Disintegrated care: the Achilles heel of international health policies in low and middle-income countries.

PURPOSE: To review the evidence basis of international aid and health policy. CONTEXT OF CASE: Current international aid policy is largely neoliberal in its promotion of commoditization and privatisation. We review this policy's responsibility for the lack of effectiveness in disease control and poor access to care in low and middle-income countries. DATA SOURCES: National policies, international programmes and pilot experiments are examined in both scientific and grey literature. CONCLUSIONS AND DISCUSSION: We document how health care privatisation has led to the pool of patients being cut off from public disease control interventions--causing health care disintegration--which in turn resulted in substandard performance of disease control. Privatisation of health care also resulted in poor access. Our analysis consists of three steps. Pilot local contracting-out experiments are scrutinized; national health care records of Colombia and Chile, two countries having adopted contracting-out as a basis for health care delivery, are critically examined against Costa Rica; and specific failure mechanisms of the policy in low and middle-income countries are explored. We conclude by arguing that the negative impact of neoliberal health policy on disease control and health care in low and middle-income countries justifies an alternative aid policy to improve both disease control and health care.

Introduction

The history of international aid can be seen as cyclical. In the 1950s and 1960s, aid policies for Africa and Asia focused on disease control. Citizens of the colonies had little political weight and access to health care was not seen as a priority by the ruling powers. In 1978, a new strategy—Primary Health Care (PHC)—was approved at the Alma Ata Conference. It promoted comprehensive care and community participation in public services, echoing the 1970's social-democrat mood. This concept led to several confrontations between WHO and multinational companies (on breast milk substitutes and essential drugs), with the US even withholding its contribution to the WHO's regular budget in 1985 [1]. This caused a return to the vertical programmes strategies of the 1950s for developing countries. Within a year of the Alma Ata conference, Walsh and Warren had argued [2] for a reduction in the scope of Primary Health Care to the control of 4–5 diseases, a strategy labelled “Selective Primary Health Care”. The Rockefeller Foundation and UNICEF stated that the public sector should be selective in the services offered. This policy was criticised on the grounds that comprehensive Primary Health Care including the same disease control objectives but securing access to health care incurred the same costs as selective Primary Health Care [3]. Numerous scientists mobilised against this initiative [4], but failed to curb the US policy which effectively soon had support from the World Bank.

This paper aims to review the evidence basis of this international aid and health policy. It does not focus primarily on understanding its causality, even though the paper addresses health policies in political terms.

Disease control, the favoured model of international aid has failed in much of its performance in low and middle-income Countries (LIC/MIC) despite intensive financing. We suggest that international aid policy shares a responsibility for this. With its promotion of commoditification and privatisation, much international aid policy can be seen to reinforce neoliberal health policies in recipient countries. In assessing its impact on disease control, we discuss a three-part hypothesis:

The vast majority of disease control interventions are clinical in essence. To be effective, they generally need to be integrated into health care delivery services. Such integration requires health facilities with patients: a sufficient pool of users needed by disease control programmes for early detection.

In theory both public and private sectors can provide integrated disease control activities. Yet international agencies have been reluctant to allocate disease control to the private for-profit sector — for good reasons.

Instead international agencies have promoted the continuing involvement of governments, non-governmental organisations (NGOs) and communities in disease control. At the same time they have applied the neoliberal principle of health care privatisation, thereby precluding integration and leading to unacceptable disease control performance.

We then describe how health care delivery has gradually shifted from the public to private sector and examine the impact of this on access to health care in LIC/MIC. Our analysis consists of three steps. Firstly, pilot local contracting-out experiments are analysed. Following this the national health care records of Colombia and Chile, two countries which have adopted contracting out as a basis for health care delivery, are contrasted with Costa Rica. Lastly, specific failure mechanisms of the policy in low and middle-income countries are explored.

The poor performance of disease control programmes

Disease control, the favoured model of international aid, in low and middle-income countries (LIC) is a weak performer. Despite a ten-fold increase in external financing for tuberculosis control in LIC over the last decade [5], only 27% of bacilloscopic positive pulmonary tuberculosis cases have access to the package set out in the DOTS strategy [6, 7]. Some experts consider that DOTS programmes are insufficient as a tuberculosis control strategy [8]. As for HIV, 3 out of 4.8 million new infections in 2003 occurred in sub-Saharan Africa [9]. Countries such as South Africa and Zambia suffer from a HIV prevalence rate of more than 20%, with Swaziland and Lesotho over 30% and Botswana approaching 40% [10]. AIDS still kills more than 8000 people every day [11], of which more than 6000 are in sub-Saharan Africa. By mid 2004, less than 5% of AIDS patients in this region were under treatment, compared to over 50% in the Americas [12]. Malaria is another problem story, with up to 3 million deaths every year [13] and an increasing population at risk [14]. The United Nations stated in September 2005 that the Millennium Development Goals related to tuberculosis, malaria and AIDS control could not be met in Latin America, Sub Saharan Africa, South and South East Asia. Furthermore, little progress was seen on AIDS and tuberculosis in Eastern Europe and Central Asia.

In the following we outline the responsibility of neoliberal health policy for this failure using a three-strand hypothesis.

Integration: a key to success for disease control

Disease control activities implemented by specialised organisational structures, sometimes bringing together several disease control programmes (such as maternal and child health) are dubbed vertical programmes. In specific cases vertical programmes can be justified on technical grounds [15] such as:

Vector control;

The control of diseases too rare for generalists to maintain the necessary specialist skills;

Outreach to specific risk groups, e.g. commercial sex workers or drug addicts;

The control of epidemics and emergencies;

The provision of health activities for which there is no demand, e.g. epidemiological surveillance.

Nevertheless, the number of diseases requiring clinical interventions makes it impossible to consider vertical programmes as the gold standard template for disease control organisation, even where these programmes are closely co-coordinated amongst themselves [16]. There are, for example:

Diseases or health issues already addressed by programmes (e.g. AIDS, tuberculosis, malaria, onchocerciasis, immunisations, family planning, acute respiratory diseases, acute diarrheal diseases, poliomyelitis, leprosy, Chagas, Guinea worm);

Virtually all neglected diseases rely partly or completely on clinical components: Soil transmitted helminths and schistosomiasis (prasiquantel, mebendazole, albendazole); lymphatic filariasis (diagnosis of acute adenolymphangitis attack, lymphoedema and scrotal swelling); leprosy (early diagnosis and multidrug therapy); visceral leishmaniasis (early diagnosis and treatment at hospital); onchocerciasis (blindness rehabilitation and ivermectin, not only in mass distribution); Guinea worm (early detection and treatment); trypanosomiasis (early diagnosis and treatment in specialised centres); trachoma (antibiotics if prevalence under 20%); Cholera (rehydration, vaccination, antibiotics), Rabies (curative vaccination) and Buruli Ulcer (tuberculostatics);

Chronic degenerative pathologies (e.g. cardio- and cerebrovascular diseases and diabetes).

The need to integrate programmes into local health facilities in order to achieve a reasonable prospect for successful disease control is stressed by many authors [17-22]. They also point to the merit of integrating curative and preventive care. Examples include the potential for detecting a patient with tuberculosis amongst those with cough, or suggesting vaccination to a patient or to a population with whom the practitioner has established trust [23].

Public rather than private disease control: understandable caution

In theory, both the public and the private sector can carry out disease control activities, but historically the public sector has taken this responsibility. Despite the widespread promotion of public-private partnerships [24-27], international aid agencies have been cautious about contracting out disease control to the private sector. Instead, such agencies have promoted continued involvement of government facilities in disease control under the general label of ‘prioritisation’ of their interventions. Their caution is understandable. The results of contracting out disease control to the private for-profit sector are not promising [28, 29], except for tuberculosis control under specific conditions [30]. Furthermore, private providers may not oppose the provision of disease control by the public sector, as they often work part-time in public services providing the opportunity to refer the selected patients to their own private consultation [31].

Public disease control & private health care: a catch-22

While disease control remains public, aid agencies have been encouraging a market approach to health care delivery in LIC for over a decade [32-35]. The transfer of ‘public’ care to the private for-profit sector is a core message of their policy on the grounds of the supposed higher efficiency of the for-profit sector and the poor responsiveness of the public one. Once predominantly providers, governments now have new roles as ‘stewards’, steering care by regulation and supervision. In theory, such privatised care could be funded publicly.

Figure 1 indicates these roles of the public and private sectors as promoted by neoliberal health policy.

Figure 1

International aid-promoted health policy.

This doctrine was introduced in LIC and MIC where the market was seen as attractive, such as in parts of Asia and Latin America [36, 37]. It was seen as less relevant in contexts where the market could not be developed easily, as in many parts of Africa. It was also not promoted in countries where geo-strategic considerations dictated an aid policy with clear social goals, as is the case in Jordan, the Southern Philippines and those central Asian republics close to Afghanistan.

One outcome of this policy was a disease control focus to Ministries of Health (MoHs), with less support for health care delivery. The continued concentration on vertical disease control efforts by the international aid community is reflected in the efforts to set up and channel significant aid through the Global Fund to Fight AIDS, tuberculosis and malaria and which has been criticised strongly for ignoring the needs of, or even weakening, the wider health system.

This aid and health policy precluded effective integration in the field and led to a true catch-22; the pool of patients was cut off from disease control interventions and ended up achieving substandard detection and follow-up rates. This point has been mathematically demonstrated in a recent paper. In order to assess the potential for integrating malaria control interventions in underused health services, a Piot predictive model was used to estimate malaria cure rates. Parameters from the best performing African malaria programmes influencing treatment at home and in health facilities were applied to a rural district in Mali, where access to care was very limited. It was demonstrated that, with a low utilisation rate, adequate control combining home treatment and professional treatment was impossible, even applying the best parameters from other countries. On the contrary, cure rates with a higher utilisation rate were 62% better. Thus, if malaria patients are to be treated and followed up early, basic health services need to deliver integrated care and be attended by an adequate pool of users [38].

Another example of failing disease control is Colombia, where TB case finding decreased after neoliberal reform [39, 40].

Furthermore, disease control programmes strained first line public health care delivery. This occurred through pressure exerted by disease control managers [41], by multiplication of disease-specific divisions in (inter)national administrations, by ill-defined priority-setting and increasing opportunity cost [41], unrealistic costing, inadequate budgets, and financial overruns [21], failure to make clear the lines of command [42]; tension between health care professionals over income disparity, treatment discrepancies and opportunity costs [43] and problems with sustainability [21, 41–46].

Management by objective, the philosophical cornerstone of such programmes, transformed health organisations into what Mintzberg classifies as mechanistic bureaucracies [47].

Poor access to health care delivery

Access to health care in LIC and MIC is also disappointing. Despite, or in some cases, because of, more than a decade of reforms, almost 50% of health systems do not provide adequate access to care for their citizens [48]. Access to care is particularly difficult in China and in the republics of the former Soviet Union. It has also deteriorated in Latin America. A third of the world's population has no reliable access to essential drugs; this rate rises to above 50% in the poorest countries of Africa and Asia [49].

We raise the question, as we did for the poor performance of disease control programmes, as to whether international aid policy, shares responsibility for this poor access to health care.

In most LIC, the use and quality of government health facilities has fallen to an all-time low. Burkina Faso saw its utilisation of health care services (expressed by the number of sickness episodes per inhabitant per year) dropping from 0.32 in 1986 to 0.17 in 1997 [50]. Uganda's University Hospital of Mbarara is pictured by Kavalier as “decades away from the end of the 20th century” [51]. Even in MIC, the problem exists. Ellen Roskam asks: “Where can you be treated by a doctor who, last year, worked a 1000 hours more than his official timetable, who has earned less than 15$/month, who has not been paid for 5 months, who has worked without drugs or bandages in an operating theatre with a leaking roof and where there has been no investment for many years? Maybe somewhere in Africa? No, in Eastern Europe” [52]. Managers of public health services faced major challenges in the light of the contradictions imposed by reforms [53].

Meanwhile, care in the private health facilities remained expensive, inefficient and unregulated. Care suffered from accessibility and quality problems. Cream skimming, semi-monopolistic settings and corruption were widespread. The increase in unnecessary prescriptions, admissions, length of stay, lab tests and medical imaging exemplifies how profitability dominated evidence-based practice. In Brazil, the caesarean section rate reached 31% in the public sector, and 72% in the private sector [54].

Admittedly, poor access to decent health care in LIC is associated with multiple and interlinked determinants:

An economic crisis, resulting in decreasing purchasing power, downsizing of public services and falling salaries, in most highly indebted poor countries [55]. In Latin America, deteriorating access to publicly delivered health care was associated with the public services' narrowing problem-solving capacity. Furthermore, professional associations sought part-time employment in public services enabling private doctors to spot profitable patients and send them to their own clinics.

Static or increasing social inequalities, as observed in the majority of countries in Latin America and the Caribbean, the region with the highest inequality in the world [56];

A combination of external and internal adverse conditions—the simultaneous effect of Structural Adjustment Programmes (SAP) and insufficient social control on administration, lack of democracy, patronage and nepotism, resulting in low priority given to the social sector [55, 57].

However, the health policy promoted by international aid agencies may also have played a specific role through the promotion of the private sector. We distinguish five different, sometimes complementary, ways in which the transfer of health care delivery to the private sector has been fulfilled:

Governments underfinanced public services, allowing the private sector to offer care without having to deal with subsidised competition. Typically, this was the scenario adopted by Sub Saharan Africa and Andean countries.

Governments, accepting the efficiency arguments of the international agencies, gradually reduced the operational role of the public sector to a greater focus on disease control programmes. International agencies financed such programmes.

Governments subcontracted health care to the private sector (in a very limited number of countries, such as Lebanon, Colombia, Zambia). In Latin America, International Financing Institutions also promoted and financed the privatisation of health care.

Governments leased or sold public hospitals to the private sector. The best-known experiences, promoted by World Bank authors as options of public–private partnerships, are Stockholm's St. Goran's and a few converted Australian hospitals [58]. In low and middle-income countries, this pathway was the exception rather than the rule. Examples included the former Soviet Union and Albania [59].

Governments granted managerial autonomy to public hospitals, blurring the boundaries between public and for-profit objectives.

Policy actors have cited all but the first of these five pathways as beneficial. In practice, however, it was the under funding of the public sector that most frequently led to privatisation. This under funding happened in many countries in the 1990s, as evidenced by the gap shown by the WHO Macroeconomic and Health Commission Report [60]. The World Bank's (WB) SAPs and the International Monetary Fund's (IMF) Enhanced Structural Adjustment Facilities (ESAFs) effectively reinforced the liberalisation of services by starving them of public resources.

One consequence of this under funding is now becoming apparent, with the haemorrhaging of professional staff to the private sector and increasingly to industrialised health systems, understandably attracted by the higher salary pull-factors.

The Bretton Woods institutes never effectively enforced the loan conditions to increase public social spending, a responsibility also neglected by LIC and MIC governments. International aid never compensated for the reduction in government health expenditure. Rather it reinforced the second scenario, driving the fading public sector to focus on externally financed disease control activities, straining the public delivery of care and thereby creating a market opportunity for the private sector.

We turn now to discussion of contracting experiences drawing on pilot experiments and national data.

Evidence from pilot experiments: contracting out and managerial autonomy

The public private partnership (PPP) approaches were developed in the context of New Public Management. PPP encompasses both Private Financing Initiatives (PFI) and contracting out. The former refers to private money being used to finance health care infrastructure that previously had been under government responsibility. So far, this technique has been mainly applied in the United Kingdom and less so in developing countries. The other approach, contracting out, has met difficulties over the last decade. In Southern Africa, researchers compared the operational cost per admission and per in-patient day between public rural district hospitals and subcontracted private for-profit hospitals [61-63]. The few studies do not support the hypothesis that efficiency increases when care is subcontracted to private companies [62-64]. These studies suggest that a similar quality of care can be achieved at a lower cost in the private sector, but at a total higher cost to the public authorities once private profit margins are included. The private providers' profit margins override the financial gains arising from improved cost-efficiency. The only positive outcomes were associated with buying care from NGOs, where there was no profit motive.

In addition, few studies have ever examined whether the pricing policy applied by the private sector is equitable [65, 66] and, to our knowledge, there has been no study to determine whether the private health sector in LIC improves patients' independence from professional care. However, the quest for gains that underpins the private sector is unlikely to favour solidarity or increase the patients' medical autonomy.

The strategy of developing hospital autonomy was expected to facilitate reinvestment and to improve staff motivation without drawing on the national budget. Managerial autonomy became a widespread policy in a number of low and middle-income countries. However, in practice government hospitals were often granted managerial autonomy without defining their objectives or providing supervision. Health staff often managed “autonomous” public hospitals as private facilities [67], enabling public hospitals to adapt a for-profit rationale while retaining public funds. This happened, for example, with Chinese state hospitals depending on profit-making private services to break even [68]. This pathway led to increased costs for users and to a reorientation of the bulk of hospital activities from secondary care to simple, first-line clinical services. Underpaid staff quickly understood that hospitalising a rich person with bronchitis, for example, instead of a poor patient in need of a surgical intervention, generated income. Opportunistic cream skimming changed the case-mix. Neither the state nor communities were able to exert control. This managerial stance created obstacles for poor patients needing secondary care and favoured instead hospitalisation of (middle class) primary cases. Similar processes can explain why private hospitals in Thailand had a shorter length of stay, a larger percentage of admissions of children and yet no increase in utilisation of their operation theatre [69].

Lastly, staff motivation rarely went beyond an understandable eagerness to access hospital profits to supplement salaries, which in turn hampered reinvestment [70].

Evidence from national data: contracting out equals inefficiency, ineffectiveness and inequity

Many international blueprints advocated subcontracting care to the private sector in LIC, but few countries actually did it. This is because most LIC and MIC were unable to generate the required funds. Government finances were mainly earmarked for wages leaving little latitude for contracting out. While in Western Europe government spending in 2001 represented more than 70% of total health expenditure, it was less than 40% in China, 30% in Vietnam and 20% in India [71].

A few countries did, however, manage to subcontract to the private health sector, and we now scrutinise health care privatisation in two of them—Colombia and Chile—contrasting them with Costa Rica which maintained publicly oriented social insurance and health services. In 1993, Colombia adopted a purchaser/provider split and contracting-out policy and committed resources to it. The State largely freed itself from direct provision of services. The exceptions were disease control activities and health care delivery for the non-insured and for pathologies not covered by health insurance—mainly in public hospitals.

Chile partly privatised its health insurance, whilst keeping a large public health sector, through its public social insurance and its public health services. Private health insurance and care have never been more than a separate, marginal health system for the well-off. ISAPRES private health insurers never covered more than 25% of the population, and are now at less than 16%. The strong backbone of a public health system has not been broken by the dictatorship of Pinochet even though public health services financing was drastically reduced. Since 1989, democratic governments have steeply increased public health expenditure.

In Costa Rica at the end of the last century, the average private health expenditure was only 25% of total health expenditure in contrast with 58% across Latin America [72]. The Social Security Administration of Costa Rica (SSAC) is the single (and public) health insurer in Costa Rica (private health insurance exists but is marginal). It both purchases and provides care.

We argue firstly that, unlike Colombia, health systems' outputs in both Chile and Costa Rica are excellent. Secondly, Costa Rica was more efficient than Chile in securing health status for its population.

The achievement claimed to illustrate the Colombian reform success was an advance in social insurance coverage which almost doubled from 31 to 62 percent of the population between 1992 and 2004 [73] leading to affiliation for 18.5 million (from a target population of 22 million), with a remarkable acceleration of affiliation in 2004 and 2005, after stagnation for several years [74]. However, caution is warranted. The benefit package for the poor is still only half that for the contributing affiliates. Furthermore, whereas before the reform, 61.7% of people needing health care were seen by a doctor, this proportion had fallen to 51.1% in 2000 [17, 24]. Each year 6.26% of the population suffered catastrophic health expenditure, and the poorest quintile had out-of-pocket payments four times higher than the richest [75]. Even more worrying, according to a periodical national survey, the proportion of people consulting in the month before the survey increased from 1993 to 1997 (from 7.5% to 23.8%), but then decreased dramatically in 2003 to 9.5%, meaning that theoretical high coverage by health insurance did not translate into higher utilisation rates of health services, in spite of increased health expenditure (from 7 to 10% of GDP) [76].

In contrast in Costa Rica, utilisation of medical health services is high with acceptable, affordable and perceived good quality health care, compared to other developing countries. Only 0.12% of the Costa Rican households suffered from catastrophic health expenditure, 52 times less than in Colombia.

Given the inequity within both Chilean society and the health system's financing, one would expect accessibility to care by the poor to be limited. However, among those who declared having felt sick in the last month, 73.9% from the poorest quintile sought care compared to 79.7% from the richest [77]. Data on utilisation rate also confirms this relatively equitable access with increases from 2.65 and 3.27 in 1990 to 3.85 and 4.12 in 1999, in FONASA and ISAPRES, respectively [78]. However, the utilisation rates still differ between poor and rich municipalities: by a factor of 2.8 for Primary Health Care, 3.9 for emergencies and 2 for inpatient care [79].

Infant mortality rate is known to reflect general social and economic conditions, and not solely access to medical care [27]. However, child mortality due to acute respiratory infections and acute diarrhoeal diseases, can be viewed as avoidable mortality and used as tracer pathologies for quality of care [28], including in less developed countries [29]. These rates have clearly increased in Colombia since 1997 [16]. Perinatal mortality is also known to be an indicator for access to quality health care. It doubled from 1996 to 1997 and continues to rise [16]. The same goes for maternal mortality, stable since 2000 at an unacceptable high level of about 100 deaths per 100,000 newborn [80]. By combining middle income with high human development since 1995, Costa Rica achieved a life expectancy at birth of 78 years (second only to Canada in the Americas), an infant mortality rate of 9/1000, equivalent to a sevenfold reduction over the last three decades (equivalent data in Colombia is 19 with a fourfold reduction) and a tuberculosis prevalence of 19/100,000 (69 in Colombia). Several of these features are related to the social commitment of successive Costa Rican governments [81] in particular public expenditure on health and education was 4.9 and 4.7% of GDP in 2001 compared to 3.6 and 4.4% in Colombia [81]. However, the pivotal role of health services and policy played by Costa Rican human development should also be recognised. Numerous indicators suggest an impact directly attributable to health services:

Comparison of IMR and MMR, with Chile and Colombia both at similar income levels, reveals the significant advance of Costa Rica (see Table 1);

Table 1

Infant and maternal mortality rates, Costa Rica, Chile and Colombia 1

			Reduction
	IMR	IMR	IMR	MMR
	1970	2001	1970–2001	2001
Costa Rica	62	9	: 7	29
Chile	78	10	: 8	23
Colombia	69	19	: 4	80

Infant mortality rate as the probability of dying between birth and exactly one year of age, expressed per 1,000 live births (1970 and 2001 figures, from UNICEF); infant mortality rate reduction 1970–2001 calculated from the 1970 and 2001 figures; maternal mortality ratio as the annual number of deaths of women from pregnancy-related causes per 100,000 live births (data refer to the most recent year available during the period specified, from UNICEF's The State of the Worlds Children 2003, based on national data, adjusted for underreporting and misclassification in a joint effort by UNICEF and WHO); health expenditure per capita in PPP US$ (from WHO); GDP per capita in PPP US$ (from WB's World Development Indicators); GDP per capita annual growth rate (aggregates from WB for the Human Development Report Office).

Sources: United Nations Development Programme (2003) Human Development Indicators. In Human Development Report 2003. Millennium Development Goals: A compact among nations to end human poverty. UNDP, New York, NY, USA, pp 237–339, http://hdr.undp.org/reports/global/2003; United Nations Development Programme (2004) Human Development Indicators. In Human Development Report 2004: Cultural liberty in today's diverse world. UNDP, New York, NY, USA, pp 139–250, http://hdr.undp.org/reports/global/2004.

Perinatal mortality rate dropped from 12.0/1000 in 1972 to 5.4/1000 in 2001 [84], which suggests obstetric improvements;

Pneumonia specific mortality in under ones dropped from 5.4/1000 in 1972 to 0.3/1000 in 2001 [82], which suggests improved and faster access to health services;

Tuberculosis specific mortality dropped from 7.2/100,000 in 1972 to 4.4/100,000 in 2001 [82], despite increased incidence, which suggests a functioning programme.

Finally, in Chile, health indicators are good, with high life expectancy and very low infant mortality and maternal mortality.

These contrasted achievements were not explained by the level of investments in health, but rather the contrary. Health expenditure rocketed since the introduction of managed competition in Colombia, with an increase from 7 to almost 10% of GDP [76]. In Costa Rica, the excellent outputs were achieved at a moderate cost; one comparison summarises the Costa Rican health policy achievements: the country spends 9 times less on health than the USA and scores better on life expectancy.

In Chile, private ISAPRES spend 3 times more on administration per affiliated than public FONASA (about 20% vs. 6%). In 2000, the GDP share allocated to health was 7.3%, of which 3.1% was public and 4.2% private. Out-of-pocket expenditures were 27% of total health expenditure.

The Colombian health policy in the last 2 years achieved a good result in health insurance coverage of the poor, but with a very limited benefit package, a reduction in real access and at a very high cost (almost 10% of GNP), due in part to the coexistence of supply and demand subsidies and to a generalised evasion of contributions to social health insurance.

In Chile, the overall good demographic and epidemiologic records may in part be due to the sustained high economic growth rate, and the spectacular reduction in poverty: from 39% in 1990 to 21% in 2000, without precedent in Chile and also exceptional internationally [83, 84]. The excellent health indicators are due to this public system, good education levels and economic growth, and hardly to ISAPRES and private providers who attended the healthy, young, and well-off urban. Instead, the private care insurance and delivery can be incriminated for the high cost of the Chilean health system, for its deficient solidarity and lack of equity.

In conclusion, the evidence of the three countries appears to contradict international recommendations to privatise health care in developing countries. The remarkable sustained achievements of Costa Rica suggest the following:

A unified public health services system, in which government expenditure represents the bulk of total expenditure, permits integration;

Dominant, though non-monopolistic publicly oriented services offer accessible health care;

Contracting-in secures both management and production targets (as opposed to the much promoted contracting-out).

A single, public insurer (private insurance being virtually non-existent), which contributes to solidarity and general access to care

Users and communities can participate in publicly oriented health services management, in contradiction to what the private for profit sector permits.

Unmet conditions for contracting-out: the mechanics of a failure

The common cause of failure (both pilot and national) of contracting-out in low and middle-income countries lies in a mix of technical and political features, as we now discuss.

If we assume that contracting-out guarantees access to good quality health care in Western Europe, why is it failing in LIC and MIC? We would argue that the technical requirements for contracting-out clash with political and administrative reality. Efficient subcontracting thrives on effective control and regulation, which is an unfulfilled condition in most LIC and MIC.

In Europe, the government regulates and controls the private sector, draws up contracts for the provision of health care, checks whether these contracts are implemented and oversees reimbursement. This situation is different from that found in LIC and MIC, where state mechanisms to control patronage, corruption and biased decisions at the administrative level are weak. Furthermore, government doctors, who are generally underpaid, expect to have a parallel private practice; as such they are unlikely to cut off the branch they are sitting on by agreeing contracts favourable to the state. Large amounts of money, as generated by contracting out, constitute a big temptation in the absence of adequate control.

Administrative and regulatory structures were rarely in place in the social sector when most LIC and MIC embarked upon neoliberal policies. This was a lethal situation, as “Failure to develop such capacity and political conditions before or simultaneously with entering into contracting and demand-side financing reforms can have negative consequences to judge from experiences in India, Mexico, Papua New Guinea, South Africa and Zimbabwe” [85]. “Contracting out clinical services is particularly complex, even when limited to non-profit providers such as church hospitals in Ghana or the United Republic of Tanzania and Zimbabwe” [85]. Only strong democratic governments with adequate regulatory resources can guarantee access and quality of care, avoid fragmentation of the system, enforce a solidarity-based financing system and keep NGOs and the private-for-profit sector under control.

Figueras and Saltman note that the reform of the medical and health sector in Europe called on public health skills to estimate the needs, evaluate the interventions and the impact of the measures [86]. As these skills are in short supply in LIC and MIC, Brugha and Zwi notice “major problems in service quality, especially in the private sector” and see the search for profit as responsible for the gap between health professionals' medical knowledge and its practice [87]. Competent professionals from LIC and MIC are numerous, but they are not in the right spot.

Discussion

Recently, a signal of policy change has been sent by the late WHO Director General Dr. Jong-wook, and by UNICEF [88]. In an effort to diversify international aid to the health sector of developing countries, these two organisations recommend reconstruction of health systems to increase access to general, appropriate health care in the services while at the same time developing disease control. The critical question is: how?

The commoditification and privatisation of services, with the public sector confined to deliver disease control activities alongside a private sector increasingly taking over health care, constrains both disease control programme performance and people's access to services. So why is neoliberal policy still actively promoted? Neoliberal health policies have aggravated the commodification and privatisation of health care, but to attribute such sweeping changes to the “imposition” of a doctrine by aid agencies would be to negate the complexity of this profoundly political process and the involvement of a set of actors that include “aid agencies”, the Bretton Woods Institutions, Western governments, the World Health Organization and developing country governments, which have either actively promoted or uncritically accepted such policies. Furthermore, the following vested interests may also contribute to this position:

The private medical sector in LIC and MIC which lobbies for access to public funds;

Local middle and upper classes which resist increased taxation to allow funding healthcare for the poor, since they can opt out for private health care;

International health care companies based in industrial countries looking for opportunities to access the Asian and Latin American markets, through the General Agreement on Trade in Services (GATS) negotiation rounds;

Pharmaceutical companies showing little interest in a public health care market that dispenses mostly generic and essential drugs and may not be very creditworthy. In contrast, disease control programmes financed by industrialised countries represent a market for the development of new products;

European politicians supporting the control of infectious epidemics in LIC and MIC that threaten the industrialised world (e.g. tuberculosis, AIDS, SARS, avian flu).

Disease control programmes enjoying high potential political visibility and hence support through mechanisms such as public–private partnerships;

Civil servants in charge of disease control programmes, enjoying privileged access to decision-makers in industrialised countries, see an emergent programme as a significant career opportunity.

These forces have systematically used recommendations of international aid agencies to promote national health policies in line with their interests. This explains the adoption of neoliberal health policies.

The GATS negotiations within the Word Trade Organization (WTO) [89] might further consolidate these results. GATS threatens to make illegal any public health care service that has not been privatised, based on the claim that the government should not offer subsidised services that the market also offers [90]. An opt-out alternative exists in theory, but LICs are unlikely to be in a strong bargaining position to use it through successive rounds of negotiation.

Conclusions

Neoliberal health policy promoted by international aid agencies in low and middle-income countries—state disease control programmes provided in services confined to deliver them and contracting out of curative care—failed to control diseases successfully, nor did it improve access to care. Its results were disappointingly out of step with those expected, its technicality out of step with the socio-political conditions and its motives with humanitarian objectives.

An alternative aid policy for LIC and MIC health sectors is badly needed if we genuinely aim to improve access to health care, controlling diseases and combating poverty.