Introduction Many low-income countries have introduced user fees for publicly provided health services, often as part of structural adjustment programmes. User fees are usually only one element in a broader package of health sector reform measures (McPake et al. 1993; World Bank 1993; Uzochukwu et al. 2002). However, user fees have an obvious drawback: their potential negative effect on access to health care, especially for women (Nanda 2002). In many cases, partly because of poor implementation, utilization decreased significantly after user fees were raised, affecting the poor in particular (Creese 1991; McPake 1993; Gilson et al. 2001). However, in some cases where fee revenues were used for quality improvements, studies have shown increased utiliz- ation, in particular by lower income groups (Litvack and Bodart 1993; Levy-Bruhl et al. 1997; Audibert and Mathon- nat 2000). On the other hand, other costs – such as transport, time, food, informal charges and drugs – may constitute even higher barriers for the poor than the user fees themselves (Abel-Smith and Rawal 1992; Hjortsberg and Mwikisa 2002; Khe et al. 2002). The official policy in many countries is to exempt the poor from payment. In practice, however, exemption mechanisms are often ineffective and generally fail to protect the poor (Creese 1991; Gilson 1997). Their major failure is in the targeting. Good targeting consists in limiting inclusion errors and exclusion errors (Gilson et al. 1995; Willis and Leighton 1995). Most exemption systems in developing countries suffer from at least one of these two errors. Inclusion errors, which lead to leakage of resources to better off people, are due to the pressure of authorities and relatives, the absence of clear criteria or the low accountability of the persons granting the exemption. Two factors cause exclusion errors. First, where health staff find themselves in a market-oriented setting, a conflict of interest may arise between granting exemptions and raising income, especially when user fees serve to top up health workers’ incomes. Each exempted patient is a loss of revenue for facilities already under finan- cial stress. Second, health staff usually do not have the exper- tise or time to assess objectively the patient’s ability to pay (Huber 1993). This requires the skills of a welfare worker. Any institutional solution really aiming at assisting the poor in their access to health care should then have two basic features. First, there must be an earmarked budget. It must be in line with the poverty profile of the covered population. Second, the fund must be entrusted to a body the interests of which do not conflict with the mission to organize effective targeting of the poor. A purchasing body, or third-party payer, able to identify the poor and to pay on their behalf doi: 10.1093/heapol/czh003 Health Policy and Planning 19(1), HEALTH POLICY AND PLANNING; 19(1): 22–32 © Oxford University Press, 2004; all rights reserved. Access to health care for all? User fees plus a Health Equity Fund in Sotnikum, Cambodia WIM HARDEMAN, 1 WIM VAN DAMME, 1,2 MAURITS VAN PELT, 1,3 IR POR, 1 HENG KIMVAN 4 AND BRUNO MEESSEN 2 1 Médecins sans Frontières, Phnom Penh, Cambodia, 2 Department of Public Health, Institute of Tropical Medicine, Antwerp, Belgium, 3 Department for International Development (UK), Beijing, China and 4 Centre for Advanced Studies, Phnom Penh, Cambodia User fees in health services are a source of much debate because of their potential risk of negative effects on access to health care for the poor. A Health Equity Fund that identifies the poor and pays on their behalf may be an alternative to generally ineffective fee exemption policies. This paper presents the experience of such a Health Equity Fund, managed by a local non-governmental organization, in Sotnikum, Cambodia. It describes the results of the first 2 years of operations, investigates the constraints to equitable access to the district hospital and the effects of the Health Equity Fund on these constraints. The Health Equity Fund supported 16% of hospitalized patients. We found four major constraints to access: financial, geographical, informational and intra-household. The results of the study show that the Health Equity Fund effectively improves financial access for the poor, but that the poor continue to face many con- straints for timely access. The study also found that the Health Equity Fund as set up in Sotnikum was very cost-effective, with minimal leakage to non-poor. Health Equity Funds managed by a local non-governmental organization seem to constitute a promising channel for donors who want to invest in poverty reduction. However, further research and experimentation are recommended in different contexts and with different set-ups. Key words: user fees, equity, access, utilization, health services (Nyonator and Kutzin 1999), may be a good solution. However, there are few reports in the literature about such schemes. In Sotnikum Operational District, Siem Reap province, Cambodia, we negotiated with the Cambodian Ministry of Health (MoH) a health-financing scheme in 1999. Informal user fees became official and the proceeds used to improve the quality of health services to the population. No exemp- tions are granted. In September 2000, to improve access for the poor, a local non-governmental organization (NGO) was contracted to manage a Health Equity Fund that identifies the poor and pays user fees on their behalf. Health staff are thus entirely relieved from the responsibilities of identifying and financing poor patients. This paper describes (1) the results of 2 years’ functioning of the Health Equity Fund in Sotnikum, along with (2) the results of an investigation in August 2001 into the constraints to equitable access to the district hospital, and the effects of the Health Equity Fund on these constraints. Context Health sector reform in Cambodia The Cambodian society has suffered from decades of inter- national conflict, and especially from the 1975–79 Khmer Rouge government. To create a so-called egalitarian social- ist rural society, and to destroy the past, cities were evacu- ated, formal health care and education abandoned, and money abolished (Chanda 1986). Over 1 000 000 people died, and educated human capital was deliberately targeted for elimination. It is only after the Paris Peace Accords and 1993 elections, that Cambodia has gradually become more stable and peaceful. The country is beginning to recover from its loss of qualified human resources and the destruction of its infrastructure. Facing the challenge of rebuilding the health system while having to coordinate extensive foreign aid contributions (Lanjouw et al. 1999), the Cambodian Ministry of Health (MoH), heavily supported by the donor community, has been implementing a Health Coverage Plan since 1996. The Plan divides the country into 69 newly created Operational Districts, each covering between 100 000 and 200 000 people. Each operational district consists of a network of health centres that provide basic health care for 10 000 to 12 000 inhabitants, one district hospital and a district office. In 1997, the National Charter on Health Financing officially introduced user fees. By 2001, thanks to the funding of international banks and donors, most facilities had been constructed or renovated, and supplied with medical equipment. Moreover, the system enjoyed a regular supply of essential drugs through a Central Medical Store. However, structural constraints prevent the Plan’s realiza- tion in terms of accessibility and quality of care, and public health facilities remain under-utilized. Government expendi- ture on health is growing but remains very low, at around US$3 per capita per year in 2001. A fundamental constraint is the extremely low salary of civil servants in Cambodia (US$10–15 per month in 2001). Consequently, as elsewhere, government health staff are forced into coping strategies, such as charging informal fees and diverting drugs, equip- ment and patients to their private practices (Roenen et al. 1997; Killingsworth et al. 1999; McPake et al. 1999; Bloom et al. 2000; Van Lerberghe et al. 2002). Households are left to private providers who aim at maximizing profits by selling whatever people are willing to pay for. Because of the unregulated health care market, household ‘out-of-pocket’ expenditure on health is very high and inequitable, estimated at US$20–30 per capita per year. As in other countries (Wagstaff 2002), in Cambodia catastrophic health expendi- ture is identified as a major cause of indebtedness and desti- tution among the rural poor (Kassie 2000). In this context, many multi-actor initiatives have tried to improve the quality and accessibility of public health care in rural areas. Several of them share common features: (1) the pursuit of a higher performance of public facilities; (2) the injection of external funding for improving staff income; and (3) the utilization of performance contracts to establish a higher accountability of the staff (Van Damme et al. 2001; Meessen et al. 2002; Ministry of Health et al. 2002; Soeters and Griffiths 2003). Sotnikum Operational District and its ‘New Deal’ Sotnikum is a poor rural area, with 220 000 inhabitants, at 30 km from Siem Reap town, in the province where the historic temples of Angkor are located. Sotnikum Opera- tional District is divided into 17 health areas, each of them having one health centre. The district hospital is in the small town of Damdek. It provides the full complementary package of activities foreseen by the national policy: internal medicine, paediatrics, obstetrics-gynaecology and surgery. An operating theatre has functioned since mid-2000. In Sotnikum, the Ministry of Health, Médecins sans Fron- tières (MSF) and UNICEF agreed in 1999 on a common approach at district level. The entry point of the experiment is a ‘New Deal’ for the government health staff, who receive a better income (US$80–100 per month on average) in exchange for respect of working hours and abiding by the new internal regulations, strictly forbidding any informal payments or prescriptions for private pharmacies. Apart from a direct positive effect on the quality of services to the population, it was assumed that in the new working environ- ment of ‘fair’ income to staff, investments in quality of care, such as training and supervision, would yield better results. Contracts were established between all actors involved, and lump sum user fees agreed upon to reduce patient’s uncer- tainty about the cost of treatment. The approach immedi- ately resulted in better staff motivation and higher user rates. In 2000, average monthly admissions to the hospital increased by 50% to 216, further increasing to 239 in 2001, and to over 300 in 2002. This ‘New Deal’ initiative has been described in detail elsewhere (Van Damme et al. 2001; Meessen et al. 2002). Health Equity Fund The ‘New Deal’ meant an increase in the official user fees, especially at the hospital level. The risk was that the poor Health equity fund and hospital access 23 would still lack access to the improved services. To allow the achievement of the two conflicting objectives by the hospital – financial viability and equity in access – MSF and UNICEF decided to introduce a demand-side initiative: the Health Equity Fund, which operates as a ‘third party payer’ for patients who cannot pay. The Health Equity Fund is managed by a local NGO, and operates independently of the hospital. MSF and UNICEF preferred a Cambodian NGO for reasons of managerial and socio-political sustainability and capacity building in the private social sector. The contracted social NGO, Cambodia Family Development Services (CFDS), established in 1993, had already extensive experience in social projects and community development programmes in the province before it started the Sotnikum Health Equity Fund. Thanks to this experience, CFDS had already well-established and field- tested guidelines for identification of and support to poor patients, using a questionnaire to score self-reported socio- economic status. In September 2000, CFDS started with a single staff member based in an office in the hospital compound. In October 2001, a second staff member was hired to improve presence in the hospital, follow-up of supported patients and information sharing at community level. The CFDS staff identify poor patients at two stages. First, the hospital admission staff refer patients arriving without enough money for the admission fee. Secondly, CFDS also actively tries to identify poor patients in the hospital wards, who although they paid the admission fee, seem to lack food or basic items such as food utensils, mosquito nets or clothing. This second round of identification is necessary because most people borrow money or sell productive assets before they go to the hospital, and thus may incur expenditure beyond their ability to repay. CFDS staff interview all patients so identified to determine their ability to pay. Important indicators are food security, ownership of land and productive assets, housing, occu- pation, as well as household size and structure. Physical appearance, including clothing, often also gives an indication of socio-economic status. The ‘target group’ of the Health Equity Fund consists of the extremely poor, as well as the poor who risk falling into extreme poverty. No fixed criteria for eligibility are used, as poverty has many dimensions that are difficult to measure, and flexibility is needed to assist according to people’s needs. The level of support is determined on a case-by-case basis, from partial payment of the admission fee to full coverage of the total cost of hospitalization, including transport, food and basic items. CFDS also carries out follow-up visits in the villages to verify the adequacy of the targeting mechanism and to identify further needs. The contracting parties, MSF and UNICEF, have been moni- toring the activities of the Health Equity Fund through quan- titative indicators and to assess the adequacy of the targeting mechanism. They have not got involved in operational issues of the local NGO. Methods We compiled data on the Health Equity Fund for the first 25 months of functioning, 1 September 2000 to 30 September 2002, from the monthly reports by the NGO implementing the Fund, from personal observations of the authors, and from in-depth interviews by the authors with CFDS staff and community leaders. In August 2001, two of the authors (WH and HK) carried out a survey in Sotnikum. The aims were (1) to analyze the constraints poor people face when in need of hospital care; (2) to assess to what extent the Health Equity Fund helped to overcome these constraints. They conducted in-depth interviews in 26 villages in five communes between 5 and 25 km from the district referral hospital. The 68 patients assessed (14 children, 40 adults and 14 elderly; 38 women) consisted of three categories: (1) 34 patients, hospitalized in the period June-July 2001, not supported by the Health Equity Fund; (2) 17 patients, hospitalized in the period June-July 2001, supported by the Health Equity Fund; and (3) 17 poor households with a person with recent serious illness, but not hospitalized and thus not supported by the Health Equity Fund. We retrieved the cases in categories (1) and (2) from the hospital administration. Category (2) consisted of all 17 patients hospitalized in the period June-July 2001 from four communes who received support from the Health Equity Fund. Category (1) consisted of 34 patients randomly selected from the same four communes, and hospitalized during the same period, but without support from the Health Equity Fund. Seven patients were difficult to trace, because of the rice-planting season, and were substituted with cases from a fifth adjacent commune or with patients hospitalized in the period April-May 2001. We found poor households with a person with recent illnesses (3) in each of the four communes. Interviewed households and local authorities could often tell where there were poor people with serious illness who stayed at home, and small thatch houses were randomly visited. The interviews with categories (1) and (2) assessed expendi- ture on hospitalization and sources of financing used, as well as the targeting ability of the Health Equity Fund. The inter- views with category (3) served to gain a better understand- ing of the constraints facing the poor in accessing hospital care. Assets, occupation and food security determined the socio- economic rank of each household. Four categories were distinguished, following the widely used classification in Cambodia (Panhavichetr 1998): rich, medium, poor and extremely poor (Table 1). No ‘rich’ households were identified in this survey. Of the category (1), hospitalized patients without support from the Health Equity Fund, it appeared that five cases had 24 Wim Hardeman et al. been outpatients only. They were excluded from the analysis of hospital expenditures. We obtained additional qualitative data through interviews with local authorities and health staff, and focus group discussions at village level. Utilization figures of the hospital were obtained directly from the hospital administration. Results Utilization of the Health Equity Fund, and its targeting ability Between 1 September 2000 and 30 September 2002 the Health Equity Fund supported 1437 patients. Their numbers per month grew steadily from some 20 to over 100 (Figure 1). The number of non-supported patients increased less. During the third quarter of 2002, the Health Equity Fund supported around 30% of all patients admitted to Sotnikum hospital. Figure 2 shows the targeting ability of the Health Equity Fund. Among hospitalized patients, only one non-poor patient received financial support, whereas the Health Equity Fund supported nearly all of the extremely poor. The extremely poor, when they arrive at the hospital, have a better chance of being selected by the Health Equity Fund than the poor who risk falling into extreme poverty (which is more difficult to assess). However, the extremely poor also face most difficulties in reaching the hospital. Therefore, improved financial access for poor patients arriving at the hospital does not yet guarantee equitable access to hospital care. Moreover, a Health Equity Fund does not directly address many non-financial constraints. Also, accessibility of health facilities depends to a large extent on measures on the supply side, such as presence of health staff, transparent user Health equity fund and hospital access 25 Table 1. Characteristics of socio-economic categories, adapted from (Panhavichetr 1998) Estimated Estimated proportion of proportion of Cambodian population in population Sotnikum Rich Good house (stone), >3 ha land, large number of animals, motorbike, TV, lending 10% <5% (‘mean’) out rice/cash, large savings/surplus assets. Medium Good house (wood), >1.5 ha land, >2 cows, sometimes motorbike, TV, no food 30% 20% (‘kuesam’) shortages, savings, borrow occasionally. Poor Wooden house, <1.5 ha land, 2 cows or less, bicycle, food shortage less than 4 40% 50% (‘kraa’) months, small savings 25 kmHospital admissions per capita per year so that patients receive treatment once they have paid the admission fee. This significantly reduces uncertainty and improves financial access. Nonetheless, even transparent, relatively low user fees (US$7–10 per admission) still appear to be a major obstacle for the poor. Moreover, as discussed above, the financial cost of hospitalization is not limited to user fees alone. According to the findings of this study, people actually spent more on transport and food than on fees (Figure 3). Figure 5 shows the average expenditure on hospitalization by socio-economic category, by all hospitalized cases, distin- guishing out-of-pocket expenditure and support from the Health Equity Fund. It clearly reflects the nature of the targeting mechanism: the poorest patients are most likely identified as unable to pay and receive the highest level of support. Indeed, 87% of the extremely poor in the sample were identified and supported by the Health Equity Fund for 50% of their total expenditures. Within this group, four patients in a state of chronic extreme poverty (dependent on charity in their daily lives) relied almost completely on the Health Equity Fund for their total expenditures. In most cases, the Health Equity Fund was crucial in provid- ing financial access to the district hospital, as otherwise patients would not have been able to pay the admission fee. It is, however, unlikely that those who borrowed or sold assets, risking falling into extreme poverty, were prevented from doing so by the assistance of the Health Equity Fund. Often, high expenditure had already been incurred before hospitalization, mainly in the private sector. Moreover, the Health Equity Fund did not address the constraint of ‘oppor- tunity cost’ of hospitalization, which remains a major obstacle for the poor, in particular during rice planting and harvesting seasons and for people living on a day-to-day basis. Effects on geographical constraints Above, we identified distance to the hospital as a major constraint to access hospital care. Figure 6 shows the proportion of patients supported by the Health Equity Fund, by distance to the hospital. Relatively few patients (less than 6%) from within 5 km of the hospital were supported, and patients coming from medium or long distance had a much 28 Wim Hardeman et al. $0 $5 $10 $15 $20 $25 $30 $35 $40 $45 Medium Poor Extremely poor Health Equity Fund Out-of-pocket Figure 5. Average expenditure on hospitalization by socio-economic category: out-of-pocket expenditure and support from Health Equity Fund Figure 6. Proportion of patients supported by Health Equity Fund, according to distance between commune of origin and hospital (January–July 2001) 0% 2% 4% 6% 8% 10% 12% <5 km 5–25 km >25 km bigger chance of benefiting from the Health Equity Fund. Apparently, the Health Equity Fund reduces inequities in health expenditure in relation to distance. However, by itself it does not correct inequities in utilization resulting from geographical constraints to access. A well-functioning referral system is of major importance. In Sotnikum, the health-financing scheme allows for incentives to government staff at health centre level to appropriately refer patients to the hospital. An ambulance service is offered according to need, and those unable to pay for the ambulance get support from the Health Equity Fund. However, this study did not assess how these mechanisms affected geographical constraints. Effects on informational constraints The Health Equity Fund did not actively promote its activi- ties at community level or in the hospital. Consequently, among the people who had not been to the hospital, the Health Equity Fund was almost completely unknown. Also, of the hospitalized cases that had not been supported by the Health Equity Fund, after their return only 35% were aware of the presence of a Health Equity Fund at the hospital. Resulting inconsistent accounts may continue to deter the poor from going to the hospital. Of the supported patients, only 12% were aware of the Health Equity Fund beforehand. Therefore, in most cases the Fund had not reduced uncertainty about access. Further- more, even the poor who are aware of the Health Equity Fund remain uncertain about whether they will receive support, because identification of poor patients is centralized in the hospital and criteria for eligibility are not made public. On the other hand, information that is shared by all patients, such as improved quality of care or the absence of informal charges, is likely to spread much faster and to enhance utiliz- ation. Effects on intra-household constraints The Health Equity Fund has no specific provision for house- holds with difficult access to the hospital because of age, gender, family size and structure. These households are more likely to be vulnerable and poor (e.g. headed by a female, many dependants). When a poor patient is hospitalized, the Health Equity Fund assists in the expenses in the hospital, but not at home, where family members may be left without someone to take care of them, and possibly short of food. Their well-being is certainly an important consideration in the decision whether or not to seek hospital care. Of the patients supported by the Health Equity Fund, 55% were female, the same figure as their share in total utilization of the hospital. Apparently, the Health Equity Fund did not affect the proportion of women hospitalized. Cost of a Health Equity Fund The total cost of the Health Equity Fund over the study period was US$27 100 to support 1437 patients; or US$1084 per month. Of this amount, US$16 260 was direct financial assistance (of which around 74% was hospital fees, 20% transport and 6% food and basic items). Per beneficiary, the average financial assistance was US$11.32 and total cost US$18.86. The cost of the NGO managing the Health Equity Fund thus represented US$10 840, or 40% of the total cost. This was mainly for staff salary and staff transport. In 2001, the total running cost of the district hospital (staff, drugs, fuel, electricity, patient food, etc.) was US$152 000, or US$53 per hospital admission, of which 62% was funded by the state, 21% through users’ fees and 17% by external parties (MSF and UNICEF). The total cost of the Health Equity Fund for one year was around US$13 000, or 8.6% of hospital cost. The total cost of the public health system in Sotnikum – the hospital, 17 health centres and the district office – was just less than US$2 per capita per year, of which US$0.69 was for the hospital. The cost of the Health Equity Fund was US$0.06 per capita per year. Discussion In Sotnikum, the introduction of a Health Equity Fund managed by a local NGO appeared to effectively improve access to hospital care for the poor. During the first year – including the period covered by the survey – the Health Equity Fund may have mainly reduced the cost of care for people who had already chosen to access care. However, during the second year of the Fund, the steep increase in utilization indicates that a considerable number of the ‘new’ patients were from poor households and would not have sought care at the hospital without financial support. During its initial 2 years of functioning, there was a steep increase in poor people accessing the hospital, while the number of patients who paid their own fees did not decrease. Most patients who managed to make it to the hospital, but could not pay the hospital fees, were identified and finan- cially supported according to need, resulting in a strongly progressive transfer of resources. There was a minimal ‘leakage’ to non-poor patients. The patients with an absolute lack of ability to pay received nearly 100% coverage of their total financial cost of hospitalization (including transport, food and basic items). Indeed, even a well-functioning exemption mechanism would fall short of giving all the support needed by the poorest patients, for whom user fees are only one of the many financial barriers to access hospital care. It is, however, important to note that in the Cambodian context by far the most household health expenditure takes place outside the public sector (National Institute of Statis- tics and ORC Macro 2001), often spent on poor quality treat- ments by informal private practitioners. Therefore, in terms of poverty prevention, the greatest potential of the Health Equity Fund seems not to lie in financing expenditure in the public sector, but in preventing unnecessary expenditure in the private sector, by encouraging the use of adequate public health care. Indeed, surveys in Cambodia have shown that, even in the absence of fee exemptions for the poor, the exist- ence of a credible public health service considerably reduces Health equity fund and hospital access 29 out-of-pocket health expenditure by the population in general, and the poor in particular (Keller and Schwartz 2001; Van Leemput and Van Damme 2002). Over the study period in 2000–02, of all the patients who arrived at Sotnikum district hospital, 16% received support from the Health Equity Fund, whereas an estimated 25% of the population is estimated to be extremely poor (Table 1), and another 25% at risk of falling into extreme poverty in case of serious illness. It seems likely that many potential beneficiaries did not reach the hospital. Remaining obstacles continue to deter many of the poor from seeking hospital care. Even though a Health Equity Fund cannot address all these constraints, it seems to have the potential to attract more of the poor, by improving on its ‘marketing’ and infor- mation sharing. Moreover, by decentralizing its mechanism of identification and support (e.g. to the health centre or commune level), it could reduce uncertainty about eligibility and better address the constraints facing the poor at home (e.g. transport). In terms of costs and benefits, through the payment of US$7–10 on user fees, an admission with an average cost of US$53 was made available to the poor, of which the govern- ment funded more than US$30. This multiplier effect made the total cost per supported patient, including the adminis- trative cost, much less than the value of hospital care obtained for the poor. A yearly investment of US$13 000, representing 8.6% of the total hospital budget, gave access to hospital care to 19% of all patients, who would otherwise not have had access or suffered serious financial loss in order to pay. The combination of low cost, well-targeted subsidies and multiple benefits seems to constitute a promising channel for institutional donors who want to invest in poverty reduc- tion. In 2000–02, the Health Equity Fund in Sotnikum had been operating as a pilot scheme, which was unknown to most of the population, run by a motivated local NGO and closely monitored by its contracting parties. In the longer term and on a larger scale, problems may arise, such as leakage to non- poor who may adapt their self-reported economic status. In addition, local NGOs may enter the scene to get a share in available budgets rather than to help the poor through a properly run Health Equity Fund. To avoid such perverse effects, Health Equity Funds need careful regulation, independent monitoring and evaluation, and accountability to the community they serve, within a broader pro-poor policy. Another effect that needs to be avoided is over- consumption of hospital care by the poor; although this seems more unlikely given the numerous constraints they face to access such care. Finally, in a fee-for-service scheme, provider-induced demand may become a concern as well, which is less probable in a scheme of fixed (‘lump sum’) user fees, as is the case in Sotnikum. In the Cambodian context, the performance of the Health Equity Fund in terms of improving access for the poor to hospital care is undoubtedly much better than the results of waivers for poor patients (Table 2). It is important to acknowledge the conditions under which the Health Equity Fund in Sotnikum is able to operate effec- tively. First, the fund is only a complement to a relatively well-functioning health service, in which health staff are present, drugs available and informal charges absent. Second, the context of substantial socio-economic differenti- ation in rural Cambodia makes it possible to charge (low) user fees to the majority part of the population, while target- ing support to those unable to pay. Third, the Health Equity Fund is managed by a local social NGO which has years of experience in the region, and demonstrates a good ability to target and motivation to serve the genuinely poor. The contracting of a local NGO to manage a Health Equity Fund seems advantageous in many respects. Whereas hospital or other government staff may have conflicting inter- ests, and often lack incentives to support the poor, an NGO contracted for this purpose does not face these constraints. NGOs may also be more easily replaceable in case of under- performance. Moreover, operating as an independent social actor in the health system, an NGO has the potential to represent the poor for whom it purchases health care and to reinforce their patient rights towards government staff. Besides improving financial access, through its social and psychological support, the patient’s dignity and confidence can be restored as well. However, one should not overesti- mate the potential role of NGOs, as they are not necessarily accountable to the population they serve. 30 Wim Hardeman et al. Table 2. Fee waiver vs. Health Equity Fund to improve access for poor patients in Cambodia Fee waiver for poor patients Health Equity Fund for poor patients Benefit for patients No fee. No fee and support for transport and other costs. Perception by health staff Negative. The hospital has to forego Positive. Increased revenue when caring for revenue from poor patients. Staff try to poor patients. avoid having poor patients. Targeting Poor, because health staff are not qualified Good, as social workers of NGO are to identify poor people, and have no specialized in identifying the poor and are interest in doing so. evaluated on their performance to do so. Cost Theoretically no cost, and in particular no Reasonably low: 8.6% of hospital costs. administrative cost. Overall results Not effective. Effective at reasonable cost. Undoubtedly, broader experience and further research into Health Equity Funds is needed, in different socio-economic contexts and in different health systems. Other set-ups can be imagined as well, for instance, where the local NGO sector is not well developed or where other alternatives seem more viable. The respective merits and costs of pre-identification of the poor (such as done in poor-card schemes) or identifi- cation at hospital entrance (as done in Sotnikum) could be compared. Comparative qualitative and quantitative studies would allow for a better assessment of the potential role of Health Equity Funds in national health and social policy (i.e. as exemption mechanism, social safety net), and their cost-effectiveness in terms of poverty reduction and preven- tion. 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. Audibert M, Mathonnat J. 2000. Cost recovery in Mauritania: initial lessons. Health Policy and Planning 15: 66–75. Bloom G, Han L, Li X. 2000. How health workers earn a living in China. IDS Working Paper 108. Brighton, UK: Institute of Development Studies. Chanda N. 1986. Brother enemy. The war after the war. A history of Indochina since the fall of Saigon. New York: MacMillan Publishing Company. Collins W. 2000. Medical practitioners and traditional healers: a study of health seeking behavior in Kampong Chhnang, Cambodia. Phnom Penh: Center for Advanced Studies. Creese AL. 1991. User charges for health care: a review of recent experience. Health Policy and Planning 6: 309–19. Criel B, Van der Stuyft P, Van Lerberghe W. 1999. The Bwamanda hospital insurance scheme: effective for whom? A study of its impact on hospital utilization patterns. Social Science and Medicine 48: 897–911. Gilson L. 1997. The lessons of user fee experience in Africa. Health Policy and Planning 12: 273–85. Gilson L, Kalyalya D, Kuchler F et al. 2001. Strategies for promot- ing equity: experience with community financing in three African countries. Health Policy 58: 37–67. Gilson L, Russell S, Buse K. 1995. The political economy of user fees with targeting: developing equitable health financing policy. Journal of International Development 7: 369–401. Hjortsberg CA, Mwikisa CN. 2002. Cost of access to health services in Zambia. Health Policy and Planning 17: 71–77. Huber JH. 1993. Ensuring access to health care with the introduc- tion of user fees: a Kenyan example. Social Science and Medicine 36: 485–94. Kassie A. 2000. Credit and landlessness: impact of credit access on landlessness in Cheung Prey and Battambang districts, Cambodia. Phnom Penh: Center for Advances Studies. Keller S, Schwartz JB. 2001. Final evaluation report: Contracting for health services pilot project. Phnom Penh: Asian Development Bank. Khe ND, Toan NV, Xuan LT et al. 2002. Primary health concept revisited: where do people seek health care in a rural area of Vietnam? Health Policy 61: 95–109. Killingsworth JR, Hossain N, Hedrick-Wong Y et al. 1999. Unofficial fees in Bangladesh: price, equity and institutional issues. Health Policy and Planning 14: 152–63. Lanjouw S, Macrae J, Zwi AB. 1999. Rehabilitating health services in Cambodia: the challenge of coordination in chronic political emergencies. Health Policy and Planning 14: 229–42. Levy-Bruhl D, Soucat A, Osseni R et al. 1997. The Bamako Initia- tive in Benin and Guinea: improving the effectiveness of primary health care. International Journal of Health Planning and Management 12 (Suppl 1): S49-S79. Litvack JI, Bodart C. 1993. User fees plus quality equals improved access to health care: results of a field experiment in Cameroon. Social Science and Medicine 37: 369–83. McPake B. 1993. User charges for health services in developing countries: a review of the economic literature. Social Science and Medicine 36: 1397–405. McPake B, Asiimwe D, Mwesigye F et al. 1999. Informal economic activities of public health workers in Uganda: implications for quality and accessibility of care. Social Science and Medicine 49: 849–65. McPake B, Hanson K, Mills A. 1993. Community financing of health care in Africa: an evaluation of the Bamako Initiative. Social Science and Medicine 36: 1383–95. Meessen B, Van Damme W, Por I, Van Leemput L, Hardeman W. 2002. The New Deal in Cambodia: the second year. Phnom Penh: MSF Cambodia. Ministry of Health/Swiss Red Cross/World Health Organization. 2002. Takéo provincial referral hospital: pioneering a health financing scheme. Phnom Penh: Swiss Red Cross. Nanda P. 2002. Gender dimensions of user fees: Implications for women’s utilization of health care. Reproductive Health Matters 10: 127–34. National Institute of Statistics C and ORC Macro. 2001. Cambodia Demographic and Health Survey 2000. Phnom Penh: Cambodia, and Calverton, MD: National Institute of Statistics and ORC Macro. Nyonator F, Kutzin J. 1999. Health for some? The effects of user fees in the Volta Region of Ghana. Health Policy and Planning 14: 329–41. Panhavichetr P. 1998. Gender relations in household food security: a case study of Kok Dong village, Cambodia. The Hague: Insti- tute of Social Studies. Roenen C, Ferrinho P, Van Dormael M, Conceicao MC, Van Lerberghe W. 1997. How African doctors make ends meet: an exploration. Tropical Medicine and International Health 2: 127–35. Soeters R, Griffiths F. 2003. Improving government health services through contract management: a case from Cambodia. Health Policy and Planning 18: 74–83. Uzochukwu BS, Onwujekwe OE, Akpala CO. 2002. Effect of the Bamako-Initiative drug revolving fund on availability and rational use of essential drugs in primary health care facili- ties in south-east Nigeria. Health Policy and Planning 17: 378–83. Van Damme W, De Brouwere V, Boelaert M, Van Lerberghe W. 1998. Effects of a refugee-assistance programme on host popu- lation in Guinea as measured by obstetric interventions. The Lancet 351: 1609–13. Van Damme W, Meessen B, von Schreeb J et al. 2001. Sotnikum New Deal, the first year. Phnom Penh: MSF Cambodia. Van Leemput L, Van Damme W. 2002. Dengue epidemic in Thmar Pouck, April-December 2001. Disease and its cost in rural Cambodia. Phnom Penh: MSF Cambodia. Van Lerberghe W, Conceicao C, Van Damme W, Ferrinho P. 2002. When staff is underpaid: dealing with the individual coping strategies of health personnel. Bulletin of the World Health Organization 80: 581–4. Van Lerberghe W, Pangu KA. 1988. Comprehensive can be effec- tive: the influence of coverage with a health centre network on the hospitalisation patterns in the rural area of Kasongo, Zaïre. Social Science and Medicine 26: 949–55. Wagstaff A. 2002. Poverty and health sector inequalities. Bulletin of the World Health Organization 80: 97–105. Willis CY, Leighton C. 1995. Protecting the poor under cost recovery: the role of means testing. Health Policy and Planning 10: 241–56. World Bank. 1993. World development report 1993: Investing in health. New York: Oxford University Press. Health equity fund and hospital access 31 Acknowledgements We thank Cambodian Family Development Services, the Ministry of Health, Médecins sans Frontières and UNICEF field staff in Sotnikum and Siem Reap for their collaboration and support, and Professor R Eeckels for useful comments on previous drafts. Biographies Wim Hardeman, BA (Economics), has worked in Nicaragua as volunteer financial and administrative advisor to micro-enterprises, and in Cambodia as financial co-ordinator with Médecins sans Fron- tières. In 2001, he obtained an MA in Development Studies from the Institute of Social Studies (The Hague), where he developed a strong interest in the relation between health care and poverty reduction. In 2002, he returned to Cambodia, to work with Médecins sans Frontières as project co-ordinator in Sotnikum. (E-mail: wim_hardeman@hotmail.com) Wim van Damme, MD, MPH, Ph.D., has worked with Médecins sans Frontières in Peru, Sudan and Guinea. In each of these coun- tries, he set to work in primary health care programmes, but because of political problems, the programmes shifted towards emergency medical assistance. He wrote a PhD thesis: Medical Assistance to Self-settled Refugees in Guinea, 1990–96. He teaches ‘Humani- tarian Assistance and Development’ at the University of Antwerp, and Public Health at the Institute of Tropical Medicine, Antwerp. Until end 2002, he worked as medical co-ordinator for Médecins sans Frontières in Cambodia, piloting a ‘New Deal’ way to improve public health services. Since early 2003, he again joined the Institute of Tropical Medicine, as professor of Public Health. (E-mail: wvdamme@itg.be) Maurits van Pelt, MSc, LL.M., has worked since 1987 with Médecins sans Frontières in several conflict and post-conflict areas. His longest stay was in Cambodia, from 1989 until 2000. After obtaining a masters degree in Health Policy, Planning and Financing in London, he went back to the field with Médecins sans Frontières, as interim-Head of Mission for Afghanistan. In April 2002, he joined the Department for International Development (UK) in Beijing to work in the Chinese Ministry of Health’s Foreign Loan Office as an adviser on rural health programmes. (E-mail: maurits@vip.163.com) Ir Por, MD, worked as chief of a Quality Improvement Programme at the Siem Reap Provincial Health Department, Cambodia. Since late 1997, he has worked with Médecins sans Frontières in Cambodia, first in health district development and STI/HIV-related programmes, then as a member of the medical coordination team, presently as deputy medical coordinator. (E-mail: depmed@msf.org.kh) Heng Kimvan, MA in Lao literature and linguistics (Vientiane University), has been a literature lecturer at the Royal University of Phnom Penh since 1992 and a socio-cultural researcher with the Center for Advanced Study in Phnom Penh since 1995. He has research experience in the fields of ethnic and indigenous people, human rights, democracy and health. He has conducted several qualitative and anthropological field studies on health-seeking behaviour and health systems for the World Health Organization and Uppsala University, Sweden. (E-mail: cas@forum.org.kh) Bruno Meessen, MA (Economics), worked with Médecins sans Frontières from 1993 till 1999. He developed his expertise in health care financing in Africa and Asia. In 1999 he joined the Department of Public Health of the Institute of Tropical Medicine, Antwerp, as a research fellow in health economics. Presently his main interests are performance contracting, social assistance and new institutional economics. (E-mail: bmeessen@itg.be) Correspondence: Wim Van Damme, Professor of Public Health, Institute of Tropical Medicine, Nationalestraat 155, 2000 Antwer- pen, Belgium. 32 Wim Hardeman et al.