted as a viable alternative to direct provision. There needs to be a supportive public policy in favour of engaging with the non-state sector. The policies on private sector contracting have seen a substantial change over the past decade in the EMR. The legal and administrative framework for contracting out health services needs updating in many countries. Many MOHs lacked a dedicated unit for contracting. Afghanistan and Egypt were two examples where donor-financed projects have supported the establishment of contracting units. We believe that the limited capacity of the public sector to design, negotiate and award contracts; undertake a cost, price and volume analysis; optimize payment methods; and effectively monitor contract performance are among the remaining problems.
We hypothesize that unless such capacities are enhanced the value of contracted out services will always be questioned. Moreover, the capacity of private providers to successfully meet the requirements of the contractual arrangement in terms of an effective process and successful outcome should also be considered. The purchaser-provider relationship is an important determinant of the success of the contracting process. We need to consider whether providers find purchasers intimidating because they have greater negotiating power or is there a mutual trust and cooperation in their relationship?10 A perception that needs to be dispelled among public sector managers about NGOs is their apparent actions as profit-making entities.
Though our study could not document it, there may be substantial transaction costs involved in creating and maintaining the contracts, which is a reflection of the capacities of public sector institutions. Thus, it is equally important to compare the transaction costs of contracting with the explicit and hidden costs of directly managed public systems that incur large costs in monitoring staff and output quality. There may also be significant costs involved in bureaucratic and administrative mechanisms as well as political interference.21
Assessing whether providers are actually providing services on the ground relies on the following: (1) whether monitoring indicators will be included in the contract; (2) whether this information will be reported by the providers or collected independently; (3) does the public sector have the means for assessing the reliability and validity of the information? (4) how will the information be used? The monitoring and evaluation aspects of the contractual arrangement seem to be deficient in most EMR countries. A study from Costa Rica has shown that the data gathered do not provide the purchaser with information directly related to all of the contract objectives nor of the contract performance.22
In the EMR, PHC services are mostly contracted out to NGOs, which is similar to the experience of contracting primary health services in other WHO regions.3,23,24 Population coverage indicators, though easier to measure, provide little information on actual service coverage, as in Afghanistan. Thus, it is important that indicators, such as effectiveness and quality of treatment or efficient resource use, are considered while preparing contracts.22
Donor agencies have promoted contracting out of health services in many EMR countries, during or post-conflict, because health systems were disrupted and MOHs were weak, and NGOs took over as providers of health care as donor funds became available. There is always the risk of the donors influencing the process in the short term, raising the question of sustainability over the long term when donor funds begin to dry up. We believe that sustainable health improvement cannot be achieved unless physical infrastructure and human resources are rehabilitated by the public sector for the poor and vulnerable.
Lebanon, which experienced a civil war in the 1970s, provides lessons on the long-term consequences of contracting out. During and following the civil war, there was rapid expansion of the private sector at the cost of the public sector, the MOH had a limited role in service provision and contracted out most services, which resulted in high cost of health care. Though Lebanon spends more than US$ 500 per capita on health, its health system is arguably the least efficient in the region. While contracting with NGOs may be the only means to improve the system in post-conflict situations, the risk of missing the opportunity for long-term health system development projects the need for future research and debate.25
We conclude that contracting as a purchasing tool, when applied judiciously, could contribute to the improvement of health system performance. Contracting does not mean privatization of health services. While the state cannot divest itself from the responsibility of ensuring essential health functions, contracting out may provide an opportunity to obtain greater control over private providers in developing countries with poor regulatory capacity. Five supporting elements have been proposed for effective contractual arrangements: a few clearly defined deliverables; supportive stakeholders; trust between contractor and agency contracted to deliver services; independent source of monitoring information; and a legal system and political environment which convinces both sides that the contract will actually be enforced.24
While effective contracting requires MOHs to have the capacity to design, award, manage and monitor contracts, enhanced capacity of the private sector to implement contracts is equally important. We suggest that contracting be used primarily to promote public health objectives. At the same time, more research is required to evaluate the impact of contracting on health outcomes in the region.
Acknowledgements are due to Professor Anne Mills, London School of Hygiene and Tropical Medicine, for comments on the first draft of this paper. We also acknowledge all the country investigators Mustafa Mastoor (Afghanistan), Ahmed Omran (Bahrain), Hassan Salah (Egypt), Abolghasem Pourreza (Islamic Republic of Iran), Musa Ajluni (Jordan), Samer Jabbour (Lebanon), Alaoui Belghiti (Morocco), Amer Khan (Pakistan), Mahmoud Dashshash (the Syrian Arab Republic) and Noureddin Achour (Tunisia) whose efforts made this study possible.
Funding: We gratefully acknowledge Dr Hossein A Gezairy, Regional Director, WHO Regional Office for the Eastern Mediterranean, for providing financial support from the Regional Office for the Eastern Mediterranean Regional Director's Development Fund.
Competing interests: none declared.
1. Jack W. Contracting for health services: an evaluation of recent reforms in Nicaragua. Health Policy Plan 2003;18:195-204.
2. Mills A. To contract or not to contract? Issues for low and middle income countries. Health Policy Plan 1998;13:32-40.
3. Abramson WB. Partnerships between the public sector and non-governmental organizations: contracting for primary health care services a state of the practice paper. Latin American and Caribbean Health Sector Reform Initiative. Bethesda (MD): Partnerships for Health Reform Project, Abt Associates Inc;1999.
4. Loevinsohn B. Practical issues in contracting for primary health care delivery: lessons from two large projects in Bangladesh. 2002. Washington (DC): The World Bank. Available from: http://rru.worldbank.org/Documents/PapersLinks/746.doc.
5. Vining AR, Globerman S. Contracting-out health care services: a conceptual framework. Health Policy 1999;46:77-96.
6. Ashton T, Cumming J, McLean J. Contracting for health services in a public health system: the New Zealand experience. Health Policy 2004;69:21-31.
7. Williams G, Flynn R, Pickard S. Paradoxes of GP fundholding: contracting for community health services in the British National Health Service. Soc Sci Med 1997;45:1669-78.
8. Allen P. A socio-legal and economic analysis of contracting in the NHS internal market using a case study of contracting for district nursing. Soc Sci Med 2002;54:255-66.
9. Christiansen T. Organization and financing of the Danish health care system. Health Policy 2002;59:107-18.
10. World Health Organization, Regional Office for the Western Pacific. Contracting for health services, Lessons from New Zealand 2004. Available from: http://www.wpro.who.int/NR/rdonlyres/B7DB4D58-7E19-4884-BD63-6F90D2DC1C47/
11. Liu X, Hotchkiss DR, Bose S, Bitran R, Giedion U. Contracting for primary health services: evidence on its effects and a framework for evaluation. Bethesda (MD): The Partners for Health Reformplus Project, Abt Associates Inc.; 2004. Available from: http://www.phrplus.org/Pubs/Tech053_fin.pdf.
12. Loevinsohn B, Harding A. Buying results? Contracting for health service delivery in developing countries. Lancet 2005;366:676-81.
13. Perrot J. The role of contracting in improving health systems performance. Discussion paper No. 1; 2004. World Health Organization, Geneva. EIP/FER/DP.E.04.1.
14. Taylor R. Contracting for health services. In: Private participation in health services. Harding AL, Preker AS, editors. Washington (DC): The World Bank; 2003:157-218.
15. England R. Contracting and performance management in the health sector: some pointers on how to do it. London: UK Department for International Development (DFID) Health Systems Resource Centre (HSRC); 2000. Available from: http://www.dfidhealthrc.org/shared/publications/Toolkits/Contracting.PDF.
16. Ministry of Health. Egypt. Family Health Fund and contracting strategy: Health Finance and Insurance Group. Health Sector Reform Program; 2003.
17. The World Bank, South Asia Human Development Unit. An external evaluation: Chief Minister's initiative on primary health care. Rahim Yar Khan District Punjab. Discussion Draft; August 2005.
18. Ministry of Health, Transitional Islamic Government of Afghanistan. Afghanistan National Health Resources Assessment (ANHRA). Prepared for the Ministry of Health by MSH, HANDS and MSH Europe; 2002. Available from: http://www.msh.org/afghanistan/ANHRA_2002_LITE.pdf.
19. Transitional Islamic Government of Afghanistan, Ministry of Health. A basic package of health services in Afghanistan. Kabul; 2003.
20. Johns Hopkins Bloomberg School of Public Health; Indian Institute of Health Management Research; Ministry of Public Health (2005). Afghanistan Health Sector Balanced Scorecard, National and Provincial Results, Round 1 (2004). Kabul, Afghanistan: Ministry of Public Health, Policy and Planning Division, Monitoring & Evaluation Unit.
21. Mills A, Broomberg J. Experiences of contracting health services: an overview of the literature. London School of Hygiene and Tropical Medicine; 1998. HEFP working paper01/98. Available from: http://www.hefp.lshtm.ac.uk/publications/downloads/working_papers/01_98.pdf.
22. Abramson WB. Monitoring and evaluation of contracts for health service delivery in Costa Rica. Health Policy Plan 2001;16:404-11.
23. Mercer A, Khan MH, Daulatuzzaman M, Reid J. Effectiveness of an NGO primary health care in rural Bangladesh: evidence from the management information system. Health Policy Plan 2004;19:187-98.
24. Mackay B, Azariah S. Contracting NGOs: DFIDs five supporting elements. Futures Group; 2005. Available from: http://www.futuresgroup.com/Documents/Contracting_NGOs.pdf.
25. Palmer N, Strong L, Wali A, Sondorp E. Contracting out health services in fragile states. BMJ 2006;332:718-21.
(Submitted: 6 May 2006 - Final revised version received: 28 July 2006 - Accepted: 4 August 2006)