1In Cambodia, the successive conflicts in the 1970s and 1980s (the Vietnamese invasion and especially the Khmer Rouge period) left the country battered at the start of the 1990s, with considerably less trained and educated human capital (Lefait, 2005). The relative political stability since 1997 has allowed some economic and social indicators to improve. In 2004, the human development index (HDI) was 0.58, compared to 0.51 in 1990 (1 being the maximum possible). Despite this improvement, poverty is still widespread and Cambodia is one of the most needy countries in Southeast Asia. Its per capita GNP is currently approximately 350 US dollars according to the World Bank Poverty Assessment Report (2006).
Table 1. Primary demographic and health statistics for Cambodia
2This economic and social situation has heavy consequences in the field of health. Today, two health care systems co-exist in Cambodia:
An under-utilised public sector that offers poor quality care at high real prices (official fees plus unofficial additional charges). This concerns three levels of intervention in Cambodia: the community health centres, the operational district referral hospitals, and the provincial hospitals.
An unregulated and growing private sector, primarily in urban areas, that offers somewhat irrational (notably over-medicated) care at very high prices. This sector is made up of private physicians, private drug sellers (with or without pharmaceutical training), and traditional therapists.
3Overall, one can speak of the inappropriate use of both private and public sectors, with considerable self-medication, multiple consultations, and seeking care late. As a result of the suppression of the modern health care system during the Khmer Rouge period, self-medication is a widespread phenomenon in Cambodia. In the vast majority of cases, the patients’ first recourse is to buy medications directly, without consulting a health care provider. If their health does not improve, then they consult a private physician, and turn to the public facilities only as a last resort. At the same time, they sometimes also consult a traditional healer in an attempt to increase their chances of recovery. Thus, the public health care system is under-utilised. Less than one-quarter (21%) of people seeking care turn to public facilities first (National Institute of Public Health and National Institute of Statistics Cambodia, 2006). This percentage is even lower for second and third consultations.
4The principal criticisms of the public sector are its geographical distance (compared to private practitioners who make house calls) and the cost of the transportation involved, staff that is barely present, very limited opening hours, medication frequently out of stock, highly variable costs, and a lack of respect for patients (Annear et al., 2006).
5The government of Cambodia has not been inactive in response to this situation. The result of a reform process, the Health Financing Charter introduced cost recovery in 1996. The purpose of introducing an official fee scale was to eliminate parallel payments for care by users and allow health care workers to be better paid. However, institutional constraints (centralised staff management, limited verification and sanction capacity) did not allow the hoped-for results to be attained. In addition, drug shortages have forced most health care centres to devote some of their revenue to additional purchases.
6Later, the Cambodian government made a genuine effort to formulate a national health policy that resulted in the definition of a “Masterplan for Social Health Insurance” in 2005 (Cambodia Ministry of Health, 2005). Nevertheless, while the country’s primary care coverage has undoubtedly improved, one can note that the government’s efforts focused mainly on developing health infrastructure and on training courses, more than on any significant improvement in access to care.
7Indeed, health care financing is still one of the main stumbling blocks in the system. Currently, the government finances only 25% of today’s total health budget and donors are estimated to finance 10%. The remainder is financed by users, which represents a very high cost for households, particularly in rural areas (approximately US$18 per person per year). Despite the reform, users continue to snub the public sector, which is seen as being poor quality and not having clear-cut fees. The consequence of the financing difficulties is that health care workers in the public sector are still paid too little. As a result, staff members show high absenteeism, a lack of motivation, and recourse to parallel private practices. Government supervision of the system is still weak.
8Since 2002, the Ministry of Health has shown an increasingly strong interest in the subject of health insurance in Cambodia and its possible contributions to the direct financing of health care by users, as can be seen in its 2003-2007 and 2008-2015 Action Plans (Cambodia Ministry of Health, 2008). In 2005, an Inter-Ministerial Committee prepared a directive on the establishment of health insurance schemes in Cambodia. Finally, a Ministry of Finances decree is being prepared to ensure regulation of the micro-insurance sector.
9Since the government of Cambodia recognises the potential interest of micro health insurance in ensuring diversified financing of health care, it seems legitimate to ask to what extent private micro health insurance can help improve public health care. While much has been written about micro health insurance and about public sector health care in Cambodia (Bigdeli, 2007; Levine, Gardner, 2008), little has yet be published on the links between micro health insurance and public health care. Our article aims to address this gap.
10After presenting the project methodology and the research methods used, we will show which project system was set up locally to coordinate micro health insurance and public health care, and what impact the implementation of this programme has had. We will then discuss the limitations of the work accomplished and the theoretical and practical lessons that can be learnt. Finally, we will conclude with the prospects for developing this type of approach.
11The SKY micro health insurance programme was launched by GRET (the Research and Technological Exchange Group, a French NGO) in Kandal Province, Cambodia, in 1998. The project initially took an experimental approach because of the lack of references on micro health insurance in Cambodia, and then slowly stabilised its intervention methodology. Today, it covers more than 40,000 people in eight operational districts in urban and rural areas (see Appendix for a map), making it the largest community-based health insurance scheme in the country. SKY’s objective for the next four years is to cover all of Takeo Province and spread to Phnom Penh in order to assess the sustainability of its system and the potential for financial cross-subsidisation between urban and rural insurance schemes and between formal and informal populations.
12SKY offers “first dollar coverage,” meaning subscribers are covered from the first dollar spent on care. It covers both primary health care at health centres and hospital care at hospitals with which SKY has agreements. Coverage also includes free emergency transportation between Health Centres and hospitals, a funeral grant and, depending on the contributors’ wish, traditional music for funerals. There is no limit to the number of visits or cost of services to which subscribers have access. Coverage is not offered for chronic diseases such as TB and HIV/AIDS (though care and treatment of opportunistic diseases are covered), plastic surgery, dental care and glasses. Membership is voluntary, family-based and on a monthly basis. Premium varies according to four family size brackets, from US$11,70 per person and per year for a single person to US$3,57 per person and per year for large families. SKY works in partnership with public health facilities using a “capitation” scheme at Health Centres and Operational District Referral Hospitals, and a third party payment scheme at provincial hospitals. This system is detailed below.
13As the regulatory framework for health insurance is currently being elaborated in Cambodia, SKY has the status of a GRET project. Nevertheless, it is structured like a non-profit company, employing more than one hundred people in the back office and on the field. An information management system was set up (currently using Access, soon using My SQL) to track enrollment and health care consumption in detail and thereby progressively establish a basis on which to evaluate illness incidence risks, and produce monthly reports. In addition, short studies are conducted yearly on subscribers’ satisfaction, care-seeking behaviour, drop out rates, SKY’s notoriety and image, etc. Furthermore, client relations are monitored closely through regular village meetings, extended contact with SKY staff assigned to health care facilities, a complaints monitoring hot line, and exit surveys of beneficiaries following hospital care. Finally, SKY employs reference physicians who specialise in relations with partner health facilities and monitoring the quality of care and service provided. The project’s approximately ten years of experience and ten technical assistance missions by the SKY team to other micro health insurance systems in Cambodia have progressively made it possible to clarify the situation of and potential for micro health insurance in the country.
14This article is based on all of the data from the SKY project’s information management system over a five-year period starting once the intervention methodology had been stabilised. The data includes the number of subscriptions, the amount of health care consumed, the capitation payments to health care facilities, real expenditure by health facilities, etc.
15In addition to this system, four in-house studies were conducted.
16First, two quantitative studies were conducted: in July 2007 surveying 216 households in Ang Roka District, Takeo Province (21% of the insured population in the district) (Barbier, Miart, 2007); and in August 2008 surveying 108 households in Kampong Thom District, Kampong Thom Province (33% of the insured population in the district) (Goursat, 2008). The aim of these studies was to obtain feedback from the insured on the relevance of the service provided. These took the form of semi-guided interviews about the interviewees’ health seeking behaviours. They were conducted by Khmer investigators trained specially for these surveys. The interviewees were selected at random from the list of insured people.
17In addition, a study to assess subscriber satisfaction in the districts of Kampong Thom, Kampot and Koh Thom was conducted. It surveyed 279 households in July 2008 (Goursat, Portejoie, 2008). The semi-guided interview technique was used for this survey as well. The questions focused on how the insured viewed the service provided and the system’s organisation. This survey was also conducted by Khmer investigators, and the interviewees were chosen at random.
18Finally, a survey of SKY staff assigned to 23 health care centres in three operational districts (Ang Roka, Koh Thom, and Kampot) was conducted by SEGA/DOMREI, an independent organisation, in October 2008 (Levine et al., 2008). Semi-guided interviews were conducted on the changes seen by SKY staff since SKY had begun its activities in the region.
19GRET has been conducting rural development projects in Cambodia since the end of the 1980s. In 1991, the association launched a microfinance programme which today has become a legally acknowledged and financially independent institution named AMRET. This experience showed GRET that a health insurance product would be relevant to protect poor rural households against severe health expenses, as a complement to micro-credit (a high proportion of non reimbursement of loans was due to health problems in the household). For this reason, GRET began developing a micro health insurance scheme, SKY, in 1998.
20Several questions need to be answered when designing a health insurance product. In addition to the type of coverage offered (small risks, hospitalisation, payment of transportation costs), one of the major issues is the choice of provider for the covered services or, in other words, the health care supply that will be selected as partner. In order to evaluate which coverage system was most effective, GRET tested two health insurance schemes in two areas near the capital, Phnom Penh (Poursat, 2004):
An insurance system that included the provision of primary care in Kandal Province (a physician was hired as part of the insurance team). This model was justified by the lack of a public health care centre and the presence of private practitioners without degrees whose practices were unsatisfying and expensive. This test, conducted from 1998 to 2004, was able to cover the health care needs of 680 people.
An insurance scheme for hospital care in Takeo Province, in partnership with the health care providers in the area. The area was chosen because of the presence of a quality hospital with clear and moderate fees that was supported by outside institutions (Swiss Red Cross, French overseas aid). This experiment, which ran from 2000 to 2004, covered 800 people.
21The primary advantages identified for the first scenario were as follows: service quality, control of supply, the insured’s satisfaction with the physician’s availability, preventive care. However, the limited hours for the physician’s presence (a problem for night time emergencies and weekends that could not be staffed), the tension generated with private practitioners who refused to provide the insured with care, the relatively limited scope of the care covered, and the cost of the service for the insurance scheme showed the limits of this type of system. A new dimension was added to the debate in 2000 when a public health care centre was built in one of the communes where the scheme was active. The question of competition between the insurance scheme and public health care emerged sharply. GRET therefore opted to coordinate its project with the Ministry of Health’s policy and establish a partnership with the new health care centre.
22The second scenario tested was to establish a contractual relationship with a public hospital. This is the scenario currently in operation. The mechanism used is capitation, that is to say a flat-rate payment calculated per insured person based on the average incidence of illnesses and cost of treatment, paid at the start of the month to the health facility. Any positive or negative balance at the end of the month remains with the health care facility, which thereby carries the financial risk.
23In practice, while capitation is included in the contracts between GRET and the public hospital, a support process was set up via a reimbursement of excess costs during the first year of the contract in order to progressively build a reliable database on the incidence of illnesses and allow the hospital staff to become familiar with the capitation logic. In addition, the need to supply care providers with appropriate financial incentives and control costs is taken into account when calculating capitation payment amounts.
24In exchange, under the terms of the contract, the hospital must commit to meeting quality criteria for services and care (constant availability of essential medications, presence of minimum staff during the night, compliance with intervention protocols, etc.). These criteria were aligned with those set by the Cambodian Ministry of Health as part of its internal assessment of health care facilities. In the contract with the insurance scheme, a financial sanction is stipulated for repeated failure to fulfil these criteria.
25This type of partnership was later extended to primary health care through a specific calculation of capitation amounts and similar contracts. A mandatory referencing system is currently in place between the different levels of care, and a third-party payer system is in place for hospital care. Today, SKY has contractual relationships with fifty-five health care centres, ten referral hospitals and five provincial or national hospitals.
26In this model, micro insurance emerges as both a tool to foster utilisation of public primary health care and a means to lift financial obstacles to access to hospital care that, even subsidised, remains difficultly attainable for the vast majority of the population.
27The success of this partnership between SKY and public health facilities required a strong involvement on the part of the authorities at various levels of government. Very early on in the process, regular discussions were held with all levels of the Cambodian hierarchy: health facility staff and managers, and representatives of the Ministry of Health at decentralised (district, province) and central (secretaries of state, technical managers) levels. Seminars, presentations, meetings, workshops and individual interviews were held.
28The support of the Ministry of Health at the central level and of NGOs working in the field of health, and the real financial difficulties of primary health care centres, had a decisive role in overcoming the resistance of some health staff more accustomed to user fees or even fees for services than to micro-insurance services as they were developed by SKY. When roadblocks were encountered at the “intermediary” levels of the decentralised health administration (operational district, health facility managers), the Ministry of Health’s direct support was crucial.
29To obtain this institutional support, the project worked simultaneously with the Cambodian government and the country’s main health actors (working group on insurance, short training courses, etc.) to develop Cambodia’s health policy. In this way, the project contributed to the elaboration of the country’s health policy.
30As mentioned in regards to methodology, GRET conducted two surveys of its subscribers’ health seeking behaviours in 2007 and 2008, one in Ang Roka (Barbier and Miart, 2007) and the other in Kampong Thom (Goursat, 2008). These two districts show major differences in the amount of time insurance has been available since the SKY project was developed in Ang Roka as early as 2000 and only starting in 2007 in Kampong Thom. They also differ in the quality of available health care, deemed good in Ang Roka (Lefait, 2005) and mediocre in Kampong Thom because of a lack of staff and equipment (Briasco and Sophat, 2006). In these two areas, the first recourses in response to illness are as follows:
Table 2. Location of First Health Contact
31In both districts, recourse to public health care is higher than the national average of 21% (National Institute of Public Health and National Institute of Statistics Cambodia, 2006), reaching 33% in Kampong Thom and 69% in Ang Roka. By offering direct coverage by the health care centre or hospital, insurance helps redirect patient flows towards these structures and away from self-medication and private practitioners.
32Quality of public health care facilities is the key factor to understand the behaviour of patients. Whereas subscribers who do not use public health care facilities in Kampong Thom deplore the poor quality of the service (42% of responses), emphasising out of stock medications and the limited skills of health care staff (Goursat, 2008), Ang Roka’s subscribers deem the quality of care good and easily go to public structures to be cured (69% of first contacts). In both cases, the perception of quality of the insured people corroborates the outside assessment of the objective quality of health care (Lefait, 2005; Briasco and Sophat, 2006).
33The success of public health care can also be measured by the attendance rate of health care centres. In the centres working with the project, the average contact rates were 3.6 visits per person per year in 2006, 3.2 in 2007, and 2.1 in 2008, showing much higher attendance than the national contact rate of less than one visit per person per year and a good control of over-consumption.
34Thus, SKY’s private micro health insurance seems able to increase attendance in public health care centres. For this trend to persist, it must be supported by an improvement in the quality of the services offered.
35The service quality criteria set contractually in the SKY project have progressively been met by the public health care centres and hospitals: longer opening hours, night time and weekend emergency on-call staff, presence of all staff in the morning at least (and two people less in afternoons), regular supply of basic medications, etc.
36The survey of SKY staff (member facilitators) with office hours in the twenty-three health care centres (Levine et al., 2008; see methodology section) gave the following results in regards to practical changes in the centres:
Table 3. Changes in Health Care Centre Conditions Since SKY Began Operations
37Overall, the situation seems to have improved since SKY began operations for all criteria, mainly dealing with perceived quality and service quality (staff politeness). In addition, the satisfaction surveys among SKY members (Goursat, 2008) indicate a good level of perceived quality. Seventy-five percent of subscribers surveyed in Kampong Thom believe that they receive better treatment than non-subscribers (Goursat, Portejoie, 2008). Moreover, satisfaction rates are high: 92% of those interviewed expressed overall satisfaction with the service, and around 90% were satisfied with the behaviour of medical staff, waiting times, and the health care centres’ diagnoses.
38Health care facilities’ financial anticipation has improved. The insurance revenues paid at the start of the month by SKY to the health facility are more stable than revenues from user fees (which would seem to follow cost curves) and are on average higher than costs evaluated on the basis of user fees (by 15% in hospitals and by 33% in health care centres, in 2006 and 2007). The system allows public health facilities to smooth out their activity, invest in new equipment, and facilitate cash flow management (on time payment of permanent staff salaries, etc.).
39The figure showing health revenue and expenses for the Ang Roka District referral hospital —a facility treating large risks— is particularly eloquent (Figure 1). Capitation revenue paid by SKY is higher than health costs and more predictable than user fees. Only in the case of an epidemic like the one between January and July 2007, health care expenses may exceed capitation revenue. Overall, capitation revenue gives to the health facility management more latitude to manage its structure than do user fees.
Figure 1. Capitation Revenue and Health Expenses for the referral hospital in Ang Roka District
40This effect has a tendency to intensify as the share of capitation in overall revenue increases for a health facility. Today, SKY can account for up to 30% of the revenue of health care centres and up to 25% of the revenue of referral hospitals.
41The progressive adjustment of the statistics used to calculate capitation payment amounts made it possible to improve the attractiveness of this mechanism for public health institutions. Its effects on attendance, financial predictability and improved quality became progressively stronger for public health care. Capitation has become the contractual mechanism recommended by the Cambodia Ministry of Health in its national guidelines for community-based health insurance (Cambodia Ministry of Health, 2006).
42However, setting off a virtuous circle such as this involves a learning curve for both health facility staff and patients.
43In regards to staff, capitation generates a managerial logic that is very different from traditional modes of cost reimbursement. It notably requires strict control of over-consumption and over-prescription. The ability to rationalise health staff behaviour and optimise management of capitation payments—or in other words, ability to manage the facility—is a decisive factor in the partnership’s success. The reform initiated in 1994 by the government of Cambodia to ensure the progressive decentralisation of health care management gives greater responsibility to the operational districts (ODs, each made up of health care centres and a referral hospital), as emphasised in the Health Strategic Plan 2008-2015. Various initiatives have been taken to improve managerial capacities in the ODs, such as internal or external contracting with specialised NGOs. Because of this, GRET works in priority with health facilities that receive outside technical support (from GTZ in Kampong Thom and Koh Thom, and from SRC in Ang Roka). Establishing consistent quality criteria between the in-house technical assistant (GTZ, SRC) and the external insurer client (SKY), defining more transparent management standards for the regular reporting of accounts to institutional partners, and the proper in-house use of additional revenues are as many factors that foster the success of capitation. As mentioned by Akashi et al. (2004), in the case of the introduction of user fees, staff awareness is the key to the effective implementation of the scheme.
44Another hindrance to good synergy between private insurance and public health care lies in the population’s habits of self-medication and multiple consultations. These practices are difficult to eliminate. Patients know the market price of medications and attempt to negotiate with care providers the treatment of their illnesses as a function of the amount of money they have and their perception of the care needed in the very short term. This attitude is encouraged by the situation of public health care workers who cannot live on their state salaries alone as their only source of income (US$30 to US$40 per month, whereas the minimum necessary income is estimated at US$120). In this situation, it is very tempting to give in to patients’ demands for treatment, and charge as much as possible for services and the medication that patients cannot buy over the counter in pharmacies. This affects the quality of care provided. As De la Seiglière (1999) summarised: “Patients and professionals maintain a relationship that is harmful to patients’ health and financial capital, but that satisfies them both in the short term.” By offering monthly salary supplements indexed on attendance, capitation slows the development of these parallel private practices.
45Another related aspect deals with the perception of patients of the quality of the care they receive. A study in 2001 emphasised that the patients’ strong demand for injectable medications was leading to the perception that SKY did not provide adequate care. A positive view by the public of SKY’s products was key to growth. However, SKY could not provide dangerous coverage and therefore education on this issue needed to continue (McCord, 2001).
46In all cases, the work accompanying and monitoring patients is crucial. The proximity of SKY staff in health care centres and the mobilisation of an advising physician allows for closer dialogue with insured people and ongoing awareness raising on better health practices.
47Until now, the Ministry of Health has emphasised micro health insurance in the aim of diversifying financing for the health care system, notably through direct financial participation by households.
48The SKY project has one point in common with this approach: selecting capitation-based partnerships for primary care makes it possible to ensure a certain degree of cost recovery for health care facilities. However, the project’s primary objective differs from the Ministry’s ambitions: the goal is not primarily to make the public health care system profitable, but to preserve household budgets. The positive impact on health facilities and the quality of care is, in some way, a hoped-for trickle-down effect and not a goal in its own right. The logical framework of the intervention is thus out of step with the government’s aims.
49Thus, deciding to build a partnership with the public health care system is a deliberate choice on GRET’s part. Indeed, to offer access to quality care, it could be entirely acceptable for an insurer to work with the private sector if the latter is more effective than the public sector and charges fees that are acceptable for the insured. A strong national policy geared towards improving the quality of public health care is therefore absolutely necessary for health insurance schemes such as the one developed by the SKY project to be able to function.
50What is the take-home message from this analysis? Based on GRET’s micro health insurance project in Cambodia, the aim was to determine the extent to which non-profit private micro health insurance can help improve public health care.
51We have seen that by taking an experimental, progressive approach, the SKY micro health insurance programme implemented by GRET in agreement with the national authorities has enabled an increased utilisation of public health care, an improved perception of the quality of care, and better financial anticipation abilities by both health facilities and the insured. However, the limitations of and obstacles facing such systems are numerous as self-medication among the insured persists, the attitude of poorly paid health care workers does not favour rational use of medicines, and the orientation of Ministerial financing for health care is not always consistent with the project’s objectives. Ultimately, the synergies between non-profit private micro health insurance and public health care will be more effective when insurance accounts for a large share of health care facilities’ revenues—that is to say when the subscriber base is large.
52Today, the challenge is to go beyond the current project and take into account the poorest populations that do not have access to the micro health insurance system. A partnership between GRET and a Health Equity Fund to attain this is currently being tested in Kampot Province. The SKY premiums of very financially disadvantaged individuals are paid by the subsidy fund (implemented by the Ministry of Health, with support from GTZ, AusAid and World Bank), providing these individuals with free access to health care. This link between a Health Equity Fund and the insurance scheme promises to increase the scheme’s leverage, triggering more quickly and more strongly the virtuous circle between non-profit private insurance and public health care.