International health policy and stagnating maternal mortality: is there a causal link?

Abstract

This paper examines why progress towards Millennium Development Goal 5 on maternal health appears to have stagnated in much of the global south. We contend that besides the widely recognised existence of weak health systems, including weak services, low staffing levels, managerial weaknesses, and lack of infrastructure and information, this stagnation relates to the inability of most countries to meet two essential conditions: to develop access to publicly funded, comprehensive health care, and to provide the not-for-profit sector with needed political, technical and financial support. This paper offers a critical perspective on the past 15 years of international health policies as a possible cofactor of high maternal mortality, because of their emphasis on disease control in public health services at the expense of access to comprehensive health care, and failures of contracting out and public–private partnerships in health care. Health care delivery cannot be an issue both of trade and of right. Without policies to make health systems in the global south more publicly-oriented and accountable, the current standards of maternal and child health care are likely to remain poor, and maternal deaths will continue to affect women and their families at an intolerably high level.

Résumé

Pourquoi les progrès vers l'OMD 5 relatif à la santé maternelle semblent-ils stagner dans la plupart des pays du Sud ? Les auteurs de l'article avancent qu'en plus des faiblesses largement reconnues des systèmes de santé, notamment les déficiences des services, l'insuffisante dotation en personnel, les lacunes de la gestion, ainsi que le manque d'infrastructure et d'information, cette stagnation est due à l'incapacité de la plupart des pays à rencontrer deux conditions essentielles : élargir l'accès à des soins de santé globaux et financés par l'État, et doter le secteur non lucratif d'un soutien politique, technique et financier cruellement nécessaire. L'article propose une perspective critique sur les politiques sanitaires internationales des 15 dernières années comme corrélat possible de la mortalité maternelle élevée, en raison de l'accent que ces politiques placent sur la lutte contre les maladies dans les services de santé publique, aux dépens de l'accès à des soins de santé globaux, et le manque de recours aux services extérieurs et aux partenariats public-privé dans la santé. Les soins de santé ne peuvent relever à la fois du commerce et du droit à la santé. Sans politiques qui orienteront les systèmes de santé du Sud vers une logique sociale et les rendront plus comptables de leurs activités, les normes actuelles des soins de santé maternelle et infantile risquent de rester médiocres, et les décès maternels continueront de toucher les femmes et leurs familles à un niveau intolérable.

Resumen

En este artículo se examinan las razones por las que los avances hacia el Objetivo 5 de Desarrollo del Milenio respecto a la salud materna parecen haberse estancado en gran parte del sur global. Argüimos que además de la existencia, ampliamente conocida de sistemas de salud débiles, con servicios deficientes, número reducido de personal, debilidades administrativas y falta de infraestructura e información, este estancamiento está relacionado con la incapacidad de la mayoría de los países para satisfacer dos condiciones esenciales: crear acceso a servicios de atención integral de la salud financiados por el sector público y brindar al sector sin fines de lucro el apoyo político, técnico y financiero que necesita. Este artículo ofrece un punto de vista crítico sobre los últimos 15 años de políticas internacionales de salud como un posible cofactor de las altas tasas de mortalidad materna, debido a su énfasis en el control de enfermedades en servicios de salud pública a expensas del acceso a la atención integral de la salud, así como a los fracasos de subcontratación y alianzas entre los sectores público y privado de salud. La prestación de atención de salud no puede ser un asunto tanto de comercio como de derecho. Sin políticas para lograr que los sistemas de salud del sur estén más orientados hacia el público y sean más responsables, los niveles actuales de atención materno-infantil probablemente continuarán siendo deficientes, y las muertes maternas continuarán afectando intolerablemente a las mujeres y sus familias.

Key Words:

• health policy and programmes
• health systems
• privatisation
• neoliberal economics
• maternal mortality and morbidity

In an address to the 61st World Health Assembly in May 2008, Dr Margaret Chan, Executive Director of the World Health Organization (WHO), acknowledged that progress towards reaching the Millennium Development Goals (MDGs) had stalled,¹ and particularly progress with improving maternal health (MDG 5), which she said would be “slow and uneven”.² This paper examines why progress towards MDG 5 appears to have stagnated in much of the global south.³

WHO, the World Bank and the European Union have long had a doctrine on aid and international health policy.⁴ This reflects the fears of industrialised countries of epidemics such as AIDS and bird flu in low-income and middle-income countries (LIC/MIC). It allocates disease control interventions to the public sector and comprehensive health care to the private sector.⁵ This is a neo-liberal doctrine, which includes:

•	the commodification of health care; and concurrently,
•	restricting public services to the delivery of disease control programmes, for which there is little demand;
•	maximising the transfer of public funds to private interests and securing outlets for privately manufactured goods;
•	the reduction of national health financing and public provision67 in favour of lowering direct taxes.

The primary health care strategy (WHO 1978) somehow receded into the background as priorities for international agencies after the United States withheld its contribution to the WHO regular budget in 1985 and the district policy (Harare Conference 1987) and Bamako Initiative (1985) proved to be short-lived.⁸ The World Bank then dismissed as irrelevant “the provision of comprehensive health care in public services”.⁹ Large sections of the academic community endorsed this new perspective,¹⁰ which generated a wealth of publications depicting public health care provision as inefficient, bureaucratic and unresponsive. These publications actively promoted the privatisation of health care via the purchaser–provider split, autonomous management of public hospitals, contracting out of services, private financing initiatives and managed care, dressed as “scientific guidance”. The Bretton Woods institutions began to condition their loans to LIC/MIC on the acceptance of limiting public health service delivery to disease control (labelled “prioritisation”). Bilateral aid agencies followed suit.¹¹ Coupled with the leveraging of aid and loan conditions, the World Trade Organization GATS negotiations began to enforce the privatisation of health care and open LIC/MIC health markets to (western) health care industries. The GATS agreement (Article 2.3.c)¹² prevents signatory Governments^* and international agencies from providing subsidised goods in the health sector for which there is market demand. These negotiations demonstrated (unwittingly) the basic commercial rationale of the preceding 15 years of international health policies.

Often mobilised outside the UN system, new alliances were set up with the private sector within international development assistance. Nearly all of them were disease-specific, public–private partnerships (PPP), named Global Health Initiatives. There were 79 of them in 2004, of which at least 20 were so-called product development partnerships (for vaccines, drugs, etc.).¹³ There are currently more than 100 that have initiated dozens of worldwide disease control programmes.^{1415 †} The proportion of development assistance disbursed through Global Health Initiatives has increased steeply over the last decade, as has total development assistance for the health sector as a whole, from just over US$ 6 billion in 1999 to US$16.7 billion in 2006.¹⁶ The proportion of development assistance to health channelled through development banks, private non-profit and transnational corporate philanthropy rose from about US$ 2 billion in 1999 to more than US$ 6 billion in 2003.¹⁷ Financing from partners such as the Bill and Melinda Gates Foundation, the Global Fund to Fight AIDS, Tuberculosis and Malaria, and the Global Alliance for Vaccines and Immunization accounted for almost 13% of development assistance to health by 2005.¹⁸ This displaced the share of development assistance to health for primary care, which declined from about 28% to 15% over the same period.¹⁷ The fastest growth was in funding for HIV/AIDS programmes from US$1.5 billion in 2002 to US$ 8.3 billion in 2006.¹⁶

While neo-liberal theories promote cost-effectiveness analysis as the scientific tool to identify disease priorities for public services, the international coalition established with “burden of disease” objectives managed to evade the need for real evidence based on how it defined its priorities.^19–21 For instance, cardiovascular diseases killed many more people than tuberculosis²² in LIC and MIC, according to Figure 1.8 in the World Health Report 2008, but has received little or no funds. The new aid configuration and Global Health Initiatives enabled multinational corporations to acquire a key role in determining health priorities and policies – mainly to liberalise trade.

Developing countries' health systems were severely segmented and fragmented. With an aid architecture based on disease-specific programmes, administrative and transaction costs soared as they fuelled an unprecedented bureaucratisation of Ministries of Health, NGOs and overall overseas development aid. The lack of coordination between disease-specific programmes and the resulting inefficiency is now largely recognised and sometimes even measured. For five neglected tropical diseases^* it was estimated that by integrating their respective control programmes (through preventive chemotherapy), a potential cost savings could be reached as high as 26–47%, or US $58–81 million annually, versus $110 million for the five stand-alone programmes.²³

From Zambia to Senegal, the bureaucratisation prompted by the expansion of disease programmes resulted in a dramatic internal brain drain: health professionals abandoned public health care and began working for these programmes.²⁴ For example, few Zambian doctors wanted to embark on a clinical career because the salaries paid by international disease programmes were much higher than those in the shrinking public and private sectors.²⁵

This brain drain also resulted in the under-utilisation and loss of doctors' skills; for example, a physician who becomes a district TB specialist is not required to use his/her knowledge of any other aspects of health (e.g. obstetrics). The loss of clinical skills among doctors, nurses and midwives has been aggravated by the weakening of academic clinical teaching and curricula and by the multiplication of commercial, unregulated and uncontrolled medical schools in universities in LIC/MIC in recent decades. These have been managed by a diverse set of authorities, funded mainly by student fees and with lax regulation.²⁶ Meanwhile, physicians graduating from top medical schools became the first to migrate to industrial countries.²⁷

In several countries in Latin America, doctors were encouraged to accept only part-time public health service employment. In order to compensate, many have been forced to pick and choose patients who come to the public health service and transfer the most profitable ones to their private clinics.^28–30 Such actions have reinforced the privatisation of health care, while paradoxically securing the existence of public services, which benefit from the political support of doctors' professional associations.

Is continuing high maternal mortality an exception or a sign of failure of the development aid paradigm?

Maternal health services have been among the most affected by the decimation of the public health sector³¹ because of their sensitivity to the overall functioning of the health system.³²³³ Today, maternal death and disability remain the leading cause of healthy life-years lost for women of reproductive age in low-income countries. In 1996, 585,000 women died of maternal causes, 99% of them in developing countries.³⁴ In 2005, 536,000 women were still dying of maternal causes, 99% in the self-same countries.³⁵ The worst numbers were in sub-Saharan Africa, where coverage for maternal health had begun stagnating in the 1990s.³⁶

The failure of progress on maternal health has been toned down by donors and national governments with claims of “uneven” achievement. Indeed, this has happened with other disease-specific programmes too. In spite of the “success story” of putting so many people with AIDS on highly active antiretroviral therapy (HAART), “success” needs to be interpreted with caution. MDG6 aims at controlling incidence, not prevalence of HIV. Thus, although the prevalence of HIV in Zimbabwe, for example, fell by more than 10% in the past decade,³⁷ in spite of political and social malaise,³⁸ there are three new HIV cases to every person put on HAART,³⁹ a continuing high incidence of new infections and very high mortality.

Why international policies have failed to reduce epidemics or make progress on maternal health

The share of total development assistance for reproductive health care and family planning (in all sectors) remained around 1% between 1993 and 2003. The share of development assistance related to health overall diminished from 22% to 16% over the same period. While the proportion of the 1% of development assistance intended for maternal and child health almost doubled from 8% in 2003 to 15% in 2006,⁴⁰ overall maternal health budgets remained substandard. India, for instance, still spends less then one US dollar per woman of reproductive age per year, according to a 2007 study.⁴¹

In 2001, the WHO Commission on Macroeconomics and Health estimated the annual incremental costs for treatment of maternity-related conditions at US$ 4 billion, a figure they said would double by 2015.⁴² Typical OECD donors, such as DFID and USAID, only spend an estimated 3% of their budgets on maternal and child health combined, or US$1.75 billion in 2003, representing a shortfall of around US$ 6 billion.⁴³ Worldwide official development aid to maternal and newborn health rose from US$ 0.7 billion in 2003 to as high as US$ 1.2 billion in 2006, but only a proportion of this is for maternal health as opposed to newborns, and the total still represents a dramatic shortfall,⁴⁰⁴³ much more than the US$ 0.06 billion that the Global Fund for HIV/AIDS, Tuberculosis and Malaria provided for maternal and child health at its height in 2006. By contrast, AIDS-specific expenditures grew spectacularly, from “millions to billions”.⁴⁴

Compared to AIDS and TB programmes, maternal health programmes have simply been unable to mobilise LIC/MIC public services to meet the MDG 5 targets, despite numerous international conferences and initiatives. There are few (pharmaceutical) magic bullets for maternal deaths. Magnesium sulphate, oxytocin and misoprostol are off patent, which deters pharmaceutical companies from investing in them. Furthermore, maternal complications do not present an epidemic-type threat, which we think reduces their priority in the eyes of industrialised countries. Long-term public funding remains grossly insufficient.⁴⁵ It has been argued that US$ 296 was being spent worldwide on every avoidable maternal death in 2005²¹ by the Global Fund for AIDS, TB and Malaria, the Gates Foundation, World Bank and US government, but the way the money has been spent, combined with health policy and health systems factors and high levels of household poverty and malnutrition in most low-income countries⁴⁶ have served to block progress.

Health systems factors: segmentation and fragmentation

Although the costs of maternal and child health (MCH) programmes have not skyrocketed, as others have, they do generate opportunity costs. Many funding agencies have set up an unintended competition between maternal and infant and child health and between skilled facility-based care and community care. Inadequate funding has been allocated to health infrastructure, training programmes and management.⁴⁷ However, achieving maternal health requires a great deal from health systems, in particular the provision of comprehensive care,⁴⁸ including effective health centres for antenatal care, treatment of complications and rapid referral; peripheral maternity units (e.g. for vacuum extraction at delivery); district hospitals (e.g. for caesarean sections), regional hospitals (e.g. for hysterectomy) and teaching hospitals (e.g. for hydatidiform mole). Sharing resources across facility boundaries for the sake of economies of scale is needed, as well as mutual support between providers, e.g. regional surgeons should train and supervise district GPs to perform caesarean sections to make them more accessible.

Over the past two decades, however, health systems have largely collapsed in most low-income countries and fragile states (altogether containing some 2.5 billion people), most of the same countries where international donor assistance has been most active. Thus, we ask: has the nature of aid and the way it has been disbursed contributed to this collapse.⁴⁹ As stated by the WHO Commission on Social Determinants of Health in 2008:

“There is… a danger that large new funding lines, running parallel to national budgeting, continue to distort national priorities for allocation of expenditure and action… While Global Health Initiatives have brought enormous new levels of funding to health-care systems within low- and middle-income countries (US$ 8.9 billion in 2006 for HIV/AIDS alone), there is a concern that their vertically managed programs have the potential to undermine the population health orientation of health-care systems and as a result exacerbate health inequity.” ⁵⁰

However, the WHO approach to strengthening health systems in low-income countries of 2007⁵¹ overlooked key systems factors,⁵² including the following:

•	Public hospitals are being starved of resources. In core services such as maternal health care;53 this has had negative implications for equity and access.54 Badly paid health professionals who are left to manage public facilities with an economic rationale tend to introduce a commercial logic to service delivery in order to maximise income.5556 Autonomously managed government hospitals in low-income settings tend to admit a middle-class child with diarrhoea rather than a township woman in need of a caesarean section, to keep costs down.32
•	Decentralisation under the form of devolution promoted by the World Bank and International Monetary Fund, without relevant evidence to support their approach, segmented the public sector into local (municipal), provincial and national facilities, with other facilities run by NGOs, which has largely had a negative effect.5758 On the other hand, cities such as Mexico City have used the devolved financial room to improve their health systems,59 but sometimes in isolation from the rest of the country.
•	Managed care (e.g. pay-for-performance, management contracts or fixed capitation fees60) was promoted in LIC/MIC to limit health care costs. However, there is growing evidence that the problems generated by these policies overwhelm the intended benefits with regard to quality of care or cost, and often both.61
•	The administration of scores of disease-specific programmes opened up avenues for the practice of “vote-catching” and nepotism.62–64
•	The centralisation of aid structures does not permit morbidity-specific epidemiological problems (e.g. the high incidence of eclampsia in Mauritania and other West African countries) to receive funds that match the extent of the problem.65
•	The reduced status of public sector health professionals is a direct consequence of structural adjustment programmes and reductions in overall social sector spending.66
•	Low-income countries have systematically failed to reduce the concentration of health staff in cities and re-deploy them to rural and urban peripheral areas; and staff resist because they are underpaid (due to structural adjustment), and career paths are often impeded.67–70
•	Since the poor rarely have access to effective legal protection or to the policy-making process,71 social welfare expenditure has been kept to a minimum and the financing of the health sector in most LICs has largely been handed over to overseas development assistance, especially in sub-Saharan Africa.

Health systems may lack responsiveness to people's need for treatment and relief of suffering because international policies no longer allow public services to play a role in comprehensive care delivery. At the same time, access to the private sector has not improved in LIC/MIC, even with State financing, because of high costs and a low capacity for regulation and control that is becoming increasingly widespread. If contracting out still guarantees access to good quality health care in many Western European countries, although on a downward slope, why is it failing in LIC/MIC? We would argue that the technical requirements for contracting out clash with political and administrative realities on the ground.⁵ Efficient sub-contracting requires stringent control and regulation, which is difficult to fulfil in most LIC/MIC settings. The World Health Report 2000 recognised that failure to develop such capacity when entering into contracting out and demand-side financing reforms can have negative consequences, to judge from experiences in India, Mexico, Papua New Guinea, South Africa and Zimbabwe.⁷²

Managerial factors

Verticalisation of services has particularly affected maternal health care.⁷³ International programmes may not, in general, be truly vertical, as their hierarchy and administration are not integrated, even though their operations are.⁷⁴ Thereby, they have overlooked the need for unity in the chain of command, one of Henri Fayol's fourteen principles of management and a key one, generally applied in the corporate sector.⁷⁵ When each dispensary has to respond to several chiefs (e.g. distinct disease and MCH programme managers) who compete with each other for public resources – in particular to monopolise the working time of nurses and midwives – it is damaging to the health system and the services in practice.

Structural disintegration of comprehensive care and disease control

International agencies have promoted the involvement of governments, NGOs and communities in disease control programmes while continuing to privatise health care. This two-track policy has systematically prevented integration. This is a problem because curative and preventive health activities need to be integrated – into comprehensive and accessible health care.⁵

Most disease control interventions are clinical. To be effective, they require services to be used by a sufficient number of patients and achieve early detection and treatment. In theory, both public and private sectors could provide integrated disease control. Yet, with the exception of tuberculosis, international agencies have been reluctant to allocate disease control to the private for-profit sector. It has been shown with malaria⁷⁶ that to be effective, disease control programmes need to be integrated in health facilities where there is a pool of patients, a principle valid for both acute and chronic conditions. This is also applicable to maternal health, e.g. effective eclampsia management requires attendance at antenatal care, ease of consultation for symptoms, appropriate treatment and referral, and attendance at a referral facility. The structural disintegration between curative and preventive maternal health care is why, in practice, we see how difficult it is for women to switch easily from antenatal care to curative care, as most women in need of antenatal care cannot be recruited through consultations for curative care – because so few of them utilise such care in pregnancy and because they are disintegrated.

Second, the boundaries between secondary prevention and curative care are not fixed, nor are those between curative care and tertiary prevention. Thus, antenatal care encompasses primary prevention (e.g. malaria prophylaxis and prevention of mother-to-child transmission of HIV), secondary prevention (e.g. early detection and treatment of pre-eclampsia) and curative care (e.g. treatment for anaemia, STIs, AIDS, urinary tract infection and malaria). Finally, patients should be able to access appropriate medical care when they need it. Therefore, referrals from curative to preventive care and vice-versa must not meet obstacles, whether related to transport, fees or drug availability.

Failures of contracting out and public–private partnerships

The need for better studies on the feasibility and impact of contracting out maternal health care is acknowledged. Peters et al⁷⁷ evaluated some 700 studies of contracting-out schemes for sexual and reproductive health services carried out between 1980 and 2003 in different regions (including maternity care, abortion, prevention or treatment of STIs and family planning). Although contracts with private sector agencies were used most frequently for maternity services (in 55% of studies analysed), these were in most cases limited to the training of traditional birth attendants (TBAs). The authors concluded that most studies left key questions unanswered on the feasibility and impact of such private sector strategies and evidence of their effectiveness was found to be weak.

Privatised training of TBAs for skilled birth attendance^78–80 has had a limited impact on maternal outcomes,⁸¹ so there is currently a focus on improving professional services. India was one of the first countries to gain experience with public–private partnerships for skilled birth attendance.⁸² But Oxfam warned: “In India, 82% of outpatient care is provided by the private sector. The number of first class private hospitals is rapidly increasing. Yet this same system denies half the mothers in India any medical assistance during childbirth”.⁸³ Yet public–private partnerships in reproductive and maternal health care continue to be advocated, especially in Asia.⁷⁷

Without evidence, it is possible to formulate hypotheses about other negative consequences of contracting out obstetric care. In theory, the separation of health care delivery and disease control programmes could have been avoided by contracting out both comprehensive care delivery and disease control interventions. In practice, this has rarely been done, except with TB control in South Asia (based on one positive pilot in India⁸⁴) and maternal and child health services in Asia, but evidence is lacking as will be seen below.

Almost all LICs have starved their public services of resources (without funding their private sectors) and not surprisingly, patients have experienced a reduction in access to public health services over the past two decades, reinforced by critical shortages of health workers⁸⁵ and shortfalls in overall performance.⁸⁶ Only a few MICs have successfully managed to finance the contracting out of some health care and disease control, usually with international loans.⁸⁷

What have international agencies said in their analyses of these experiences? The International Finance Corporation acknowledged that lower-income and rural populations have been most profoundly affected in sub-Saharan Africa by the failings of private health care.⁸⁸ A World Bank report suggested that government services generally perform far better than the private sector for rich and poor women alike.⁸⁹ Based on an extensive literature review of efficiency in contracting out, the WHO recognised that “the conditions necessary for competition… were generally absent from most areas of most low- and middle-income countries.⁹⁰ More recently, Oxfam noted: “Lebanon has one of the most privatized health systems in the developing world. It spends more than twice as much as Sri Lanka on health care yet its infant and maternal mortality rates are two and a half and three times higher respectively”.⁸³

Contracting out maternal health services on a fee-for-service or lump-sum payment basis for individual clinicians in developing countries in a context of weak stewardship is also to be feared. Here are two examples of what has happened:

•	In a fee-for-service system, health professionals unduly increase the indications for procedures such as caesarean sections, especially among more affluent women. For example, Demographic and Health Survey data show that private hospitals in Brazil (36% to 72%),91 México (52%) and Colombia (59%) have the highest caesarean section rates in the world.92
•	Conversely, when a system of giving a lump-sum payment for 100 deliveries was instituted in a scheme in Gujarat, India, the caesarean section rate dropped disproportionately among some segments of the population, mainly due to private physicians refusing to perform them at the given rate and referring patients to the public health service (Werner Soors, personal communication).

The professional private sector is absent from LIC/MIC rural areas and shantytowns, and access to privately delivered multifunction curative health care financed by user charges in LIC is severely restricted.^93–97 Although depicted by numerous publications as offering flexible access with shorter waiting times and greater confidentiality,⁹⁸⁹⁹ private practitioners remain out of reach for many; only higher income groups are able to afford them.¹⁰⁰

Colombia since 1993 has been distinguished by a decentralised system of social/health security; contracting out of health care; and for those who can afford it, a complete package of health interventions in a “contributory system”. This regime should have covered the rest of the population with a minimal package of interventions – however it failed to do so. In contrast, Costa Rica did not privatise health financing but maintained a public delivery system alongside the private one. It has a dominant public health system and a monopoly over public health insurances. It spends nine times less on health per capita than the USA yet manages to secure a similar life expectancy. Its remarkable output and outcomes in the health sector may be linked to its health service organisation and the way it is financed. Tellingly, in 2005, Costa Rica had a maternal mortality ratio of 30 maternal deaths per 100,000 live births, four times lower than Colombia's at 130.

In fact, both quantitative and qualitative studies provide evidence of the failure of contracting out of obstetric care. While consuming a lot of resources,¹⁰¹ privatisation and medicalisation of childbirth (with or without public funding) has not improved key obstetric indicators. In 1999, the UN General Assembly agreed on a global target of 80% coverage of skilled birth attendance by 2005, 85% by 2010 and 90% by 2015.¹⁰² In 2008, the coverage was estimated at 46.9% in south-central Asia and 33.7% in Eastern Africa, both falling compared to 2005, not improving.¹⁰³ Furthermore, the proportion of skilled birth attendance may well be less than is officially recognised.¹⁰⁴ Lastly, unlike with not-for-profit, private organisations,¹⁰⁵ attempts to get private practitioners to comply with national guidelines in LIC/MIC have often failed,^106,107 and they are sometimes reluctant to refer their patients to public facilities in a timely way, even with serious conditions.¹⁰⁸ We conclude that when it comes to training skilled birth attendants in LIC/MIC, the commercial sector has shown itself unable to substitute even for ailing public services.

Figueras and Saltman note that the reform of the medical and health sector(s) in Europe called upon public health skills to estimate the needs, evaluate the interventions and the impact of the measures proposed.¹⁰⁹ These skills are in short supply, including for maternity care, in many LIC and MIC. Brugha and Zwi point out “major problems in service quality, especially in the private sector” and see the search for profit as responsible for the gap between the medical knowledge of health professionals and their practice.¹⁰⁶

We therefore call for policies that are adapted to the environment of low- and middle-income countries. They will necessitate strong monitoring and regulation, independence from donors and most of all political commitment to the population at large. Neoliberal health policies (particularly contracting out) appear to have failed to deliver in most LIC/MIC settings in the past two decades. Local experts need to acknowledge these failings and the price that has been paid by the millions of women who have died avoidable maternal deaths.

If the prospects for alternative international health policies are gloomy, the UN should take maternal mortality out of the MDGs, as they did for access to care, adult male mortality and chronic degenerative diseases.¹¹⁰ However, that would be unacceptable to the women's health community. On the other hand, there are a wealth of interesting, heterodox experiences and local initiatives in the developing world that have worked in providing access to maternity care for poor women, but political awareness at international level (to acknowledge failings) and State democratisation and commitment will remain pivotal to overcome the challenges. The rest of this paper will address these.

Publicly-oriented health policies and services to improve maternal health

Many lessons can be learned from countries that did not adopt the dominant international health policies but have successfully experimented with the delivery of health care with a social mission – e.g. Costa Rica, Cuba, Kerala State (India), Sri Lanka, Lesotho, Brazil, Sweden, Spain, Jordan and Bahrain. Criteria defining medical practice with a social mission and health management premised on a social rationale exist.^105,111,112

All health facilities with a social mission should be coordinated within the publicly-oriented sector. A broad spectrum of maternal health care services, however, must be divided between primary or health centre level, including maternities attached to health centres for normal deliveries, and district and regional hospitals to handle complications that cannot be dealt with at primary level. The backbone is the development of district health teams that are defragmented. Integrated local health systems or districts are centred around one hospital¹¹³ and gather also the not-for-profit maternities and health centres belonging to municipalities, regions, Ministry of Health, churches, social security, NGO and community mutual aid associations. Their efforts should aim at making maternal health care (curative and preventive) attractive to women and their families, answering to both their needs and demands, overcoming services segmentation and providing them with technical support and evaluation.

The aim of a publicly-oriented health system (even operating alongside a commercial one) is to secure universal access to comprehensive care (initiated both by a health worker or needs-based, as well as patient or demand-based), which includes guaranteed access (depending upon context and resources) to a district hospital capable of tackling some medical and surgical obstetric emergencies, and to a regional hospital, for others. Crucially, district hospitals require far more funding and appropriate technology¹¹⁴ than what they receive today.

Such comprehensive health care aims at balancing the concerns of the patient, the community, the state and health professionals. These criteria can be used in contracts and to inform the training and evaluation of the delivery of care and management. These criteria are fundamentally opposed to for-profit medical practice and commercial health management, which focus on financial profitability, treat corporate and health professionals' income as ends in themselves and emphasise maximising the financial returns from professional practice. With a profit motive, there is not one medicine or one management science for health but two, according to whether a profit will be pursued or not.

Instead of classifying “priority” conditions to determine “priority” disease control programmes (on the basis of DALYs, for instance) and structure publicly-oriented services accordingly, the “package of activities” may be better defined on the basis of the full use at primary level of nurses', midwives' and GPs' skills. It would then be delineated according to what these health workers can do and what they can learn through in-service training, technical/psychological coaching and practical rotations in a referral hospital. A clinical practice improvement plan can be used to secure knowledge transfer in all districts, building on expertise available locally (e.g. in the hospitals, while recognising that obstetrician–gynaecologists themselves may have to be retrained in the handling of surgical, obstetric and other medical emergencies.

The Immpact project¹¹⁵ is one of a number of examples of an evidence-based global research initiative, started in 2003, that has taken a fresh look at robust evidence for safe motherhood strategies. It has produced excellent standards for protecting access to essential and emergency obstetric care for the poorest women and households, working on financial, physical and functional barriers together. Delivery fee exemption policies; village-based, resident, trained midwives; and improved health insurance coverage were combined, with a special interest in getting strategies out to the poorest as well as other women, so as not to inadvertently widen the gap between rich and poor. The combination of measures developed has resulted in a sustained increase in use of services and institutional deliveries over more than four years in Ghana, Burkina Faso and Indonesia.

Conclusion

Health care delivery cannot be an issue both of trade and of right. These “alternatives” are incompatible from many points of view – medically, politically, socially, ethically and economically. In low-and middle-income countries, health care as a right¹¹⁶ is an issue of development and of political and economical stability. Without policies to make health systems in the global south more publicly-oriented and accountable, the current standards of maternal and child health care are likely to remain poor, and maternal deaths will continue to affect women and their families at an intolerably high level.

Young mothers wait at a pregnancy clinic, S Corner Project, Kingston, Jamaica

Notes

* Some 52 countries have already signed the GATS agreement, affecting their health sectors.

† For instance, the Global Fund to Fight AIDS, TB and Malaria; European Malaria Vaccine Initiative; Global Alliance for Vaccines and Immunization; Global Business Coalition on HIV/AIDS; Global Public—Private Partnership for Hand Washing with Soap, etc.

* Lymphatic filariasis, onchocerciasis, intestinal helminthiasis, schistosomiasis and trachoma.