Safeguarding the Health Sector in Times of Macroeconomic Instability
This page intentionally left blank
Policy Lessons for Low- and Middle-Income Countries
Copyright © 2008 International Development Research Centre (IDRC)
First Printing 2008
Jointly Published by
AFRICA WORLD PRESS
P.O. Box 1892, Trenton, New Jersey 08607
awprsp@verizon.net/www.africaworldpressbooks.com
and the
International Development Research Centre
PO Box 8500, Ottawa, ON Canada KIG 3149
info@idrc/www.idrc.ca
ISBN (e-book): 978-1-55250-370-6
The findings, interpretations and conclusions expressed in this study are entirely those of the authors and should not be attributed in any manner to the institutions with which they are, or have been, affiliated or employed.
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means electronic, mechanical, photocopying, recording or otherwise without the prior written permission of the publisher.
Book and cover design: Saverance Publishing Services
Library of Congress Cataloging-in-Publication Data
Safeguarding the health sector in times of macroeconomic instability : policy lessons for low- and middle-income countries/editors: Slim Haddad, Enis Baris, Delampady Narayana.
p. ; cm.
Includes bibliographical references and index.
ISBN-10: 1-59221-596-3 (hardcover)
ISBN-10: 1-59221-597-1 (pbk.)
1. Health care reform--Economic aspects--Developing countries. 2. Medical policy--Economic aspects--Developing countries. 3. Public policy--Developing countries--Finance. 4. Structural adjustment (Economic policy)--Developing countries. 5. Economic stabilization--Developing countries. 6. Medical economics--Developing countries. I. Haddad, Slim. II. Baris, Enis. III. Narayana, D.
[DNLM: 1. Health Care Reform--economics. 2. Developing Countries. 3. Public
Policy. 4. Socioeconomic Factors. WA 530.1 S128 2007]
RA395.D44S34 2007
362.109172'4--dc22
2007028214
In memory of our friend and esteemed colleague
Stephen Chandiwana,
an ardent advocate for health policy and systems research
as a key ingredient in improving the health of sub-Saharan
African communities.
This page intentionally left blank
Structural adjustment, macroeconomic adjustment, and health sector reform are loaded terms. Each has become a kind of shorthand for diverse and frequently opposing health and development constituencies, signalling for each constituency a very different constellation of goals, assumptions, values, ideologies, methodologies, technical content and results. At one fundamental level, these differences reflect very different visions of how societies should function. Recent and ongoing debates about globalization, market-based approaches to service delivery, and roles of the state and active civil society are helping to sharpen these alternative visions and continue to play an important role in unpacking the contents of such seemingly bland terms as macroeconomic adjustment and health sector reform. At another level, the debates and growing evidence base around globalization highlight how entangled national and local policy and practice are with decisions and political and economic imperatives elsewhere. Yet at a third level, the level on the ground where macroeconomic policies and realities, health policies, health-care delivery systems, and people intersect, we see a much more complex and subtle situation unfolding.
At this third level, contexts, policy choices, personalities, policy-implementation capabilities, cultural and social norms, and political and economic history interact to produce outcomes that are both promising and troubling. It is this reality on the ground that is too often neglected or avoided by sweeping triumphalist or catastrophist narratives, the reality where governments must make choices in circumstances they may ardently wish were different, where ministries and clinics alike are short of skilled staff, where sudden shifts in commodity prices or weather may precipitate a crisis with which a health department must somehow deal, where both noble and venal motives and actions are in evidence, but where there may just be ways to safeguard the health sector while the higher-level debates about politics and vision continue both in-country and around the world. It is this third level which the researchers of the MAPHealth study so rigorously and thoughtfully address, and the result is an exceptional collection of analyses, reflections, and recommendations. Some of the results confirm prior findings, while others are surprising and on the surface may be counterintuitive. The conclusions and recommendations within each case study, and of the overall comparative work, make for sobering but also hopeful reading. There are no easy answers. The continued importance of political will points to the critical role that broad-based public debate and real—not pro forma—civic engagement play in informing and shaping power and politics, in and between societies, to the benefit of all, especially the poor. "Governance" is indeed political, despite what the more technocratic among us might wish. But as all the case studies show, action to safeguard the health sector is possible, even in countries as poor as Burkina Faso, as crisis-ridden during the period examined here as Zimbabwe, Mexico, Colombia, and Thailand, or as complex as India. The vision is crucial but the devil, as always, is in the details—hence the importance of the careful, systematic work presented here.
This book and other studies support the view that some kinds of imposed policy choices, such as the first generation of structural adjustment, leave very little room for action or mitigation at the country level. However, neither the study design nor the analysis presented here are aimed specifically at these macro-level political and philosophical debates, though the reader will find frequent, thoughtful, and challenging contributions to these debates throughout the chapters. The debates are longstanding, with modernization and dependency theory still being played out under new names as the pendulum swings back and forth in global development policy.
It is perhaps worth noting, in the foreword to a book which reports on many years of careful effort to collect and interpret evidence of policy impact, how seldom these sweeping and enthusiastic policy shifts are actually informed by careful ex ante analysis, or even concurrent rigorous evaluation of implementation and impact. This is one admonition the book's authors delicately avoid making, but which needs to made. Fortunately, recent initiatives around the world reflect a growing impatience with making the developing world a kind of global social policy laboratory, minus the careful lab science. We hope the methodological challenges experienced by these and other researchers, discussed in Chapter 3, will become less of an obstacle to future researchers and decision-makers.
As this book goes to press, the WHO Commission on Social Determinants of Health is winding up its deliberations—a fortuitous marker for the end of this chapter of the work described in this book, which also coincided with the work of the WHO Commission on Macroeconomics and Health. While the latter Commission played a seminal role in shifting the world's attention to the importance of health not only as an outcome of development and economic growth, but also as an essential input, our growing understanding of the social determinants of health equips us to address better some of the challenging findings and recommendations of the MAPHealth study. Research on population health and the social determinants of health has repeatedly demonstrated that social and income inequalities translate into health and mortality inequalities along the full gradient of inequality (and not just between the ends), as biological and social processes interact and are embodied in vulnerability or resilience to illness and death. However, recent data, notably from Canada, also demonstrate that deliberate social policies to transfer wealth and protect broad-based social benefits can significantly, or even entirely, flatten this gradient effect. This is good news because it suggests that, even in the absence of the more complete social and economic equity to which many of us are committed, there is room for action and significant gains in health. This news is echoed by several of the chapters in this book, though tempered at times with the sobering reality that good intentions—or even individually good policies—are no guarantee of good outcomes.
The MAPHealth team put into practice a key desideratum of current health policy and global health research thinking: that country work should be led as much as possible by national teams and that both Southern and Northern research partners should collaborate in the intellectual effort of global policy research and analysis. The collegial and collaborative spirit among all the country teams and with the Université de Montreal research team is in evidence throughout the book. The International Development Research Centre is privileged to have been able to support and participate in this important work.
This is an important and challenging book. If you are looking for a broad, high-level diagnosis or an easy-to-swallow generic prescription to immunize health systems against macroeconomic instability or crisis, this book is not for you. If, however, you are ready to tackle the complexities of understanding real health systems, real policies, and real politics in six very different countries facing similar constraints, this analysis of a collaborative, comparative, multi-year initiative is an invaluable contribution.
Christina Zarowsky, MD, PhD
Research for Health Equity
International Development Research Centre
The Editors and the members of the MAPHealth team extend their heartfelt thanks to all those persons who provided invaluable support in the preparation and execution of this research program across the several countries involved. The Editors also gratefully acknowledge the important contribution of the International Development Research Centre (IDRC), which provided both technical and financial support for the studies, and offer our warmest appreciation to Randy Spence, Christina Zarowsky, Marie-Claude Martin, Rohinton Medhora, Rosemary Kennedy, and Marta Melesse, who were a constant source of support over the years and without whom this work would not have been possible. Special thanks go to Donna Riley, our editor, translator and overall coordinator for her invaluable assistance in the production of this work. Publication of this book was also made possible in part by the Quebec Population Health Research Network.
This page intentionally left blank
Health sector reform has been a recurring theme over the past two decades in many low- and middle-income countries, especially in the broader context of macroeconomic instability. Although countries have adopted both macroeconomic and health policies to stimulate growth and improve health system efficiency, implementations and outcomes have varied considerably depending upon several factors: the origin, extent, and duration of the instability; the relative importance of economic, social, and political factors at play; and the scope and severity of sectoral challenges in health and health care. These policies, termed macroeconomic adjustment policies (MAPs), often involved adopting short-term austerity measures and longer-term structural and sectoral adjustment programs with financial assistance and policy guidance from international finance institutions (IFIs). MAPs have since been the subject of intense debate and remain controversial not only in terms of the scope and nature of policy dialogues, interventions, and implementation processes, but also because of their purported effects on the overall economy, on the poorest and most vulnerable households, and on the financing and provision of social services such as education and health.
Despite the existence of several studies, the outcomes of this unique natural experiment also remain controversial. Data are often unreliable and results vary widely. This is not surprising, given the great diversity of policies that vary from one country to another in terms of content, severity, implementation, and, of course, the contexts within which they are applied. Divergence of opinions is also to be expected for reasons ranging from ideology and underlying values to more credible ontological and epistemological explanations. More specifically, as described in Chapter 2, these studies differed not only in scope and purpose, but also in their design and methodology. Here it should simply be noted that the usual problems of validity and inference in attributing causes and effects proved to be more daunting in the health sector, where the reasons for—and the timing, sequencing, and degree of implementation of—sectoral reforms seldom lend themselves to a thorough and rigorous evaluation.
Not surprisingly, even if the consequences of MAPs on the overall economy in developing countries are relatively well documented, their impacts on the health sector proper, and more specifically on access to health care, utilization, and quality—and indeed on health outcomes—are still debated. Many policy analysts, researchers, and activists claim that the health sector in adjusting countries has been seriously undermined either as a direct result of MAPs or due to the deterioration of socioeconomic conditions associated with their stabilization measures. Following UNICEF's 1987 report Adjustment with a Human Face, many studies demonstrated a disturbing deterioration in health conditions in countries facing stringent adjustment policies, generating an intense debate regarding the real origin of this decline. Some ascribed it to the adjustment itself, and others, notably the IFIs, attributed it to the original crises that led to the need for adjustment, or else to an implementation of MAPs that was too timid or inappropriate. In this way, with examples at the ready, each side was able to delve into the diversity of experiences and the heterogeneity of situations in the countries under adjustment to support their arguments either that the adjustments were intrinsically deficient or that the original economic crisis was to blame, or to label those countries "poor adjusters."
Given that these adjustment policies were largely recycled within the context of recently resuscitated strategies in the war against poverty, our proposal is not, in this text, to re-examine these arguments and the value of the premises underlying them.
It must be noted that in the great majority of cases, analysis of the effects of MAPs on the health sector was primarily focused on one specific consequence—health conditions—and on one specific mechanism, i.e., deterioration of households' living conditions as the agency by which adjustment policies led to decline in the conditions of health. This perspective was justifiable in terms of studying the social consequences of adjustment and their impacts on health. However, this approach also served to distract attention away from the influence of MAPs and socioeconomic context on the health-care system itself, and on its performance in terms of its principal functions of regulation, financing, and provision of services.
Conversely, these questions have been at the heart of research into the effects of sectoral reform undertaken over the past 15 years by the Southern countries. However, the tendency has been to observe these transformations through a primarily sectoral lens, which does not sufficiently take into account the socioeconomic context within which health sector reform takes place, nor the influence and roles of various non-sectoral stake-holders in its design and implementation.
Whether integral to or concomitant with MAPs, health sector reform initiatives often arise from sectoral needs to improve the health status of the population and to improve system performance. Yet the timing, extent, and scope of health sector reform vary across countries. Therefore, a thorough impact assessment of health-care outcomes—such as accessibility, utilization, and quality of care—would require not only examining relationships between the reform parameters and the outcomes, but also carrying out a comparative assessment of the macroeconomic and sectoral contexts within which they occur in various countries. With this information, we would be better able to draw lessons from both failures and successes.
The research initiative that led to this book was designed to provide a comparative ex-post evaluation of health sector reform initiatives in middle- and low-income countries so as to document which measures, and which policy environments, are more likely to be successful in safeguarding the health sector. This evaluation is intended to offer some perspective on how and why changes in the health-care system manifested in the ways they did in different countries under similar adjustment policies. It also provides evidence to help in defining the appropriate mix of policy instruments and adaptations to be made according to the particularities of each country. As such, this book provides a fresh look at the relationships between public policies and the health system in general, and presents new and detailed observations to inform policy choices within the health-care sector worldwide.
The results presented here come out of studies carried out by researchers in Africa, Latin America, Asia, and Canada within the framework of a research initiative entitled "Macroeconomic and adjustment policies and health care: studying the macro-micro links," also referred to as the MAPHealth project. This initiative was developed through a cooperative network of eight participating countries, the International Development Research Centre (IDRC), which initiated and funded the project, and the Groupe de recherche interdisciplinaire en santé (GRIS) at the Université de Montréal.
The MAPHealth project examined the evolution of policies and health services in eight Southern countries (of which six are presented in this book) in light of transformations in their macroeconomic environments, and especially with regard to the crises and sectoral adjustment policies they faced, i.e., how health systems responded, at different times and in different contexts, to major transformations of their environments. The countries associated with this IDRC initiative were Burkina Faso, Colombia, India, Kenya, Mexico, Thailand, Uganda, and Zimbabwe. Given that the majority of these countries had begun or carried out, in the previous 20 years, one or more health system reforms, particular attention was given to analyzing the links between these reforms, macroadjustment policies, and the services provided.
Chapter 2 surveys the literature for two key areas of study—the effects of MAPs on health outcomes and the effects of health sector reform on system performance. The impacts of MAPs are considered from both the demand side (e.g. health, nutritional status, education) and the supply side (e.g. health expenditures, access to good quality care). The evidence on the effects of health sector reforms is reviewed for three core measures: financing, decentralization, and integrating competition into the health sector. The literature, while significant, indicates a need for stronger evidence on the impacts of reforms driven by macroeconomic instability, as well as for a better understanding of how health systems are influenced by various factors within and external to the health sector.
Chapter 3 focuses on methods. In that chapter, we present briefly the main features of multiple-country and single-country studies used to assess the effects of reforms and adjustment policies in the health sector. We examine the challenges faced in evaluating outcomes of large-scale policies, with special attention to issues related to the assessment of counterfactual situations, the attribution of observed changes to specific measures or global policies, and the choice of appropriate designs and measurement tools. We then describe the methodology of the MAPHealth project, which builds upon our efforts to draw lessons in methodology from studies carried out over the past 15 years on the effects on health of structural adjustment.
The experiences of the six countries, summarized in the following paragraphs, are presented in greater detail in chapters 4 through 10 of this book. Note that, in India, the experience of the state of Kerala was found to be sufficiently informative that it has been described in a chapter of its own.
To move from a state-dominated to a market-based economy, the government of Burkina Faso decided, in 1991, to implement a structural adjustment program (SAP) with the aid of the International Monetary Fund (IMF) and the World Bank. This involved stabilization measures as well as structural and sectoral reforms. The most significant measure was the deep devaluation the CFA franc BCEAO (XOF) in 1994 to one-half its value.
Health sector reform was one of the sectoral reforms within the adjustment package. It focused on decentralization and implementation of cost-recovery mechanisms, the supply of essential generic drugs, the liberalization of health services, and the hospital sector. Laws and regulations were passed giving public facilities a great deal of management autonomy and establishing the principle of cost recovery through fees for services, and particularly for essential drugs. Within a few years, fees for services were generalized in public facilities.
Chapter 4 studies the implementation and the outcomes of the ambitious health sector reform initiated with the SAP in 1992. The authors look at decentralization and the other elements of the health system reform: development of the private health sector, reform of the pharmaceutical sector and hospital reform. They examine the impacts of the reform on the allocation of financial resources and on the availability and distribution of personnel, facilities, and services; they also consider the effects on prices, economic barriers, and access to health care services. The analysis includes an assessment of the economic burden of health care on families and of the impoverishing effects of out-of-pocket expenditures in urban and rural areas.
Colombia underwent adjustment programs during the 1980s in the areas of fiscal exchange, trade, and foreign debt, but adjustment was less drastic compared to other countries. Health sector reform was not an immediate consequence of MAPs, but was rather the product of a new Constitution (1991) that led to a profound transformation from a highly centralized government to a decentralized one. The new Constitution authorized the participation of the private sector in the provision of public services, set the basic rights to education, health, and social security, and made decentralization a constitutional mandate. In 1993, a wide and ambitious "pro-equity" social security reform aimed at increasing health insurance coverage and ensuring access to a basic package of services was adopted and has been progressively implemented.
The Colombian study, in Chapter 5, analyzes the equity effects of health system reform, looking especially at the decentralization of the system with social participation by local health committees and at the social security reform. In particular, the study focuses on whether coverage of the poor has improved under the subsidized system, where there has been an increase in public health expenditure in a context of low overall economic growth.
A national MAP was implemented in 1991 in India in response to the balance of payment crisis following the Gulf War. Reform measures were undertaken in fiscal policy, industrial and trade policy, exchange rate policy, foreign investment, and technology import policy. State governments, which have a significant degree of autonomy in the Indian federal system of governance, have taken their own policy stances. They have been led to reform the state apparatus, invest in infrastructure, and provide incentive frameworks for private enterprise.
With the support of the World Bank, seven states have initiated health sector reforms, called Health Systems Development Projects: Andhra Pradesh, Karnataka, Maharastra, Orissa, Punjab, Uttar Pradesh, and West Bengal. Although these projects have been implemented differently and at different times, they have objectives in common such as strengthening the performance, accountability, and efficiency of health systems. The projects seek to improve not only access to health care, but also the quality of health services, by renovating and expanding district, subdistrict, and community hospitals.
The Indian study, presented in Chapter 6, focuses on the issue of public expenditure. Trends in expenditure are analyzed over a long period—before MAPs and after—across the diverse states. Public health expenditure during the adjustment period is compared between two groups of states, i.e., those that implemented the World Bank-funded Health Systems Development Projects and those that did not. The study looked at questions such as whether the states used the World Bank loans for capital spending in the health sector and whether the hope of external funding for disease control programs led to reductions in allocation from state budgets, with possible adverse effects on programs.
While the Indian study focused on an inter-state analysis of public health-care expenditures, Chapter 7 takes up the analysis of utilization and quality of health care and out-of-pocket health-care expenditures in the state of Kerala. Kerala is often taken as a model of good health at low cost. But creeping privatization can lead to private financing of health care, with an attendant heavy burden and impoverishment of vast segments of population. Decentralization of the public health-care sector does not help greatly in mitigating this situation. The chapter explores the reasons for high utilization of private health care facilities in Kerala in terms of quality differences between the public and private sectors and analyzes the burden of health care costs.
Several adjustment packages have been implemented in Mexico since the early 1980s, each lasting about five years. In 1994, a severe financial crisis led the government and the international community to set up a financial rescue package, conditional upon a stringent adjustment program under IMF control. This crisis and subsequent adjustments led to a decline in general public expenditures and a profound deterioration in socioeconomic conditions among the population, prompting the authorities to implement social policies targeting the most vulnerable groups. A new program was created—Progresa (Program of Education, Health and Nutrition)—to improve maternal and child nutritional conditions and the quality of health and education services, as well as to provide financial support to families in the program. Expanding services to the poor and decentralizing public health services to the state level were major directions of the National Development Plan of 1995–2000, with health sector reform aimed at improving quality and efficiency and extending coverage, especially among the uninsured.
Chapter 8 examines the role of the state in providing primary care and a basic package of services to the poor. It analyzes the variation in access to primary care across the regions of Mexico and differences in access to secondary care between the poor and the rich. The perceived quality of service among private, social insurance, and SSA providers is also assessed.
In recent history, Thailand experienced three macroeconomic crises, all followed by structural reforms: the first (1973–1975) and second (1979–1985) oil crises and the 1997–1999 currency crisis. The latter led to a severe contraction of the economy. The government's response was to implement a strict stabilization program supported by the IMF. However, after being criticized for following the IMF program, which was deemed inappropriate for the Thai situation, the government reversed its macroeconomic policy and launched a stimulation package intended to bring the economy out of recession and resume economic growth.
There have been three cycles of health sector reform (1975–1980, 1990–1995, and 1996 to the present), each following a period of economic adjustment. The most recent cycle included autonomous management of public hospitals, reform of the Civil Servant Medical Benefits Scheme (CSMBS), and introduction of the Universal Coverage Scheme, or the so-called 30 Baht Scheme, offering a comprehensive benefits package for 45 million Thais.
Chapter 9 examines these three cycles of health sector reform in relation to Thailand's major economic and social transformations and, in particular, the three economic crises of 1973, 1979, and 1997. The authors explore the linkages between macroeconomic policies and changes in financing, resource allocation, coverage, and provision of services. Special attention is given to successful strategies adopted to mitigate the adverse effects of external shocks.
After severe foreign currency shortages, the Zimbabwe government launched, in 1991, its first Economic Structural Adjustment Programme (ESAP). The package was aimed at increasing investment and restoring economic growth through a market-based economy. This was followed, in 1995, by a second phase of reforms embodied in the Zimbabwe Programme for Economic and Social Transformation (ZIMPREST) that were designed to stimulate broad-based macroeconomic stability, accelerate poverty reduction, and alleviate hardships arising from ESAP.
In the period following independence, the health sector became the target of ambitious and proactive public policies, health being seen as a key component of the development process. Under pressure from the World Bank, the IMF, and bilateral agencies, health system reform was introduced in the late 1980s. Its main objective was to decentralize power and decision-making in health services to the operational level. Cost-recovery and market-based mechanisms were promoted favouring the private and non-governmental sectors.
Chapter 10 analyzes the detrimental effects of MAPs on access to and utilization and quality of health care services in Zimbabwe and the government's strategy for mitigating adverse impacts on the poor. The authors assess structural changes within the health-care system resulting from socioeconomic pressures arising out of MAPs, such as the migration of skilled health manpower and the requirement of using foreign currency to acquire the inputs for drugs.
These six countries represent diverse cases in terms of population size, level of income, human development, and implementation of MAPs and health system reforms. At one end are medium-income countries like Mexico, Colombia, and Thailand, and at the other end is the extremely poor country of Burkina Faso, with India and Zimbabwe lying in between. The diversity of cases and of health system characteristics allows for useful and informative comparisons within the framework of a common approach evolved through numerous interactions among the researchers involved in the MAPHealth project.
This page intentionally left blank
By the time the MAPHealth initiative was launched in the mid-1990s, an extensive body of literature had accumulated on macroeconomic adjustment policies (MAPs) and health sector reform. Two major streams of research had emerged. The first focused on the impacts of adjusting economies on health outcomes. In the late 1980s, UNICEF sounded a warning bell in its report Adjustment with a Human Face (Cornia et al. 1987), that captured the attention of the international community. UNICEF pointed out that the earlier gains of the 1970s in child nutrition and health were eroding under poor economic conditions. Public health activists and researchers responded with a call to protect the health of communities, particularly poor and vulnerable populations such as children. For this, MAPs would need to take into account social aspects.
Increasingly, studies have been aimed at deepening our understanding of MAPs and their impacts on equity, health, and social welfare. Since the late 1980s, researchers have attempted to assess whether MAPs have had negative consequences on health, either through the deterioration of socioeconomic conditions or the dismantling of public health-care systems. Surveying the literature, we encounter proponents of MAPs and (more often) opponents, both of whose arguments are often ideologically driven rather than empirically based. The literature, sometimes referred to as "inconclusive and often contentious" (Anand and Chen 1996, p.1), is peppered with lamentations of the lack of timely and reliable data, as well as of the methodological difficulties impeding the research process (Birdsall and Hecht 1995; Costello et al. 1994; Lundy 1996; Peabody 1996).
The other stream of research concentrated on changes in health-care systems among countries undergoing health sector reforms. During the 1980s and 1990s, health-care systems, particularly in sub-Saharan Africa, were collapsing (Simms et al. 2001; Turshen 1999). Multiple pressures, including deteriorating economic conditions, demographic transitions, HIV/AIDS, and other emerging or re-emerging diseases, were too much of a burden on already strained health-care systems. Health sector reforms, the new imperative, encompassed a range of strategies aimed at improving the financing, organization, and delivery of health services. The World Bank offered a number of policy ideas to mobilize resources and improve the efficiency of the systems (World Bank 1987, 1993).
By the 1990s, the World Bank had become a major player in the field of international health, both as donor and as policy adviser (Fournier et al. 2001). In 1998, Kamran Abbasi interviewed Richard Feachem, former director of the Bank's Health, Nutrition, and Population (HNP) sector, about the World Bank's engagement in health. Feachem gave three reasons for this new interest:
The first one is that, in all countries, health sector expenditure is a major slice of GDP. And therefore for an organisation interested in macroeconomic policy it matters. The second factor comes from the fact that the two main parts of the bank, IBRD and IDA [the International Bank for Reconstruction and Development and the International Development Association], lend to governments; and that leads us into those sectors where the free market cannot work without major government involvement… Thirdly, investment in people and the formation of "human capital" is an essential prerequisite for sustained economic growth and social development (Abassi 1999, p. 1206).
Not everyone celebrated the entrance of the World Bank into the health sector. Many feared that its market-driven approach created more problems than solutions for under-resourced health-care systems (Banerji 1994; Laurell and Arellano 1996; Loewenson 1994; Simms et al. 2001). Researchers turned to studying the effects of health-care reforms to assess changes in access and utilization of health services and to see whether their quality was improving.
However, these two areas of study—the effects of MAPs on health outcomes and the effects of health sector reform on system performance—did not tend to intersect. Comparatively less attention was paid to how MAPs transformed health-care systems in terms of access, utilization, and quality of care. Nor was there sufficient analysis of health sector reforms within the broader context of MAPs (Sen and Koivusalo 1998). The research has been criticized for being "technocratic," ignoring the wider social, economic, and political contexts (Berman 1995).
In this chapter, we do not present an exhaustive review of the material to date. Rather, we re-examine the literature, highlighting these two important knowledge gaps, from which the theoretical underpinnings of the MAPHealth conceptual framework are derived (see Chapter 3). First, we present an overview of MAPs and health sector reforms. Then, we consider the impacts of MAPs from both the supply and demand sides. We then look at health sector reforms and at strategies and assumptions most often espoused in their formulation. In conclusion, we point out areas where outstanding issues remain.
Proposed as a cure for economic woes, MAPs allowed international lenders to support a country's balance of payments on the condition that certain economic policy interventions were undertaken (World Bank 1994a). These may be divided into two main types. The first were measures, typically guided by the IMF, to correct for internal and external macroeconomic disequilibria over the short term and were aimed at contracting demand, principally through restricting public expenditures. Once an economy had stabilized, policies could then be implemented to address structural barriers to production (Husain 1994). This involved medium- to long-term measures for stimulating economic growth. These "supply" policies relied on a variety of directions recommended by the World Bank that included, among others, liberalization, privatization, and deregulation. MAPs varied with respect to their content, scope, intensity, duration, and ultimate outcomes, leading authors to propose different approaches to classifying MAPs and the countries adopting them, for purposes of comparison.
One approach used a conceptual classification to provide a framework for understanding the evolution of policies. Cornia (1992) described the progression of MAPs over time, distinguishing among three different generations. During the early 1980s, a first generation of MAPs maintained purely fiscal and monetary objectives. The main goals were to stabilize the economy and liberalize the market. There were no stipulations to protect social sectors or vulnerable populations. The second generation expanded supply policies across sectors, including industry, agriculture, energy, finance, banking, health, and education. In their 1990 World Development Report, the World Bank officially recognized the need for social funds (Cornia 2001), and by 1991, MAPs included at least one condition specific to health (Van der Gaag and Barham 1998). Certain policies included compensatory measures for temporary relief from the costs of reforms. Finally, the third generation aimed to integrate economic and social objectives, with explicit intentions of protecting health and education.
An operational approach was used by two sets of authors in their respective cross-country studies looking at conditions that enabled MAPs to be operationalized. Kakwani's (1995) system divided countries according to whether they underwent adjustment or not. He further divided adjusting countries according to whether they had received single or multiple loans. This approach measured the intensity of MAPs. Van der Gaag and Barham (1998) divided countries into four groups. The first two included adjusting countries, further subdivided by time period: Early Adjusting Countries (EAC) and Other Adjusting Countries (OAC). Countries that did not implement MAPs were divided between those having achieved positive economic growth and those where growth had been negative. This system also captured the intensity of MAPs. Although these systems of classification facilitate comparisons among countries by capturing certain characteristics of MAP implementation, they fail to represent the complexity and diversity of country experiences and ultimately provide limited perspectives in terms of political and social contexts.
A third approach categorized MAPs according to outcomes, using a normative classification system based on judgments related to the adherence to MAPs, as illustrated in a World Bank evaluation study of the period 1980–1993 (Jayarajah et al. 1996). Countries were classified into four categories according to their experiences with policy instruments. The first, labeled "good cases," included those countries that followed the "right" policy and achieved the "right" outcome. The second, called "bad cases," followed the "wrong" policy and got the "wrong" outcome. The third and fourth categories are the "ugly cases," which included "unlucky" countries that implemented the "right" policy but achieved the "wrong" outcome and "lucky" countries that pursued the "wrong" policy but got the "right" outcome. The "ugly cases" were a result of exogenous events—the unlucky cases faced negative circumstances, such as adverse trade conditions, while the lucky cases experienced favourable events. The rather unfortunate and simplistic labeling of the cases (the good, the bad, and the ugly) fails to acknowledge that governments can autonomously design and implement policies leading to positive outcomes and makes it impossible to extract lessons from these cases. Nor does this system address the shortcomings of MAPs and the crucial dimension of country diversity in shaping public policies.
These approaches—conceptual, operational, and normative—illustrate the complexity associated with the analysis of MAPs. Although policy differences should be considered in evaluating MAPs, none of these approaches sufficiently takes into account the diversity of country experiences.
Most countries that underwent MAPs during the 1980s and 1990s also initiated health sector reform. Sometimes these reforms were stimulated by problems related to the economy, or by specific causes within the health sector, or combinations of both. Financial crises in the public sector led to severe consequences, such as deterioration of public health infrastructures and quality of health services, and declining capacities of governments to provide free health care for their citizens. Policies shaping the organization and delivery of quality health services were inadequate, focusing largely on curative care within hospital settings.
Health sector reforms adopted different forms and varied with respect to content, timing, and implementation. Generally, these reforms were aimed at increasing health sector financing; reallocating public budgets to promote primary health-care services; reorienting public funds to benefit the poor; improving the quality and range of services; increasing utilization; decentralizing the public sector through institutional reforms; reinforcing managerial capacities; and liberalizing the health sector by expanding private sector involvement (Gilson and Mills 1995).
In some countries, health sector reform was an integral part of adjustment. In others, reforms were undertaken independently, although MAPs could either precede or follow. In certain rare cases, MAPs were implemented without concomitant reforms of the health sector. In studying the first two types, particular efforts are required to dissociate influences arising from MAPs and from health sector reforms.
The effects of MAPs on the health sector are multiple and interrelated, affecting the macro (e.g. macroeconomic or sectoral policies), meso (e.g. health-care systems, communities), and micro levels (e.g. households, individuals). MAPs may operate through either the supply or demand side (Anand and Chen 1996; Costello et al. 1994). Many studies explored the impacts of MAPs on the demand side, specifically by assessing changes in health outcomes. Comparatively less effort was made to understand change processes in the health sector by analyzing factors influencing health, including both medical and non-medical determinants of health. The supply side of MAPs was also explored. However, except for a few notable studies on changes in quality of care, most concentrated on the evolution of public health-care expenditures. Assessing the health impacts of MAPs is plagued with methodological difficulties; weighing the evidence requires explicit knowledge of methods employed, data sources used, and the context in which the study took place.
A recent review of the literature prepared for the World Health Organization (WHO) Commission on Macroeconomics and Health surveyed the debate, including the empirical evidence and methodological approaches adopted (Breman and Shelton 2001). Several trends are worth mentioning. First, of the 72 articles reviewed, although normative articles generally argued that MAPs had negative consequences on health, the empirical evidence presented both negative and positive outcomes. Second, among 28 empirical studies, the two most widely studied outcome variables were public health expenditures and child mortality. Other variables studied were malnutrition, life expectancy, and maternal mortality. A few studies also analyzed quality of health care. Both positive and negative outcomes were described in these empirical studies. Third, studies were divided evenly into whether a case study approach or cross-country analysis was used. Negative effects tended to be reported in case studies, whereas cross-country comparisons illustrated both positive and negative outcomes on health. Finally, there were no discernible trends with respect to region, except in Africa, the most widely studied region, that generally displayed negative outcomes. The review does not highlight differential impacts of MAPs on distinct populations such as women, indigenous populations, or people living in urban versus rural areas, hiding possible inequalities arising from, or worsened by, the implementation of MAPs. This omission is most likely due to a lack of disaggregate data at the time, but these potential inequalities should be kept in mind when assessing the health impacts of MAPs.
As highlighted by the review of Breman and Shelton, the impacts of MAPs on health status and malnutrition, particularly among children, were intensively studied. Child mortality and malnutrition are indicators sensitive to changes in health policies and programs, and these indicators are generally available in poor countries (Garenne and Gakusi 2000; Musgrove 1986). The results were, however, mixed. The literature is complicated by disagreements among researchers at two different levels. First, the evidence diverges on whether health status deteriorated or improved following the period of adjustment. This may be related, in part, to regional or local conditions. For example, increases in malnutrition were noted, at least for a short period, in most African countries during the 1980s (Stewart and Van der Geest 1995), while in Latin America, nutritional status improved in some countries but declined in others (Tardanico and Larín 1997). UNICEF (1984) found increasing malnutrition in Costa Rica and Bolivia during the early 1980s, while in Chile and Cuba, child nutritional status was improving and infant mortality rates were declining (Musgrove 1986).
Different results may also arise depending on how health or nutritional status is measured. Even though many studies indicated that health deteriorated in sub-Saharan Africa (Breman and Shelton 2001), Garenne and Gakusi (2000) found child survival either stabilized or improved among 10 African countries using "innovative methodology." Positive and negative outcomes were often found within a single study, illustrating the challenges of assessing "health." For example, a survey of child nutritional status in a community in the Congo noted an increase in wasting but a decline in levels of stunting (Cornu et al. 1995). This type of pattern, demonstrated elsewhere, can be explained by the fact that the indicators measure different dimensions of malnutrition; wasting captures short-term impacts, whereas stunting is an indication of long-term undernutrition (Mohindra 1999). This reaffirms that, in studying the impacts of MAPs on health, distinction must be made between short-term and long-term effects (Musgrove 1986).
Another area of disagreement among researchers is in how changes in health status are explained. Certain authors attributed deterioration of health and nutritional status directly to MAPs (Kanji et al. 1991; Loewenson 1993), whereas other, more nuanced, interpretations inferred that MAPs did not protect populations from wider economic conditions (Cornu et al. 1995). MAPs have also been credited with stimulating improvements in health. For example, a multicase study of African countries argued that health improved because "positive forces associated with SAP's were at work," particularly better governance and management (Garenne and Gakusi 2000, p.15). Some argued that causal links could not be made because the study designs were incapable of discounting rival hypotheses. Other explanations were offered for the reversal of trends in health gains. For example, the World Bank proposed that improvements in health status (measured as life expectancy) of populations in countries undergoing adjustment were declining because infant and child mortality rates were stabilizing (World Bank 1996).
The lack of clear evidence makes it difficult to formulate general statements about the health effects of MAPs. Methodological difficulties in linking MAPs to health status are extensive (see Chapter 3). Measuring health status is not enough. Explicit pathways between MAPs and health outcomes must also be described and analyzed.
Health is determined by a number of factors both within and external to the health sector (Evans et al. 1994; Kjellstrom et al. 1992). Although deterioration in living conditions was recognized as a potential link between MAPs and health outcomes, a more complex analysis of how MAPs influenced the determinants of health was lacking (Yach and von Schirnding 1994). There were, however, a number of studies outside of public health that examined the impact of MAPs on key inputs for both health and health care, such as economic status, employment, and education. Not surprisingly, the evidence was contradictory and inconclusive.
Increases in poverty have enormous implications for health; poverty is both a fundamental cause of disease and a barrier to health care (Wagstaff 2001). Increased poverty is especially associated with malnutrition, and some argued that MAPs led to higher food prices and reductions in subsidies, affecting the diet of the poor (Schoepf et al. 2000). Poverty studies varied in their results and were hindered by methodological problems and data deficiencies. As in the health sector, studies diverged on the nature and severity of changes in poverty and on the extent to which these could be linked with MAPs. Differences occurred both across and within countries and depended on how the poor were defined and measured. While certain studies reported increases in poverty, particularly during the early stages of MAPs (ECLAC 1996; Stewart and Van der Geest 1995), World Bank studies showed that poverty was declining in countries undergoing adjustment (Jayarajah et al. 1996). Sahn and colleagues found not only that the poor did not suffer from the effects of MAPs, but also that their lives improved slightly. They analyzed 10 African countries and concluded that:
… reforms primarily hurt those who receive rents that arise when goods are rationed, benefit from guaranteed public employment, gain a disproportionate share of subsidies on poorly targeted public services, and profit from interference in agriculture and food markets. Africa's poor rarely belong to these groups. As a result, policies that reduce or eliminate these privileges in the name of economic efficiency are not usually prejudicial to the poor (Sahn et al. 1997, p. 247).
Others have indicated that MAPs created a new class of poor, the "adjustment poor," as opposed to the "chronic poor," who were poor before MAPs. The adjustment poor spurred the development of social funds (Cornia 2001). Eventually, the distinction between the two types of poor became blurred and the social funds required expansion because the chronic poor were being excluded from employment and welfare programs. Although debates continue regarding the impact of MAPs on the poor, whether pre-existing or newly created, it is clear that MAPs had a large influence on poverty through changes in employment policies (Gershman and Irwin 2000).
In many low-income countries, particularly in sub-Saharan Africa, agriculture was the most important employment sector during the 1980s (Basu and Stewart 1995). The informal sector played an increasingly important role in the 1980s and 1990s, providing an important refuge for people during times of economic crisis and restructuring. Following the implementation of MAPs, there were reports of declining wages and increasing levels of unemployment. For example, declining wages were observed in Latin America during the 1980s, and although they began to reverse in the 1990s, unemployment rates remained high (ILO 1991; Morley 1995). Using general equilibrium models, a multicase study on MAPs also noted increasing unemployment, to the extent that the number of unemployed would have tripled if people had not moved over to the informal sector (Bourguignon and Morrisson 1992). Changes in the workforce affect women in particular, in two ways. First, due to male bias in the labour market, women are more vulnerable to declining wages and loss of jobs (Elson 1992; Stewart 1992). Second, they are more likely to spend their time coping with reductions in household income and increasing health needs of family members. Many case studies demonstrated women not only continued their unpaid functions, such as domestic chores and care of family members, but also took on paid employment to supplement family income, often at the expense of abandoning educational goals (Garcia and Oliveira 1997; Gindling 1993; Lam and Levison 1997; Minujin 1992).
The effect of MAPs on education is also a controversial issue. It was argued that, as in the health sector, education was protected (Ferroni and Kanbur 1990; Sahn et al. 1997). However, others suggested that MAPs had negative impacts on the education sector (Reimers 1994; Reimers and Tiburcio 1993). Reimers (1994) argued that cuts in recurrent expenditures lessened the quality of education systems, thereby reducing demand for schooling. The combined effects of MAPs on the demand and supply of education reduced primary school enrolment rates, particularly among girls (Cornia et al. 1987; Jayarajah et al. 1996). Reimers (1994) noted that, at the primary level, enrolment rates were twice as high in non-adjusting than in adjusting countries, but that this trend was not found at secondary and tertiary levels. Grootaert (1994) underscored the importance of analyzing differences in area of residence, socioeconomic status, and gender when assessing the impact of adjustment on education. He studied the Ivory Coast, often cited as an example of a country that invested heavily in education and was superior to other countries in protecting this sector following adjustment. The analysis of household survey data between 1985 and 1988 indicated there were important inequalities in school enrolments. School attendance declined, but only among poor households, and particularly for girls, and in urban areas.
Declining participation of girls in the education system is particularly troublesome to the development community, because education of women is one of the key determinants of household health and a major contributor to lowering fertility rates (World Bank 1993). Education also improves women's chances of participating in stable and well paid jobs (Buchmann 1996). Removing girls from school became an important coping strategy for households under financial constraints, either because of increasing school fees or declining household incomes.
When school expenses increase dramatically (as government subsidies are eliminated), parents may decide to keep their daughters at home. If parents know that women have limited opportunities in the formal market, they may reason that the benefit of educating their daughters no longer outweighs the expense. Moreover, daughters may be kept at home to help with the increasing amount of household labour. Thus, household strategies to survive during periods of austerity may inhibit female access to education (Buchmann 1996, p. 8).
Studying changes in public health expenditures was popular among researchers, particularly at the World Bank and IMF. These studies were facilitated by the relatively wide availability of this data (Costello et al. 1994). During the early 1980s, governments were supposed to cut expenditures, including health, to improve their economy. The World Bank suggested that these cuts would not necessarily be detrimental because a more efficient use of resources would improve health outcomes (Alderman and Lavy 1996; World Bank 1993). A number of authors pointed to a decline in public health expenditures following the adoption of MAPs during the 1980s (Cornia et al. 1987; Kanji et al. 1991; Stewart 1989; Woodward 1992). Others concluded that, in countries where public health expenditures declined, there was no evidence this was caused by budget cuts related to MAPs (Sahn 1992; Sahn and Bernier 1995; World Bank 1993).
Jayarajah and colleagues (1996) found that between 1980 and 1993, expenditures declined in most countries, although considerably less among the "good adjusters" compared to both non-adjusting countries and those that did not adjust properly. The World Bank argued that MAPs had evolved to consider social sectors, such that public health expenditures were protected during later generations (World Bank 1990). A study comparing health expenditures among non-adjusting and adjusting countries that differentiated "early" adjusters from others came to the same conclusion (Van der Gaag and Barham 1998). This study found that health expenditures declined more in adjusting countries than in non-adjusters during the 1980s, but that this was reversed during the 1990s. However, these results should be interpreted with caution, as the authors themselves mention their data was of poor quality. The evidence is conflicting across studies, and within studies there are differences across countries. Several explanations have been proposed, including government's level of commitment to protecting the health sector, the severity of the economic crisis, and the manner in which the MAPs were implemented (Breman and Shelton 2001).
The picture is clearer with respect to the impact of MAPs on the allocation of public health resources. Most studies, including those of the World Bank, concluded that MAPs did not improve allocative efficiency in the health sector. The World Bank observed that hospital care continued to receive the lion's share of health resources at the expense of primary care, even in countries where there was a decline in health-care expenditures (Jayarajah and Branson 1995). In the African context, Sahn and Bernier (1995, p. 254) concluded that "while not being able to resolve the accounting and definitional problems, our review of existing data paints a picture that despite the assertions of most governments, and years of engaging in structural adjustment, preventive and promotive care remains neglected at the expense of curative care." This is problematic because most of the health needs in the South, particularly among the poor, require primary health care.
Studies examining the supply of health services were overwhelmingly centred on health expenditures, with studies linking MAPs to access, utilization, and quality of health care being more rare. Two studies demonstrated associations between perceived declines in quality and the implementation of MAPs. One examined nurse professionalism, patient satisfaction, and quality of care in Zimbabwe following a structural adjustment program in 1991, and others demonstrated drops in utilization rates (Bassett et al. 1997). Bassett and colleagues conducted focus groups with nurses and with community members, in which there was consensus that people were dissatisfied with clinic service fees, drug shortages, and long waiting times for care. Although such problems were not new, the participants perceived a notable increase. There were also more conflicts between nurses and patients. Patients felt nurses were indifferent toward them, and nurses complained of being overworked and undervalued. The worsening patient–nurse relationship was clearer to patients than nurses. The authors argued that these conflicts were not created, but rather exacerbated, by adjustment.
Lundy (1996) examined the case of Jamaica, conducting semi-structured interviews with various health professionals, as well as clinic interviews, and surveying participants' observations of clinic conditions. Most described a deterioration of public health services, which they associated with Jamaica's debt crisis and specifically with MAP-related cuts to social expenditures. Although the departure of health personnel had already been problematic, this exodus intensified following adjustment. Informants observed that patients had become disillusioned with public services, complaining of longer waiting times and shortages of supplies. They also noted declining standards among health professionals and increasing environmental health risks. Bassett and Lundy documented important declines in quality of care, although it is difficult to ascertain the specific role of MAPs as opposed to economic crisis.
Even if health sector reform does not include any "consistently applied, universal package of measures" (Haran 1998, p. 768), it generally comprises three core measures: financing reforms, decentralization, and integrating competition into the health sector (McPake and Machray 1997).1 We examine each of these in turn, beginning with financing reforms, which have accumulated the most evidence thus far.
Driven by a scarcity of resources in the health sector, governments faced the dilemma of how to provide adequate health services to their citizens. The solution proposed by the World Bank (1987) was to encourage governments to apply a user fee at the point of service delivery, both to increase revenue and to create a greater rationalization of services, because the cost would incite people to "think twice about demanding it" (Shaw and Griffin 1996). The World Bank further argued that cost recovery would improve equity because as the contributions of households (especially richer ones) increased, there would be more money in the system, making possible more and better services for all people, including the poor (Griffin 1992; World Bank 1993). The severity of the financial crisis in the health sector and, in certain cases, explicit loan conditions left governments little choice but to follow the route of cost recovery. In 1987, African Ministers of Health launched their own cost-recovery program, with the support of UNICEF and WHO, known as the Bamako Initiative (UNICEF 1992). This initiative sought to mobilize resources through community financing to improve primary health-care services and especially to ensure availability of essential drugs. The scarcity of resources for African health systems led UNICEF to campaign for a selective approach to primary health care, by improving people's health through a limited number of targeted interventions. UNICEF was criticized for following in the ideological footsteps of the World Bank, and it was argued that the Bamako Initiative undermined poor people's access to health services (Kanji 1989).
Before the merits of cost recovery had been sufficiently proven, however, governments across the developing world were implementing this policy. In 1997, over 30 African governments were operating cost-recovery programs (Watkins 1997). The hasty introduction of user fees raised concerns, particularly regarding equity (Gilson et al. 1995), and the dialogue became focused not on whether but on how user fees should be implemented. The vast array of practical cost-recovery experiences has provided fertile ground for testing the efficiency and equity claims made in the 1980s. These claims were based on a number of assumptions, of which the following four are most prominent.
Early World Bank proponents of user fees argued that the demand for health care was price inelastic, such that implementing or increasing costs did not deter people from using health services when in need (de Ferranti 1985; World Bank 1987). Since then studies, including those by the World Bank, have repeatedly demonstrated that demand for health care is in fact price elastic, especially among the poor (Creese 1994; de Béthune et al. 1989; Gertler and Van der Gaag 1990; Haddad and Fournier 1995; Mwabu and Wang'ombe 1995; Sauerborn et al. 1996; Sahn et al. 2003; Waddington and Enyimayew 1989; Yoder 1989). Studies of the demand for health care illustrated that user fees were sensitive to price, but other alternatives for increasing resources were scarce. Some authors proposed that the focus be shifted toward minimizing the regressive nature of cost recovery and, in particular, to improving quality (McPake 1993).
User fees were supposed to increase resources in the health sector that could subsequently be used to improve the range and quality of services (World Bank 1987). It was further argued that better quality of health care would lead to increased use of services, including among the poor (Alderman and Lavy 1996). A review of user charges noted that early studies largely ignored the question of quality, focusing instead on financial gains (McPake 1993). Later studies considering quality, notably the often-cited study by Litvack and Bodart (1993) in Cameroon, observed higher utilization when quality improvements accompanied the introduction of user fees. Some studies supported the notion that quality improvements can negate rising fees (Chawla and Ellis 2000; Diop et al. 1995; Knippenberg et al. 1997). Others, however, suggested the issue is more complex. For example, in a longitudinal study in Zaire, utilization rates declined by 40% within a five-year period, even with quality improvements (Haddad and Fournier 1995). Cost of care explained 18–32% of the drop.
Effective exemption policies would ensure that the poor would not have to pay the full cost of care. Equity would improve because richer patients would subsidize the poor through user fees at public health-care facilities, thereby raising quality to the benefit of all users. Although many governments designed exemption policies, they were difficult to put into action (Kaddar et al. 1997). Administrative costs were high, and many countries had neither the capacity nor the political will to put in place exemption policies (Watkins 1997). More importantly, the systems were simply ineffective (Gilson and Mills 1995; Gilson et al. 1995; Watkins 1997). Exemptions were only weakly linked to the poverty status of users (Ensor and Bich San 1996; Newbrander et al. 2000; Nolan and Turbat 1995). For example, in Kenya, only 1–5% of users were exempted because of their poverty status, although 11–34% of the population fulfilled the criteria for exemption (Huber 1993). In a multi-country study, fees were not waived among the poor because either the poor were unaware of such waivers or an exemption system was not in place (Newbrander et al. 2000). Leakage also occurred, such that non-eligible users were being exempted (Gilson et al. 1995). For example, one study found the non-poor benefited from exemptions in 88% of cases in Senegal, 75% in Burkina Faso, and 73% in Niger (Leighton and Diop 1995). Thus, without effective exemption policies, gains in equity become illusory.
Cost recovery relies on the principle that people are willing to pay for health services (Shaw and Griffin 1995). It was argued that people were already paying for health care, whether for private services or unofficial fees. Furthermore, studies indicated that people were willing to pay for health services if there were quality improvements and that utilization rates did not necessarily decline among the poor following the implementation of user fees and quality improvements. This was considered to be sufficient evidence that there were no equity trade-offs to cost recovery. There was, however, conceptual confusion between the concepts of willingness to pay and ability to pay (Russell 1996; Watkins 1997). Watkins (p. 14) observed that "markets do not reflect human need, but monetised consumer preference, as expressed through the price system." Russell urged that studies should analyze the impact of user fees not only on utilization, but also on household processes. This has been corroborated empirically. For example, a study in Vietnam concluded that "while charges may have a relatively small impact on current willingness to pay, they do have a substantial impact on income, wealth and the ability to pay for treatment in the future" (Ensor and Bich San 1996, p. 79).
In examining the equity impacts of user fees, the financial burden of care must also be considered. Studies have demonstrated that, in countries where no effective national health insurance system is in place, the poor have a greater financial burden of health care than the non-poor because they pay proportionately more of their household income for health care. In Vietnam, about 19% of household incomes of the very poor was allocated to health care, compared to almost 4% for the rich (Ensor and Bich San 1996). This phenomenon has been noted in a number of countries (Berman 1995; Fabricant et al. 1999; Pannarunothai and Mills 1997; Sen 1997).
Furthermore, studies have shown that user fees elicit household coping strategies that can be destructive to families, particularly over the long term. These include delaying or not seeking treatment, seeking informal care, or self-treatment (Sahn et al. 2003). Health services may be obtained at the expense selling off assets, removing children from school, reducing food consumption, or borrowing from friends and relatives. Gender-based coping strategies also contribute to increasing inequalities. During periods of financial stress, girls were more likely to be removed from school than boys (Buchman 1996). Also, women were more likely to take on employment to raise income for families, while continuing unpaid work, such as domestic chores, child rearing, and caring for family members who had fallen ill (Garcia and Oliveira 1997; Gindling 1993). Kanji emphasized that our understanding of user fees will depend on the questions asked. For example, "How much would you be willing to pay for an x% reduction in the risk of dying?" elicits different information than "How much can you afford to pay for a particular treatment without getting into debt or cutting down on your food bill?" (Kanji 1989, p. 118). To assess the full impact of user fees, studies must recognize "the multi-faceted nature of vulnerability and responses" (Watkins 1997, p. 15), by considering not only impacts on health-seeking behaviours, but also the concessions made by the poor to cope with costs.
In sum, cost recovery has not proven to be a panacea for cash-strapped health systems in poor countries. Equity and efficiency are current objectives in improving health-care systems, and access to health care for the poor is at the forefront of policy discussions (Kaddar et al. 1997). Certain governments are beginning to reverse the policy on cost recovery. For example, in 2001 the Ugandan government removed their cost-sharing program and supplemented the health budget with funds from a World Bank project (Burnham et al. 2004). Some argue that cost recovery remains essential but must be done humanely (Van der Geest et al. 2000), and others argue that cost recovery is simply not the best policy option for poor countries and that other financing mechanisms should be considered, such as community health insurance (Arhin-Tenkorang 2001).
According to Gilson and Mills "the main public sector organizational reform … is the decentralization of planning and management, which aims to push responsibility for at least some decisions down the administrative hierarchy" (Gilson and Mills 1995, p. 292). The four main objectives of decentralization were to improve planning and management of services, to create a more effective and efficient organization of services, to increase accountability, and to encourage community participation in decision-making (Gilson and Mills 1995). Although not a new concept, decentralization received enormous support during the late 1980s and 1990s from donor institutions, including the World Bank, WHO, and USAID. Even if decentralization is purported to be inherently good, benefits remain for the most part speculative (Bossert 1998; Collins 1996). It is possible that decentralization merely provides a channel for central governments to shift financial and administrative burdens to lower levels because of limited resources at the national level (Sahn and Bernier 1995), as illustrated in Nicaragua (Birn et al. 2000).
Because decentralization involves reshaping institutional arrangements and redistributing powers and tasks within the health sector, it was difficult to link impacts with changes in access to and utilization of services (World Bank 1994b). Decentralization is best viewed as a complement to other reform measures, such as user fees and increased market competition. Although there are few studies linking decentralization with health system performance (Bossert 1998), various experiences with decentralization do offer lessons by highlighting problems and successes that arose in practice.
To ensure proper functioning and acceptance by health workers, roles and regulations related to decentralization must be clear. A review of 10 countries highlighted this as a problematic issue (Kolehmainen-Aitken 1998). The problem is amplified by the number of organizations involved. Larbi (1998) noted that the multiplicity of institutions involved in Ghana—Ministry of Health, Ministry of Finance, Auditor General, Bureau Chief, and Commission of Public Services—contributed to policy confusion, conflicts of authority, and other administrative problems and consequently led to only a partial decentralization. Conflicts also arose in Uganda between the Ministry of Health and district staff, particularly related to budget priorities (Jeppsson 2001). These conflicts were further exacerbated by conditions imposed on the districts by central government.
Decentralization also provoked resistance. In some countries, the Ministry of Health was less than enthusiastic, partly because this reform was externally imposed, entailing a shift in powers (Collins 1996). Health providers, such as hospitals and NGOs, also resisted coming under the control of districts or municipalities and losing privileges they had enjoyed under the Ministry of Health (Prud'homme 1995). In Nicaragua, the central government had to withdraw local control of five hospitals because of problems between hospital directors and local health authorities (Kolehmainen-Aitken 1998).
To improve efficiency, resources were to be decentralized to the periphery. However, political obstacles presented significant challenges to the equitable redistribution of resources. Also, many governments did not possess sufficient information to complete such transactions, nor were procedures, such as planning and budgeting, properly in place. Tanzania was a case in point, where communities continued to rely on central government funding after decentralization (Gilson et al. 1994). Decentralization in Papua New Guinea led to resources being allocated to economically and politically stronger regions (Campos-Outcalt et al. 1995). In contrast, decentralization in Bolivia resulted in the creation of 200 new municipalities that adopted a per capita formula for reallocating resources that enabled previously disadvantaged areas to receive more resources (Holley 1995).
Other important concerns continue to be raised about decentralization. Community participation, advocated by the World Bank (1994), was often a top-down process, such that formal consultations with communities were largely absent (Smithson et al. 1997). Decentralization also had negative impacts on human resources (Kolehmainen-Aitken 1998) and information systems, which were already ineffective in poor countries (Collins et al. 1999; Thomason et al. 1991). Studies are also emerging on experiments with autonomous hospitals, a form of decentralization in which hospitals are given greater authority to pursue their own goals (Collins et al. 1999).
In sum, although decentralization has had many proponents, concerns remain. In situations where reforms have led to lesser forms of decentralization, the question may not be whether to decentralize, but rather how to ensure that "whether or not authority and accountability for health services delivery are fully transferred to subnational levels, parallel changes in financing and organization must not be ignored" (Birn et al., p. 126).
The World Bank encouraged reducing the role of the state and introducing competition into the health sector by expanding the private sector. This strategy is based on the belief that the market is a more efficient producer of goods and services.
For some health services provided by the public sector, the system of provision is so grossly inefficient that it is unlikely to be cost-effective no matter what interventions the system tries to provide. Such inefficiencies have been criticized so clearly and for so long that it is evident they will only be overcome by radical changes in the organization of the health care—such as a shift in the government's role from providing care to financing care and stimulating competition among providers (World Bank 1993, p. 65).
The World Bank also argued that private health expenditures were already high in the South and that many people would be willing to pay for better quality services in the private sector, freeing up resources for the poor. Unfortunately, a lack of even basic data, such as on levels and composition of private health care provision, has hampered a better understanding of the private sector in the South (Hanson and Berman 1998). Several authors point out the important and growing role of the private sector in many poor countries (Berman 1998; Bhat 1996; Gilson and Mills, 1995; Ogunbekun et al. 1999; Sahn and Bernier, 1995; World Bank, 1994b). Two types of privatization, passive (or incremental) and active, have been identified in the literature (Uplekar 2000). Passive privatization is largely unplanned and is a result of public sector failure, whereas active privatization is advanced by purposeful public policy. The deterioration of public health services in many countries has also contributed to more people seeking health care in the private sector, including the poor (Bloom and Lucas 2000; Bloom and Standing 2001).
A number of strategies have been used to stimulate private sector involvement, including subsidizing NGOs (Gilson et al. 1994; Laffont and Tirole 1993), contracting out (McPake and Ngalande-Banda 1994; Mills 1998), introduction of managed care (Iriart et al. 2001), integration into planning processes (Bennett et al. 1994; Larbi 1998); and institutional reforms (Bossert et al. 1998; PAHO 1998). There are few studies evaluating the impact of increased competition in the health sector on access and quality of care. Concerns have been raised about increasing private participation, particularly when this occurs rapidly without proper regulation in place.
These concerns may be illustrated by two examples from the literature. First, looking at countries that have integrated contracting, we consider the claim that there will be efficiency gains in the private sector. Second, we look at the quality of private provider practices in the management and treatment of tuberculosis (TB) in India. Such examples highlight the important counterbalancing role of regulation.
Governments can stimulate private sector involvement by contracting the services of private providers. In a study done in South Africa, Thailand, and China of contracting-out practices, efficiency gains were partially or completely offset by administrative costs (Mills 1998). Large transaction costs were also found in Chile and the Philippines, on the order of 30% and 45%, respectively (Tollman, Schopper, and Torres 1990).
Negotiating, preparing, and managing contracts also add to the strain on human resources, because these skills were not needed when health care was publicly provided (McPake and Ngalande-Banda 1994). It is not clear that public services are less efficient than private services. Mills (1998, p. 39) wrote that "it is far from self-evident that efforts to establish contracting will provide greater returns than efforts to improve public service provision." Transaction costs must also be included in any comparison of efficiency between the private and public sectors (Hsiao 1995).
The profit motive can encourage unethical practices of private providers (Bennett et al. 1994). Providers may cut costs by reducing quality of care, or create profits by doing more expensive tests. Uplekar (2000) provided an interesting example in his studies of private providers and tuberculosis (TB) in India. TB is manageable and curable if diagnosis is accurate and treatments closely monitored. There were positive observations, such as proximity of private providers to patients, patients' confidence in their providers, and acceptance of services despite costs. There were also a number of what Uplekar called "ignorant practices." For example, private providers tended to use X-rays instead of sputum examinations for testing, thereby increasing costs and reducing reliability of diagnosis. Many providers did not reveal the prognosis of TB to their patients. Many inappropriate and costly drugs were prescribed and follow-up with patients was minimal. Record-keeping was also not a strong point. There were discrepancies between provider advice and patient behaviors, and essential records were not kept. Uplekar also cited similar performance of providers in other countries. For example, in Korea, in a study of 923 private providers, 50% stated that sputum examinations were not essential, 75% said treatments did not need to be monitored, 73% gave non-recommended regimens, and 16% provided unacceptable regimens. It is clear that state involvement is required through regulative practices to discourage poor quality and unethical practices.
The state is recognized as playing an important role in addressing problems arising from market involvement, particularly through regulation (Söderlund and Tangcharoensathien 2000; World Bank 1997). Regulation involves control through either incentives or legal restrictions. However, even when regulations are in place, they are often outdated and rarely effective (Hongoro and Kumaranayake 2000; Kumaranayake 1997). In Zimbabwe and Tanzania, legislative regulation existed but important gaps remained, particularly related to consumer protection and development of new private markets (Kumaranayake et al. 2000).
In sum, despite increased marketization of the health sector, there is still no evidence that the private sector is more efficient. Furthermore, marketization has rarely been accompanied by effective regulation, thereby failing to protect patients, particularly the poor, from inappropriate care. More resources are needed to improve regulatory practices to keep up with the growing private sector (Kumaranayake 1997).
Improving health sectors in the South remains a critical issue for the 21st century. The 1990s produced a vast number of publications but, curiously, attention has since waned even as more data become available and better tools are developed for evaluating macroeconomic and sectoral policies. One editorial stated that "it seems as if the academic community lost interest in writing about health sector reform at about the same time that it became possible to study its results" (Blas and Hearst 2002, p. 2). Many questions remain in understanding the full impact of MAPs on the health sector, although certain knowledge gaps are perhaps more pressing. Sen and Koivusalo summarized a number of priority issues:
It is important to develop case studies from different regions of the world allowing comparison of the processes and possible impacts of changes upon vulnerable groups and health systems as a whole. In view of the evidence already emerging, there is also a need to support evaluations and descriptive studies of process rather than focusing on single measures of outcomes such as clinic attendance, drug availability or waiting times, in order to assess the overall impact of these changes. Furthermore, the cumulative effects of these changes need to be better understood. More information is needed on the meso and macro-level impacts of the reforms on health systems development, as well as on micro level impacts at the household level, analysed by gender and socio-economic status over a long-term perspective (Sen and Koivusalo 1998, pp. 209–210).
Two areas for further study would be the differential impact of reforms on rural and urban areas with respect to issues such as ability to pay and use of services and the impact of reforms on indigenous populations.
The MAPHealth initiative grew out of a need to gather stronger evidence on the impacts of reforms driven by macroeconomic instability. Protecting the health sector requires a better understanding of how health systems are influenced by factors within and external to the health sector at both national and international levels. Insights gained from the experiences of countries undergoing macroeconomic and health reforms are a valuable resource for policymakers, international donors, and civil society, by highlighting which policies are most equitable and effective in turbulent economies.
1. Other authors have provided further breakdowns of reform measures. See especially Cassels (1995).
Abbasi, K. 1999. The World Bank and world health: interview with Richard Feachem. British Medical Journal, 318, 1206–1208.
Alderman, H.; Lavy, V. 1996. Household responses to public health services: cost and quality tradeoffs. The World Bank Research Observer, 11(1), 3–22.
Anand, S.; Chen, L. 1996. Health implications of economic policies: a framework of analysis. Discussion paper series, no. 3. United Nations Development Program, Office of Development Studies, New York, NY, USA. 34 pp.
Arhin-Tenkorang, D. 2001. Mobilizing resources for health: the case for user fees revisited. CID Working paper no. 81. Center for International Development (CID), Harvard University, Cambridge, MA, USA. 21 pp.
Banerji, D. 1994. A simplistic approach to health policy analysis: the World Bank team on the Indian health sector. International Journal of Health Services, 24(1), 151–159.
Bassett, M.T.; Bijlmakers, L.; Sanders, D.M. 1997. Professionalism, patient satisfaction and quality of health care: experience during Zimbabwe's structural adjustment programme. Social Science and Medicine, 45(12), 1845–1852.
Basu, A.; Stewart, F. 1995. Structural adjustment policies and the poor in Africa: an analysis of the 1980s. In Stewart, F., ed., Adjustment and poverty: options and choices. Routledge, London, UK. pp. 138–170.
Bennett, S.; Dakpallah, G.; Garner, P.; Gilson, L.; Nittayaramphong, S.; Zurita, B.; Zwi, A. 1994. Carrot and stick: state mechanisms to influence private provider behaviour. Health Policy and Planning, 9(1), 1–13.
Berman, P.A. 1995. Health sector reform: making health development sustainable. In Berman, P.A., ed., Health sector reform in developing countries: making health development sustainable. Harvard University Press, Boston, MA, USA. pp. 13–36.
_____ 1998. Rethinking health care systems: private health care provision in India. World Development, 26(8), 1463–1479.
Bhat, R. 1996. Regulation of the private health sector in India. International Journal of Health Planning and Management, 11(3), 253–274.
Birdsall, N.; Hecht, R. 1995. Swimming against the tide: strategies for improving equity in health. Inter-American Development Bank, Human Resources Development and Operations Policy (HRO), Washington, DC, USA. 23 pp.
Birn, A.; Zimmerman, S.; Garfield, R. 2000. To decentralize or not to decentralize, is that the question? Nicaraguan health policy under structural adjustment in the 1990s. International Journal of Health Services, 30(1), 111–128.
Blas, E.; Hearst, N. 2002. Health sector reform and equity – learning from evidence? Health Policy and Planning, 17, 1–4.
Bloom, G.; Lucas, H. 2000. Health and poverty in sub-Saharan Africa. IDS Working Paper no. 103, Institute of Development Studies, Brighton, UK. 44 pp.
Bloom, G.; Standing, H. 2001. Pluralism and marketisation in the health sector: meeting health needs in contexts of social change in low and middle income countries. IDS Working Paper no. 136, Institute of Development Studies, Brighton, UK. 36 pp.
Bossert, T. 1998. Analyzing the decentralization of health systems in developing countries: decision space, innovation and performance. Social Science and Medicine, 47(10), 1513–1527.
Bossert, T.; Hsiao, W.; Barrera, M.; Alarcon, L.; Leo, M.; Casares, C. 1998. Transformation of ministries of health in the era of health reform: the case of Colombia. Health Policy and Planning, 13(1), 59–77.
Bourguignon, F.; Morrisson, C. 1992. Adjustment and equity in developing countries: a new approach. Development Centre of the Organisation for Economic Co-operation and Development. OECD Publications and Information Centre, Washington, DC, USA. 111 pp.
Breman, A.; Shelton, C. 2001. Structural adjustment and health: a literature review of the debate, its role-players and presented empirical evidence. Commission on Macroeconomics and Health (CMH), Working Paper Series, no. WG6:6, World Health Organization, Cambridge, MA, USA. 53 pp.
Buchmann, C. 1996. The debt crisis, structural adjustment and women's education: implications for status and social development. International Journal of Comparative Sociology, 37, 5–30.
Burnham, G.M.; Pariyo, G.; Galiwango, E.; Wabwire-Mangen, F. 2004. Discontinuation of cost sharing in Uganda. Bulletin of the World Health Organization, 82(3), 187–195.
Campos-Outcalt, D.; Kewa, K.; Thomason, J. 1995. Decentralization of health services in Western Highlands Province, Papua New Guinea: an attempt to administer health service at the subdistrict level. Social Science and Medicine, 40(8), 1091–1098.
Chawla, M.; Ellis, R.P. 2000. The impact of financing and quality changes on health care demand in Niger. Health Policy and Planning, 15(1), 76–84.
Collins, C. 1996. Decentralization. In Janovsky, K., ed., Health policy and systems development: an agenda for research. World Health Organization, Geneva, Switzerland. pp. 161–178.
Collins, C.; Green, A.; Hunter, D. 1999. Health sector reform and the interpretation of policy context. Health Policy, 47(1), 69–83.
Cornia, G.A. 2001. Social funds in stabilization and adjustment programmes: a critique. Development and Change, 32, 1–32.
Cornia, G. A.; Jolly, R.; Stewart, F., ed. 1987. Adjustment with a human face. Clarendon Press, Oxford, UK. 340 pp.
Cornia, G.A.; van der Hoeven, R.; Mkandawire, T. 1992. Africa's recovery in the 1990s: from stagnation and adjustment to human development. St. Martin's Press, New York, NY, USA. 375 pp.
Cornu, A.; Massamba, J.P.; Traissac, P.; Simondon, F.; Villeneuve, P.; Delpeuch, F. 1995. Nutritional change and economic crisis in an urban Congolese community. International Journal of Epidemiology, 24(1), 155–164.
Costello, A.; Watson, F.; Woodward, D. 1994. Human face or human facade? Adjustment and the health of mothers and children. Centre for International Child Health, Institute of Child Health, London, UK. 68 pp.
Creese, A.L. 1994. User charges for health care: a review of recent experience. Health Policy and Planning, 6(4), 309–319.
de Béthune, X.; Alfani, S.; Lahaye, J.P. 1989. The influence of an abrupt price increase on health service utilization: evidence from Zaire. Health Policy and Planning, 4(1), 76–81.
de Ferranti, D. 1985. Paying for health services in developing countries: an overview. Staff Working Paper no. 721. World Bank, Washington, DC, USA. 140 pp.
Diop, F.; Yazbeck, A.; Bitran, R. 1995. The impact of alternative cost recovery schemes on access and equity in Niger. Health Policy and Planning, 10(3), 223–240.
ECLAC (Economic Commission for Latin America and the Caribbean). 1996. Preliminary overview of the economy of Latin America and the Caribbean 1996. United Nations, Santiago, Chile. 53 pp.
Elson, D. 1992. From survival strategies to transformation strategies: women's needs and structural adjustment. In Beneria, L.; Feldman, S., ed., Unequal burden: economic crises, persistent poverty, and women's work. Westview Press, Boulder, CO, USA. pp. 26–49.
Ensor, T.; Bich San, P. 1996. Access and payment for health care: the poor of northern Vietnam. International Journal of Health Planning and Management, 11(1), 69–83.
Evans, R.; Barer, M. 1994. Producing health, consuming health care. In Evans, R.; Barer, M.; Marmor, T., ed., Why are some people healthy and others not? The determinants of health of populations. Aldine de Gruyter, New York, NY, USA. pp. 27–66.
Fabricant, S.J.; Kamara, C.W.; Mills, A. 1999. Why the poor pay more: household curative expenditures in rural Sierra Leone. International Journal of Health Planning and Management, 14(4), 179–199.
Ferroni, M.; Kanbur, R. 1990. Poverty-conscious restructuring of public expenditures. Working Paper no. 9, Social Dimensions of Adjustment and Development Unit, Africa Technical Unit. World Bank, Washington, DC, USA. 34 pp.
Fournier, P.; Haddad, S.; Mantoura, P. 2001. Réformes des systèmes de santé dans les pays en développement: l'irrésistible emprise des agences internationales et les dangers de la pensée unique. In Rainhorn, J.; Burnier, M., ed., La santé au risque du marché. Incertitudes à l'aube du XXIe siècle. Presses Universitaires de France, Paris, France. pp. 71–84.
Garcia, G.; Oliveira, O. 1997. Economic recession and changing determinants of women's work. In Tapinos, G.; Mason, A.; Bravo, J., ed., Demographic responses to economic adjustment in Latin America. Clarendon Press, Oxford, UK. pp. 229–251.
Garenne, M.G.; Gakusi, E. 2000. Health effects of structural adjustment programs in sub-Saharan Africa. Working paper, French Center for Population and Development Studies (CEPED), Paris, France. 26 pp.
Gershman, J.; Irwin, A. 2000. Getting a grip on the global economy. In Irwin, A.; Kim, J.Y.; Gershman, J.; Millen, J.V., ed., Dying for growth: global inequality and the health of the poor. Common Courage Press, Monroe, ME, USA. pp. 11–43.
Gertler, P.; Van der Gaag, J. 1990. The willingness to pay for medical care: evidence from two developing countries. Johns Hopkins University Press, Baltimore, MD, USA. 139 pp.
Gilson, L.; Kilima, P.; Tanner, M. 1994. Local government decentralization and the health sector in Tanzania. Public Administration and Development, 14, 451–477.
Gilson, L.; Mills, A. 1995. Health sector reforms in sub-Saharan Africa: lessons of the last 10 years. In Berman, P.A., ed., Health sector reform in developing countries: making health development sustainable. Harvard University Press, Boston, MA, USA. pp. 277–316.
Gilson, L.; Russell, S.; Buse, K. 1995. The political economy of user fees with targeting: developing equitable health financing policy. Journal of International Development, 7(3), 369–402.
Gilson, L.; Sen, P.D.; Mohammed, S.; Mujinja, P. 1994. The potential of health sector non-governmental organizations: policy options. Health Policy and Planning, 9(1), 14–24.
Gindling, T.H. 1993. Women's wages and economic crisis in Costa Rica. Economic Development and Cultural Change, 41(2), 277–297.
Griffin, C. 1992. Welfare gains from user charges for government health services. Health Policy and Planning, 7(2), 177–180.
Grootaert, C. 1994. Education, poverty, and structural change in Africa: lessons from Côte d'Ivoire. International Journal of Educational Development, 14(2), 131–142.
Haddad, S.; Fournier, P. 1995. Quality, cost and utilization of health services in developing countries. A longitudinal study in Zaire. Social Science and Medicine, 40(6), 743–753.
Hanson, K.; Berman, P. 1998. Private health care provision in developing countries: a preliminary analysis of levels and consumption. Health Policy and Planning, 13(3), 195–211.
Haran, D. 1998. Health sector reform. Journal of Epidemiology and Community Health, 52(12), 768–769.
Holley, J. 1995. Estudio de descentralización de la gestion de los servicios de salud. Territorio de Capinota, Bolivia. Latin American Health and Nutrition Sustainability Project. University Research Corporation, Washington, DC, USA.
Hongoro, C.; Kumaranayake, L. 2000. Do they work? Regulating for-profit providers in Zimbabwe. Health Policy and Planning, 15(4), 368–377.
Hsiao, W.C. 1995. Abnormal economics in the health sector. In Berman, P.A., ed., Health sector reform in developing countries: making health development sustainable. Harvard University Press, Boston, MA, USA. pp. 161–179.
Husain, I. 1994. Structural adjustment and long-term development in sub-Saharan Africa. In van de Hoeven, R.; van de Kraaij, F., ed., Structural adjustment and beyond in sub-Saharan Africa: research and policy issues. James Curry, London, UK. 290 pp.
ILO (International Labour Organization). 1991. Employment policies in the economic restructuring of Latin America and the Caribbean. Tripartite Symposium on Structural Adjustment and Employment in Latin America and the Caribbean. WEP 1-4-07 (Doc. 2), Geneva, Switzerland.
Iriart, C.; Merhy, E.E.; Waitzkin, H. 2001. Managed care in Latin America: the new common sense in health policy reform. Social Science and Medicine, 52(8), 1243–1253.
Jayarajah, C.; Branson, W. 1995. Structural and sectoral adjustment: World Bank experience, 1980–92. World Bank, Washington, DC, USA. 358 pp.
Jayarajah, C.; Branson, W.; Sen, B. 1996. Social dimensions of adjustment: World Bank experience, 1980–93. Operations Evaluation Division, World Bank, Washington, DC, USA. 218 pp.
Jeppsson, A. 2001. Financial priorities under decentralization in Uganda. Health Policy and Planning, 16(2), 187–192.
Kaddar, M.; Schmidt-Ehry, B.; Stierle, F.; Tchicaya, A. 1997. Indigence et accès aux soins de santé en Afrique sub-saharienne: situation et perspective d'action. Gesellschaft für Technische Zusammenarbeit (GTZ), Eschborn, Germany. 65 pp.
Kakwani, N. 1995. Structural adjustment and performance in living standards in developing countries. Development and Change, 26(3), 469–502.
Kanji, N. 1989. Charging for drugs in Africa: UNICEF's Bamako Initiative. Health Policy and Planning, 4(2), 110–120.
Kanji, N.; Kanji, N.; Manji, F. 1991. From development to sustained crisis: structural adjustment, equity and health. Social Science and Medicine, 33(9), 985–993.
Kjellstrom, T.; Koplan, J.P.; Rothenberg, R.B. 1992. Current and future determinants of adult ill-health. In Feachem, R.G.A.; Kjell-strom, T.; Murray, C.J.L.; Over, M.; Philips, M.A., ed., The health of adults in the developing world. Oxford University Press, New York, NY, USA. pp. 209–260.
Knippenberg, R.; Reinke, W.; Hopwood, I. 1997. Sustainability of primary health care including expanded program of immunizations in Bamako Initiative programs in West Africa: an assessment of 5 years' field experience in Benin and Guinea. International Journal of Health Planning and Management, 12(suppl.), S9–S28.
Kolehmainen-Aitken, R.-L. 1998. Decentralization and human resources: implications and impact. Human Resource Development Journal, 2(1), 1–16.
Kumaranayake, L. 1997. The role of regulation: influencing private sector activity within health sector reform. Journal of International Development, 9(4), 641–649.
Kumaranayake, L.; Lake, S.; Mujinja, P.; Hongoro, C.; Mpembeni, R. 2000. How do countries regulate the health sector? Evidence from Tanzania and Zimbabwe. Health Policy and Planning, 15(4), 357–367.
Laffont, J.-J.; Tirole, J. 1993. A theory of incentives in procurement and regulation. MIT Press, Cambridge, MA, USA. 705 pp.
Lam, D.; Levison, D. 1997. Structural adjustment and family labour supply in Latin America. In Tapinos, G.; Mason, A.; Bravo, J., ed., Demographic responses to economic adjustments in Latin America. Clarendon Press, Oxford, UK, pp. 201–228.
Larbi, G.A. 1998. Institutional constraints and capacity issues in decentralizing management in public services: the case of health in Ghana. Journal of International Development, 10(3), 377–386.
Laurell, A.C.; Arellano, O.L. 1996. Market commodities and poor relief: the World Bank proposal for health. International Journal of Health Services, 26(1), 1–18.
Lavy, V.; Germain, J.-M. 1994. Quality and cost in health care choice in developing countries. Living Standards Measurement Study, Working Paper no. 105. World Bank, Washington, DC, USA. 48 pp.
Leighton, C.; Diop, F. 1995. Protecting the poor in Africa–impact of means testing on equity in the health sector in Burkina Faso, Niger, and Senegal. Health Financing and Sustainability (HFS) Project, Abt Associates Inc., Bethesda, MD, USA. 29 pp.
Litvack, J.I.; Bodart, C. 1993. User fees plus quality equals improved access to health care: results of a field experiment in Cameroon. Social Science and Medicine, 37(3), 369–383.
Loewenson, R. 1993. Structural adjustment and health policy in Africa. International Journal of Health Services, 23(4), 717–730.
_____ 1994. World Bank policy and health policy in Africa. Paper presented at the Economic Policy and Health Network Conference, 27–29 June 1994, Johannesburg, South Africa, 12 pp.
Lundy, P. 1996. Limitations of quantitative research in the study of structural adjustment. Social Science and Medicine, 42(3), 313–324.
McPake, B. 1993. User charges for health services in developing countries: a review of the economic literature. Social Science and Medicine, 36(11), 1397–1405.
McPake, B.; Machray, C. 1997. International comparisons of health sector reform: towards a comparative framework for developing countries. Journal of International Development, 9(4), 621–630.
McPake, B.; Ngalande-Banda, E. 1994. Contracting out of health services in developing countries. Health Policy and Planning, 9(1), 25–30.
Minujin, A. 1992. From "secondary workers" to breadwinners: poor and non-poor women facing the crisis. Paper presented at the IUSSP (International Union for the Scientific Study of Population) Seminar on Demographic Consequences of Structural Adjustment in Latin America, 29 September–2 October 1992, Ouro Preto, Brazil.
Mills, A. 1998. To contract or not to contract? Issues for low and middle income countries. Health Policy and Planning, 13(1), 32–40.
Mohindra, K. 1999. Socio-economic determinants of child growth in five developing countries: implications for health and development. Unpublished MSc. thesis. University of Northern British Colombia, Department of Community Health Science, Prince George, BC, Canada. 112 pp.
Morley, S.A. 1995. Poverty and inequality in Latin America: the impact of adjustment and recovery in the 1980s. Johns Hopkins University Press, Baltimore, MD, USA. 222 pp.
Musgrove, P. 1986. The impact of the economic crisis on health and health care in Latin America and the Caribbean. WHO Chronicle, 40(4), 152–157.
Mwabu, G.; Wang'ombe, J. 1995. User charges in Kenya: health service pricing reforms in Kenya: 1989–93. International Health Policy Program, Washington, DC, USA. 16 pp.
Newbrander, W.; Collins, D.; Gilson, L. 2000. Ensuring equal access to health services: user fee systems and the poor. Management Sciences for Health, Boston, MA, USA. 187 pp.
Nolan, B.; Turbat, V. 1995. Cost recovery in public health services in sub-Saharan Africa. EDI Technical Materials Report no. 14520, World Bank, Washington, DC., USA. 118 pp.
Ogunbekun, I.; Ogunbekun, A.; Orobaton, N. 1999. Private health care in Nigeria: walking the tightrope. Health Policy and Planning, 14(2), 174–181.
PAHO (Pan American Health Organization). 1998. The response of the health systems. In Health in the Americas. PAHO, Washington, DC, USA. pp. 210–337.
Pannarunothai, S.; Mills, A. 1997. The poor pay more: health-related inequality in Thailand. Social Science and Medicine, 44(12), 1781–1790.
Peabody, J.W. 1996. Economic reform and health sector policy: lessons from structural adjustment programs. Social Science and Medicine, 43(5), 823–35.
Prud'homme, R. 1995. The dangers of decentralization. The World Bank Research Observer, 10(2), 201–220.
Reimers, F. 1994. Education and structural adjustment in Latin America and sub-Saharan Africa. International Journal of Educational Development, 14(2), 119–129.
Reimers, F.; Tiburcio, L. 1993. Education, adjustment and reconstruction: options for change. UNESCO, Paris, France. 116 pp.
Russell, S. 1996. Ability to pay for health care: concepts and evidence. Health Policy and Planning, 11(3), 219–237.
Sahn, D.E. 1992. Public expenditures in sub-Saharan Africa during a period of economic reform. World Development, 20(5), 673–693.
Sahn, D.E.; Bernier, R. 1995. Has structural adjustment led to health sector reform in Africa? In Berman, P.A., ed., Health sector reform in developing countries: making health development sustainable. Harvard University Press, Boston, MA, USA. pp. 247–275.
Sahn, D.E.; Dorosh, P.; Younger, S. 1997. Structural adjustment reconsidered: economic policy and poverty in Africa. Cambridge University Press, New York, NY, USA. 304 pp.
Sahn, D.E.; Younger, S.D.; Genicot, G. 2003. The demand for health care services in rural Tanzania. Oxford Bulletin of Economics and Statistics, 65(2), 241–260.
Sauerborn, R.; Nougtara, A.; Hein, M.; Diesfeld, H.J. 1996. Seasonal variations of household costs of illness in Burkina Faso. Social Science and Medicine, 43(3), 281–290.
Schoepf, B.G.; Schoepf, C.; Millen, J. 2000. Theoretical therapies, remote remedies: SAPs and the political ecology of poverty and health in Africa. In Irwin, A.; Kim, J.Y.; Gershman, J.; Millen, J.V., ed., Dying for growth: global inequality and the health of the poor. Common Courage Press, Monroe, ME, USA. pp. 91–125.
Sen, B. 1997. Health and poverty in the context of country development strategy: a case study on Bangladesh. Macroeconomics, Health and Development Series, no. 26. World Health Organization, Geneva, Switzerland. 25 pp.
Sen, K.; Koivusalo, M. 1998. Health care reforms and developing countries–a critical overview. International Journal of Health Planning and Management, 13, 199–215.
Shaw, R.P.; Griffin, C.C. 1995. Financing health care in sub-Saharan Africa through user fees and insurance. World Bank, Washington, DC, USA. 99 pp.
Simms, C.; Rowson, M.; Peattie, S. 2001. The bitterest pill of all: the collapse of Africa's health systems. Medact and Save the Children Fund, London, UK. 27 pp.
Smithson, P.; Asamoa-Baah; Mills, A. 1997. The case of the health sector in Ghana experience. University of Birmingham Series, The role of government in adjusting economies, Paper 26. University of Birmingham, Birmingham, UK. 48 pp.
Söderlund, N.; Tangcharoensathien, V. 2000. Health sector regulation – understanding the range of responses from government. Health Policy and Planning, 15(4), 347–348.
Stewart, F. 1989. Recession, structural adjustment and infant health: the need for a human face. Transactions of the Royal Society of Tropical Medicine and Hygiene, 83, 30–31.
_____ 1992. Can adjustment programmes incorporate the interests of women? In Afshar, H.; Dennis, C., ed., Women and adjustment policies in the Third World. Macmillan, Hong Kong. 291 pp.
Stewart, F.; Van de Geest, W. 1995. Adjustment and social funds: political panacea or effective poverty reduction? In Stewart, F., ed., Adjustment and poverty: options and choices. Routledge, New York, NY, USA. pp. 108–137.
Tardanico, R.; Larín, R.M. 1997. Global restructuring, employment and social inequality in urban Latin America. North-South Center Press, Coral Gables, FL, USA. 320 pp.
Tollman, S.; Schopper, D.; Torres, A. 1990. Health maintenance organizations in developing countries: what can we expect? Health Policy and Planning, 5(2), 149–160.
Turshen, M. 1999. Privatizing health services in Africa. Rutgers University Press, Piscataway, NJ, USA. 200 pp.
Uplekar, M.W. 2000. Private health care. Social Science and Medicine, 51(16), 897–904.
UNICEF. 1984. The state of the world's children, 1984. Oxford University Press, New York, NY, USA. 126 pp.
_____ 1992. The Bamako Initiative: progress report, UNICEF Executive Board 1992 session. E/ICEF/1992/L.6, 3 February 1992. UNICEF, New York, NY, USA. 40 pp.
Van der Gaag, J.; Barham, T. 1998. Health and health expenditures in adjusting and non-adjusting countries. Social Science and Medicine, 46(8), 995–1009.
Van der Geest, S.; Macwan'gi, M.; Kamwanga, J.; Mulikelela, D.; Mazimba, A.; Mwangelwa, M. 2000. User fees and drugs: what did the health reforms in Zambia achieve? Health Policy and Planning, 15(1), 59–65.
Wagstaff, A. 2001. Poverty and health. Commission on Macroeconomics and Health Working Paper no. WG1:5. World Health Organization, Geneva, Switzerland. 52 pp.
Waddington, C.; Enyimayew, K. 1989. A price to pay. Part 1: the impact of user charges in Ashatin-Akim district, Ghana. International Journal of Health Planning and Management, 4, 17–47.
Watkins, K. 1997. Cost-recovery and equity in the health sector: issues for developing countries. Paper presented at World Institute for Development Economics Research (WIDER) Project on Provision and Financing of Public Goods in Developing Countries, 27 February 1997, London, UK. Oxfam, Oxford, UK, and Ireland Policy Department. 31 pp.
Woodward, D. 1992. Debt, adjustment and poverty in developing countries: national and international dimensions of debt and adjustment in developing countries, vol. 1. Save the Children Foundation (SCF) and Pinter Publishers, London, UK. 300 pp.
World Bank. 1987. Financing health services in developing countries: an agenda for reform. World Bank, Washington, DC, USA. 108 pp.
_____ 1990. World development report 1990: poverty. Oxford University Press, New York, NY, USA. 272 pp.
_____ 1993. World development report: investing in health. Report no. 12183. Washington, DC, USA. 344 pp.
_____ 1994a. Adjustment in Africa: reforms, results, and the road ahead. Oxford University Press, New York, NY, USA. 304 pp.
_____ 1994b. Better health in Africa: experience and lessons learned. Report no. 13488. World Bank, Washington, DC, USA. 304 pp.
_____ 1996. World development report: from plan to market. Report no. 15892. World Bank, Washington, DC, USA. 260 pp.
_____ 1997. World development report: the state in a changing world. Oxford University Press, New York, NY. 354 pp.
Yach, D.; von Schirnding, Y.E.R. 1994. Towards a higher priority for health on the development agenda. Public Health Review, 22, 339–374.
Yoder, R.A. 1989. Are people willing and able to pay for health services? Social Science and Medicine, 29(1), 35–42.
Producing a solid base of factual evidence to use in assessing the impacts of reforms and adjustment policies is a complex undertaking fraught with difficulties. These have to do mainly with the attribution, in an uncontrolled context, of cause–effect relationships between public interventions and dimensions of health and well-being that are themselves the products of multiple processes. There are also difficulties inherent in isolating the interventions from their presumed consequences, as well as in generalizing observations from very specific contexts and environments. The issue of methodology is therefore central to any evaluation of public policy. There would no doubt be considerable divergence of opinion on the social consequences of structural adjustment even without these difficulties, although the terms of the debate would probably be greatly simplified.
Before introducing the MAPHealth project, we review briefly, in the following section, the methodological approaches currently used to evaluate the effects of macroeconomic adjustment on the health-care sector.1 We then examine the challenges in identifying counterfactual situations and in choosing appropriate measurement tools for evaluating the impacts of public policies. We draw lessons in methodology from studies carried out over the past 15 years on the effects on health of structural adjustment that inform and enrich the final design and methodology of the MAPHealth project, as described in the latter part of the chapter.
Studies on structural adjustment policies in relation to health have been carried out in single as well as multiple countries, with single-country studies being more frequent. Table 1 presents the main features of each approach and the elements that may vary among studies. For each type, examples are drawn from the literature. Single case studies are often based on before–after approaches, i.e., pretest–posttest non-experimental designs, and some use simulation approaches based on modeling techniques such as CGE (computable general equilibrium). Interrupted time series are eventually part of the analyses. Both primary and secondary data may be used, and we found an increasing use of primary data in the later 1990s. Some surveys, such as Patricia Lundy's (1996) in Jamaica, use qualitative approaches, employing semi-structured interviews and participant observation to assess the quality of health-care services. A well-known case study in Zimbabwe used complex, multiple methods that combined quantitative and qualitative data with household and clinic data (Bijlmakers et al. 1998).
In multiple-country studies, there are many factors to be considered, including the number of countries, the specificity of each case, and the types of generalizations that can be made. For this discussion, we have categorized multiple-country studies into two types. The first, Type I, compare outcomes levels between adjusting and non-adjusting countries. Typically, a limited number of outcomes is examined among a large number of countries. The evaluation is broad, in that it does not look at specific policies or macroeconomic measures, but at adjustment policies as a whole.
TABLE 1. COMPARISON OF RESEARCH APPROACHES
Approach |
Typical features |
Differences among studies |
Examples |
Case study |
Before–after Primary and secondary data Quantitative and/or qualitative Several specific outcomes Sponsored by international organizations, universities Low to high cost |
Use of modeling and econometric tools (e.g. CGE–computable general equilibrium) |
Bijlmakers, Bassett, Sanders 1996 Cornu 1995 Lundy 1996 |
Multiple country studies Type I |
With–without Posttest-only designs Secondary data from countries and IFI sources Quantitative Limited number of outcomes Sponsored by IFIs Low cost |
Number of countries Regional diversity Statistical tools Outcome types |
Garenne and Gakusi 2000 Jayarajah et al. 1996 Kakwani 1995 Makinen et al. 2000 Mwabu 1996 Noorbakhsh 1998 Van der Gaag and Barham 1998 |
Multiple country studies Type II |
Shared methodology between countries Complex designs Primary and secondary data Quantitative and qualitative Multiple outcomes Sponsored by international organizations, universities Large teams High cost |
Number of countries Regional diversity Use of qualitative techniques Level and type of collaboration between teams Household surveys Health facility surveys |
Bourguignon and Morrisson 1992 Hanson 1999 Newbrander et al. 2000 Reed 1999 Sahn et al. 1997 |
Based on posttest-only designs, these studies draw upon secondary data taken mostly from international financial institution (IFI) sources. These are low-cost studies conducted by individuals or single teams from IFIs, other organizations, or universities. Often, they aggregate countries from widely divergent contexts, making generalizations difficult. Typical studies of this approach include Mwabu's (1996) cross-country regression analysis of 103 countries. The author used aggregated household data from the World Bank, the International Monetary Fund (IMF), and the United Nations, performing econometric estimations to complete missing data from these sources. Four outcome variables were included in the analysis: life expectancy, child mortality, infant mortality, and maternal mortality. Another example is Van der Gaag and Barham's (1998) study of health and health expenditures in adjusting and non-adjusting developing countries.
Type II multiple-case studies also compare countries but adopt a different approach. These studies have a smaller number of cases (each country being a case) but analyses are more in-depth, and specific aspects of public policies are analyzed. Methodologies are often multiple and more complex than Type I studies. Study designs are shared by the cases, and efforts are made to compile and synthesize results. These studies often rely on before–after observations, collecting primary as well as secondary data. Extensive investments in time and money are needed, particularly for initiatives involving a high level of country collaboration. Country teams often conduct their own research, with the assistance of a technical advisory committee. The topics covered are variable. For example, Sahn and his colleagues (1997) examined trade, exchange rate, fiscal, and agricultural policy outcomes related to macroeconomic adjustment policies (MAPs) among 10 African countries, using CGE modeling. A study by UNICEF coordinated an eight-country team effort to examine health sector reform in Africa, where teams used qualitative approaches (Hanson 1999). Reed's comparative survey (1999) estimated structural adjustment impacts on the environment. These three studies resulted in in-depth analysis of a relatively small number of cases that synthesized results and extracted common policy lessons. Other examples include a study of user fees and access to health care among five countries, using a common methodology under the guidance of a technical advisory committee (Newbrander et al. 2000). This study also demonstrated the inter-country variation and the role of local context, while providing a synthesis of results and policy recommendations.
The primary challenge is to assess appropriately which effects are attributable to the reforms, i.e., given a set of outcomes, to differentiate between the situation post-intervention and the situation that would have prevailed had the intervention not occurred (Shadish et al. 2001). It is a relatively simple task to assess the post-intervention situation. On the other hand, one can never observe, nor accurately measure, what the situation might have been without the intervention—the counterfactual situation that "remains forever in the darkness of the unfulfilled" (Thomas Hardy, quoted by Mohr, 1995, p. 4.) We are thus constrained to develop a source of counterfactual inference to serve as a plausible foundation from which to determine those changes that might be attributed to the policies implemented. We will generally find this source in one or another of the following situations: (i) the situation as it existed in the same place, before intervention (the before situation); (ii) the situation observed elsewhere, where the intervention did not occur (the situation without); (iii) a virtual situation, neither observed before nor elsewhere, but rather simulated (the as-if situation); and (iv) a virtual situation derived from repeated observation of the evolution of states similar to that in which the intervention was implemented, that represents the natural evolution of the situation (the situation as it might have been).
The "before" situation is that which prevailed immediately before the implementation of the changes under evaluation. For every indicator of a presumed outcome, the effect of a policy is derived by comparing the indicator before and after the policy's implementation. Thus, the effect of the adjustment on the health of a population would correspond to the average of the changes in health status of individuals in that population. This approach forms the basis of the simplest quasi-experimental evaluations, based on before–after observations but without control groups. These types of assessments are very common because of their appealing simplicity and because of the great advantage of using as the source of counterfactual inference the same as that used to assess the situation after the intervention. However, these advantages are also their greatest weakness, in that their internal validity is poor and their findings are generally considered to be less robust.
The primary threat to internal validity lies in the possibility that changes that occurred between the two points of observation are the result either of events other than the intervention itself (history bias) or of the natural evolution of the phenomena under consideration (maturation bias) (Campbell and Stanley 1963; Cook and Campbell 1979). These assessments also do not take into consideration changes induced by exogenous factors—political, economic, climatic, social, etc.—that inevitably arise, nor the long-standing tendencies of existing natural cumulative processes, whether social or physiological. Therefore, although the approach may be appealing in its simplicity, it cannot, in itself, guarantee acceptable scientific results (Bourgignon and Morrisson 1992).
In this approach, for a given indicator, the effect is deduced from differences observed between countries that have been exposed to reforms and those that have not.3 The appeal of this approach is that it offers a control group of non-adjusting countries to test the counterfactual argument (Costello et al. 1994) and also offers the possibility, at least in principle, of generating more generalized findings than single case studies. Even here, there are major weaknesses in the design. The internal validity is undermined by the heterogeneity of the exposed and non-exposed groups, the inability to control for selection bias, and the excessive simplification of comparisons based on aggregate data.
The first problem encountered is that situations differ greatly among countries making up any one particular group, whether they are the early adjusters, the late adjusters, the non-adjusters, or any other grouping of countries. Among the adjusting countries, there are differences in the conditions leading to adjustment, the policies implemented, the rhythms of implementation, and, of course, the economic and sociopolitical contexts in which they occur. Thus, in aggregating the data from these heterogeneous groups of countries, the specific realities of each are masked by the resulting forced homogeneity, as are the unique features of their reforms and the conditions that prevailed at their implementation (Sahn et al. 1997). In other words, many key endogenous factors are overlooked that determine, in large measure, the economy, the level of well-being, and the living conditions of the country. In so doing, there is the risk that the analysis of the presumed effects of reform will be biased.
Another problem may be found in the conditions under which countries are designated as exposed-adjusters and controls-non-adjusters and the resulting selection biases that may emerge from the start (P selection bias) or during the course of implementation of the reforms (Q selection bias). The adjusting countries have had to deal with critical situations from which most non-adjusting countries have been spared. Often they have experienced turbulent cycles of crisis and reform. Their starting conditions and their particular histories set them apart from those who have never entered into structural adjustments. As such, the two groups of countries present distinctly different profiles with regard to external conditions and the outcomes of interest that tend to become accentuated during reform. It is not entirely reasonable to believe, as this type of approach would require, that adjusting and non-adjusting countries differ only in terms of their policies and the adjustment measures implemented; therefore, it is difficult to see the non-adjusting countries as constituting a plausible counterfactual base of reference. Additionally, selection biases can be reinforced by the non-availability of data. It frequently happens, as Reed found in his multi-country survey on adjustment and environment, that those countries for which there is little or no data have particular profiles that in themselves are associated with specific levels of health or well-being; very poor countries, or countries under reconstruction where the public statistical apparatus is not functioning, are also likely to be performing less well in terms of human development (Reed 1999).
A third problem is the frequent existence of contamination bias. The fact that certain countries have not received formal adjustment credits from IFIs does not mean they have not, or would not, put in place policies or measures emanating from the same reasoning. During the 1980s, for example, Burkina Faso undertook a so-called auto-adjustment policy that is comparable in many respects to those promoted by the Bretton Woods institutions.
In the end, what are the consequences, in these multi-country studies based on aggregate data, of such threats to internal validity? For some authors, these approaches result in reducing the sensitivity of the results, as well as introducing systematic bias "in a positive direction, so that favourable effects are overstated and unfavourable results are understated" (Costello et al. 1994, p. 9). This assertion is supported in a recent review of studies on health and adjustment, in which the authors demonstrate, notably, that in multi-country studies of the with–without type, there is less tendency to highlight negative impacts of adjustment on health than in single-country studies, that more often use a before–after design (Breman and Shelton 2001).
To produce acceptable evidence using with–without approaches, a number of conditions would have to come together that are difficult to achieve: a significant sample of countries regrouped into homogeneous categories; natural histories that are similar; data of sufficiently good quality; and information that is both longitudinal and transversal (Bourguignon and Morrisson 1992, Lundy 1996). Another condition would be that the countries' economies behave in similar ways, which is an unrealistic hypothesis (Bourguignon and Morrisson 1992). For these reasons, contrary to the IFIs, most scientific researchers look suspiciously upon these approaches based on minimalist counterfacts, which produce tenuous results (Sahn et al. 1997; Lundy 1996; Narayana and Navaneetham 1998). "While such aggregate analysis is very valuable, it does not help much in determining whether the effects that we believe are caused by structural adjustment actually happen" (Peabody 1996, p. 827).
Simulation studies attempt to compare the post-adjustment situation with that which would have prevailed if the adjustment had not occurred. Thus, these are one-group designs in which the validity is strengthened by the addition of a counterfactual situation created through modeling. With the development of CGE modeling techniques, this approach is now commonly used to evaluate or predict the consequences of macroeconomic policies. The study may focus on a single country, or, if there is sufficient available data, on multiple countries, as in the OECD's international study (Bourguignon and Morrisson 1992). The popularity of simulation is explained by its advantages. First, it is easier to control biases that would threaten internal validity. The result is somewhat like having a with–without study that controls for history and maturation biases, but in which, furthermore, the country being studied is matched to itself. As in a before–after study, this matching would eliminate the problem of designating exposed and control groups, thereby limiting selection biases. Next, the evaluation possibilities are nearly infinite because once the causal structure has been defined, it is possible not only to define a counterfactual scenario, but also to elaborate upon it and evaluate any number of alternative scenarios. Finally, these analyses require only secondary data and can, insofar as the base data is available, be carried out at a limited cost.
It is important to keep in mind that any simulation is based upon a causal structure derived from a series of hypotheses about the mechanisms of macro-meso-micro transmission. Therefore, predictions are necessarily limited by the validity of these hypotheses and the ability of the researchers to produce a good scaled-down model of reality (Bourguignon and Morrisson 1992). This requires a deep understanding of the functioning of causal mechanisms and, more generally, of the world being modeled, that is really only possible in areas where there is a substantial base of empirical knowledge (Rossi and Freeman 1993). This is true, to a point, for general modeling of an economy. However, it is less true when it comes to the social sector, and it is not true at all when it comes to modeling functions of individual and collective production in health or the behaviour of the health services market. Therefore, although promising, simulations are still not very useful for evaluating the effects of public policy on health.
The fourth approach to building a counterfactual scenario is to base it upon a series of historical observations. Using appropriate statistical techniques, the impact is deduced by comparing several time series before and after the intervention (interrupted time series). The main advantage to this approach is in the enriched information that makes it possible to construct a more elaborate and credible counterfactual scenario. Biases that could undermine internal validity are minimized. Maturation bias is controlled, and even if historical bias is still possible, it is significantly reduced in comparison with before–after designs. In practice, historical bias can be nearly completely eliminated through the use of comparative time series, which are more demanding but always possible (Mohr 1995). In this case, the analysis is based on the comparative analysis of several chronological series, each related to a specific result.
One of the primary advantages of time series is their capacity to integrate temporal dimensions into the analysis. In the first place, they make it possible to take into account the temporal nature of the response to the intervention. Some interventions will have immediate, or near-immediate, impacts, as with devaluation and the price of imported medicines. However, this is not always the case, and it is very useful to be able to take into account delayed responses when studying phenomena such as health. Sometimes the response is intrinsically delayed because the action mechanisms take time, or the changes require a long exposure, or several other prerequisite mechanisms or mediators need to be activated first. In other cases, the response is shifted because of adaptive strategies, as in the case, for example, of households reallocating their resources to protect themselves from the negative effects on their service consumption of a reduction in revenue or in the number of professionals (Costello et al. 1994). As long as these adaptive strategies work, no change in outcomes may be observed. The first changes emerge when the adaptive strategies become ineffective. In some cases, transformation occurs but the changes are modest or progressive and the effects are not immediately measured or even measurable. Finally, when it comes to health, past investments in infrastructure, education, vaccination, etc., because of their long-term influence, can obstruct, modify, or delay changes that might result from any modification of the economic environment and public policies (Grosh 1990). Certain legacies are long-lasting and even irreversible, such as vaccinal immunity, and are therefore not susceptible to changes in context or policy. Other indicators, such as mortality, are known not to be very sensitive in the short or long term.
Chronological series also make it possible to test different hypotheses regarding the permanent or temporary nature, cumulative or not, of the effects. However, it is not always possible to draw full advantage of these possibilities and assessing the medium- and long-term effects of reforms remains difficult. First, the reforms are often undertaken intermittently (Reed 1999) and precise temporal reference points are not always available for analyzing response times or cause–effect relationships, nor for projecting a counterfactual scenario. To be fully exploitable statistically, temporal series must be sufficiently long and based on a large number of observation points, both before and after the intervention. Nevertheless, even short time series with only summary statistical analyses are preferable to simplistic before–after designs, and it is therefore unfortunate that the use of chronological series for evaluating public policies is still somewhat limited. Finally, the researcher generally has only a limited number of time series available, most often built upon routine administrative data whose reliability is questionable.
We can see that none of the counterfactual scenarios is completely satisfactory and that recourse to research designs combining multiple approaches is inevitable. There are also practical considerations to this decision. The available quantitative indicators are often quite limited and their reliability doubtful, such that they need to be complemented with both quantitative and qualitative primary data. The problem of reliability in data is particularly acute when it comes to aggregate measures of health status. A recent review of maternal mortality in Africa demonstrated that aggregate estimates can be extremely imprecise and require great caution when used for evaluation (WHO 2004). Maternal mortality estimates varied from one to five times for the African region of the World Health Organization (390–1 500 per 100 000) and up to twice for the WHO American region (82–210 per 100 000). The same databases can give rise to very different estimates, as is demonstrated by the following example. To estimate differences in infant mortality in Burkina Faso, Lachaud used data from the 1998 Demographic and Health Survey (Lachaud 2001). He estimated the coefficient of concentration to be–0.241. The World Bank, using the same data, estimated this coefficient to be–0.06. The former group concluded that post-adjustment differences in infant mortality were significant, whereas the latter found that there were no differences!
First of all, we cannot ignore what has already been learned from two decades of studies on the social effects of adjustment, given that so much of the controversy surrounding these effects stems from the disputed internal validity of the studies (White 1996) and their tenuous results. As was pointed out by Brennan et al., methodology effects are generally present in any study of the links between adjustment and health, as "the outcome depends on the variable used in the study, the method and whether or not it is a country specific case study or a cross country study" (Breman and Shelton 2001, p. 14).
The second lesson is the importance of specifying the object of the evaluation. Reforms are generally made up of a variety of macroeconomic measures that affect different elements of the economy or the health-care system. Using generic terms related to adjustment and reform can mask that plurality, leading researchers to question whether they should be evaluating policies in general or specific components. The idea of carrying out a comprehensive evaluation of the reforms is obviously appealing. However, this is not always possible, nor even desirable: first, because being policy-specific makes creating the counterfactual easier and certainly better (Sahn et al. 1997); then, because it is more informative and more efficient to establish the links between specific measures and their presumed consequences than to determine the aggregate effects (sometimes antagonistic) of the changes, which themselves may be staggered over several years. A good example of this concept is the analysis of the effects of public policy reforms on access to medications. Certain macroeconomic measures, such as a correction of the exchange rate and a reduction in public subsidies, when combined, may converge and result in an increase in the price of medications, thus reducing access to the poorer population. Let us then imagine that, several months later, still in the context of adjustment, the government introduces other measures designed to promote the offer of generic medications at lower cost. The result will likely be a reduction in the cost of medications in the market and a lowering of financial barriers to access. An evaluation looking at the comprehensive results of the reforms would probably not render adequately the effects on access of the various measures undertaken. It is therefore more efficient to concentrate on concrete aspects of reform policies. However, this being said, when implementation is progressive, or if reforms overlap or are carried out intermittently, this dissociation is not always possible. This is illustrated in Figure 1, which shows the principal measures carried out in Thailand over the course of three crises, three macroeconomic reforms, three sectoral reforms, and numerous changes affecting the health-care system since 1974. The search for specific links between measures and consequences in this case is greatly impeded by the multiplicity of changes and the rapid pace of events and reforms.
FIGURE 1. CRISES, MACROECONOMIC REFORMS, AND HEALTH SECTOR REFORMS IN THAILAND (1974–2000)
The third lesson is that it is vital to take into account the contingent nature of reforms. They are many-sided, with components that are very diverse in terms of context, scope, and the extent of transformation they generate. The initial conditions from which the reforms emerged vary from one country to another, as do the conditions within which the interventions are carried out. Some reforms extend over more than five years; others have a more limited duration. As is shown in Figure 2, some countries have engaged in a single reform, whereas others have experienced several adjustment cycles, with measures implemented at different times and at varying levels of intensity. Finally, all reforms undergo some type of amendment once they are put in place. Governments modify them, and it is not uncommon for the initial reforms to be substantially transformed in the wake of adjustments and debt rescheduling. This has two consequences for evaluation. First, of course, this obliges us to recognize the circumstantial nature of reforms. Given that public policies are embedded within a specific reality, it is not reasonably possible to judge their results outside of this context and without taking into account their specific characteristics and their own pace of implementation, if we are to retain their essence. This therefore restricts, as we have seen, the use of comparative quantitative studies based on aggregate data. Again, this touches upon the objectives of the evaluation; we must decide whether it should consider only the impact of what was actually implemented or also the processes—and particularly any discrepancies between the initial program and the implementation of the policies. We also must consider whether to limit ourselves to those consequences expected at the start or also to take into account the unexpected effects of modifications introduced during the course of implementation. Reforms are dynamic, some
FIGURE 2. STRUCTURAL ADJUSTMENT AND HEALTH SECTOR REFORM IN THE MAPHEALTH COUNTRIES OF STUDY.
measures are unforeseen, and "the expected can give way to the unexpected" (Peabody 1996, p. 828).
The fourth lesson is that we must consider the influence not only of macroeconomic reforms, but of sectoral reforms as well. Setting aside sectoral reform, the effects of MAPs can be analyzed relatively independently. If, however,—as happens in most countries—the implementation of MAPs at any time precedes, follows, or is concurrent with reform of the health sector, this duality must be taken into account. What would we make of an analysis of the effects of adjustment on health that disregards the coexistence of a sectoral reform that, as seen in the previous chapter, greatly affects the allocation of resources, system performance, and equity of access? Conversely, any study that purports to examine the effects of health system reform would have little credibility if it were to disregard the contingent effects of adjustment on the general economy and household demand, as well as the multiple impacts—whether direct, indirect, immediate, or staggered—on the health sector of the tax system, the government's rate of expenditure, foreign exchange policies, reforms in the labour market, i.e., on the whole ensemble of macroeconomic adjustment measures. The regulation and distribution of supply, the allocation of resources, the price of services, and ultimately the production and consumption of health inputs are, in effect, largely influenced by government arbitrations and economic policies. It is therefore imperative that this double perspective (sectoral and macroeconomic) be integrated to avoid the inevitable distortions that would result from a myopic evaluation centred on purely endogenous or exogenous causes of change in the health sector.
The fifth lesson is that it is essential to be able to identify the causal pathways linking public policies with their possible consequences. Generally, measuring the micro effects of an adjustment requires modeling of macro–meso–micro links, i.e., the mechanisms by which the macroeconomic environment is internalized, through a variety of incentives and constraints, by those acting at the microeconomic level (Morrissey 1997). This is fundamental. To guide the analysis and the search for consistencies and to restrict the universe of alternative explanations, we require a system of reference that makes it possible to retrace the effects of adjustment on the constraints and behaviours of the actors and on the opportunities available to them. We come close to achieving this in accounting for the effects of adjustment on a country's economy. However, the problem expands when it comes to identifying and understanding the multiple causal pathways and leverage mechanisms, both direct and indirect, that are mobilized in health production and assessing the consequences of adjustment outside of the strictly economic sector. The complexity of this undertaking is illustrated by Sahn and colleagues, on the link between adjustment and poverty: "To understand how a particular adjustment policy affects the poor, we need to trace a complex path accounting for direct and indirect effects. In doing so, we must recognize the mediation of policy change by various economic institutions and account for other exogenous influences before we can understand the distributional and poverty effects" (Sahn et al. 1997, p.7). In other words, to assess the effects on health, we first need to identify the direct and indirect mechanisms that link macroeconomic developments with the production and consumption of health inputs.
However, to uncover these transmission mechanisms is a complex undertaking, particularly because we must take into account the many causal influences, as well as the interconnectedness of the causal pathways that may have been set in motion. As Chen puts it, "cause and effect linkages between the dynamics of the macro-economy and the health of populations are multiple, difficult to pinpoint and undoubtedly complex" (Chen 1986, p. 279). Certainly, public policies, through a variety of mechanisms, will model the opportunities for being in good or poor health. But these opportunities arise from a tangle of complex influences in which are interwoven public policies, lifestyles, characteristics of individuals and families, households' allocation decisions, etc. Then we must be able to distinguish between the changes that have come out of the reforms and the causes that provoked the reforms in the first place, i.e., the crises and imbalances that the adjustment programs were intended to correct (Sahn et al. 1997). Even there, the undertaking is complex, because what we are observing is the cumulative effect of the crises and the institutional responses to them (Genberg 1995). The interconnected nature of crises and reforms lends a certain artificiality to their separation, and, much as we might wish to, it is not always possible to distinguish their respective contributions to any eventual deterioration in health status.
The challenge, however, is not only in the multiplicity of causal mechanisms involved in health production; it is also in their interconnectedness and overlapping. We would need not only to model the causal links, but also to anticipate the consequences of the coexistence and interaction of the mechanisms of health production and the additive or synergistic, agonistic or antagonistic, nature of their behaviours. Ideally, we would also need to be able to judge the effects of (i) the intensity of the changes implemented, looking for possible dose–effect relationships; (ii) the sequence of implementation of macroeconomic measures; (iii) any interference from exogenous variables (e.g. natural disasters, social troubles, changes in governments, etc.); and (iv) the time delay between the adoption of a measure and the expression of its consequences. We also need to consider the recursive nature of certain causal mechanisms, as when changes at the micro level (poverty, productivity, input factors availability, etc.) feed back to the macro level.
All of this makes it especially complicated to estimate what the situation would have been in the absence of reforms (Behrman 1995), and identifying effects that are specific to the adjustment presents formidable methodological problems that some consider absolutely insoluble: "Changes in economic policy and Health Sector Reform are only a part of this very large set of forces influencing health outcomes, isolating the effect of which is virtually intractable" (Narayana and Navaneetham 1998, p.15).
The analysis of public reforms in relation to health status is thus strongly limited in terms of both internal validity and generalizability. It may seem more productive to concentrate on the study of modifications that, although not addressing changes in well-being, provide more prosaic information on institutional responses. This means judging the effects of reforms less on the basis of outcomes and more on outputs, i.e., changes induced in the production of services (availability, price, quality) and their interface with populations (accessibility, utilization). These indicators are, in fact, more specific, more sensitive, and more rapidly reactive to institutional changes and structural reforms, especially those targeting supply, than are health and well-being. Even this approach, however, has its limitations. Of course, it helps lift some of the constraints that weigh upon the analysis of causal links between adjustment and the health sector.
However, to a certain extent, what applies to the production of health (outcomes) applies as well to the production of health systems (outputs). Because these influences can also be numerous, any study of the effects of adjustment on system outputs and of the resulting snowball effect must disaggregate the mechanisms having direct or indirect influence on supply and demand in health services (Figure 3). The supply response to production incentives (e.g. devaluation, easing of price restrictions) is conditioned by many contextual factors such as the degree of liberalization of internal markets; the tax system; the quantity and quality of existing infrastructures; access to credit; the efficiency of financial institutions; the mobility of the labour force; the concentration and distribution of property; or again, the strength of the private sector and the extent of the informal sector. Similarly, the effects on demand are also dependent upon context, economic situation, political and institutional factors, and other factors.
FIGURE 3. LINKING MACROECONOMIC ADJUSTMENT POLICIES, HEALTH SECTOR REFORMS, AND HEALTH SYSTEMS OUTCOMES.
In practice, no evaluation design is perfectly satisfactory. However, by triangulating different designs and methodologies and using a conceptual base that allows us to anticipate and understand macro–micro links, we can strengthen the internal validity of the evaluation. Our aim is to go well beyond the superficial approach of comparing aggregate indicators between adjusting countries and non-adjusters and to ensure a more serious analysis that recognizes the diversity of sociopolitical contexts, the content and pace of implementation of public policies, and the intercurrence of a variety of exogenous factors.
To be able to document policy outcomes and formulate recommendations, country-focused studies are primordial. However, any study that relies only on before–after analyses will have insufficient internal validity (Sahn et al. 1997). By using data and analytical tools that allow us to model, through simulation, multivariate analyses or, with the help of time series, the likely counterfactual scenario, we can greatly improve the quality of our results. Also, doing multiple-case studies, as proposed by Yin (1984), makes it possible to bring more depth to single case studies, identify potential recurrences, and avoid inferences based on anecdotal results.
With regard to evaluation criteria, it is important to measure the absolute consequences of the reforms and their effects in terms of how they respond to needs. Experience shows, however, that these consequences are differently distributed among economic sectors, social groups, or territories. Some persons will see their situation deteriorate, whereas others may not be affected by the reforms or may even benefit from them. Therefore, it is also important to be able to assess the distributive effects of the policies. Finally, as we have seen, the available quantitative indicators are often limited and poorly reliable and must be complemented by quantitative and qualitative primary data (Lundy 1996).
The MAPHealth project is an international study that aims to contribute to a better understanding of the links among macroeconomic policies, health sector reform, and the organization and delivery of health services in developing countries. The project has a set of interwoven objectives. Its primary objective is to obtain as clear a picture as possible of the health systems of Southern countries whose economic environments, after a succession of crises and structural reforms, are considered to be turbulent. This picture includes an examination of health reforms undertaken in these countries—reforms that preceded the upheavals in some cases and followed in others. We were also interested in articulating these diverse elements and in studying how health systems responded, at different times and in different contexts, to major transformations of their environments. For the reasons mentioned earlier, the project is focused on the health-care system, and particularly on the aspects of quality, accessibility, and utilization of health services. The study of changes attributable to the implementation of overall or sectoral policies is therefore centred on the organization and delivery of services, both within each country and between participating countries.
The study originally included eight countries: four African (Burkina Faso, Kenya, Uganda, Zimbabwe); two Latin-American (Colombia, Mexico); and two Asian (India, Thailand). Clearly, this selection is not fully representative of all situations of turbulent environments or of crisis–adjustment cycles. This would be extremely difficult to achieve, given the vast heterogeneity of situations and contexts among countries. The group is nevertheless sufficiently demonstrative of the great diversity of public policy reforms, particularly in the health sector, carried out in recent years. These countries embarked on this project in response to the invitation of the International Development Research Centre (IDRC). This development agency, which initiated and funded the study, invited research teams in Africa, Asia, and Latin America with an interest in the proposed area of study to submit research proposals. These proposals were subsequently evaluated by a peer review committee and eight teams were selected based on numerous criteria, including multi-disciplinary research capacity, political stability, availability and reliability of country-level data, level of country development, and background and experiences with MAPs and health sector reform. Six of these teams, with the support of a team of researchers from the Université de Montréal, in Québec, Canada, undertook and completed the process that has resulted in this report. Although case studies were carried out in Uganda and Kenya, the teams in these countries did not complete the process that would have led to inclusion in this book. Therefore, the material presented in the remainder of this chapter and in subsequent chapters refers only to six countries.
The participating countries, with some key development indicators, are listed in Table 2. Three countries fall within the group of intermediate economies (Colombia, Mexico, Thailand), and one, Burkina Faso, is among the least advanced countries. These countries have different development profiles and experiences with the implementation of MAPs and health sector reform. All of them were led to implement structural reforms of their economies, preceded or followed by more or less drastic reforms of their health-care systems. These reforms, whose intensity, content, and conditions of implementation are presented in the following chapters, are summarized in Table 3.
TABLE 2. SELECTED DEVELOPMENT INDICATORS FOR THE MAPHEALTH STUDY COUNTRIES, 2000
Country |
Type of government |
Population (000) |
Human development index |
Human poverty index |
Illiteracy rate-adults |
Infant mortality rate |
Corruption perception index rank* |
% of poor |
Burkina Faso |
Parliamentary Republic |
11 553 |
0.325 |
58.4 |
76 |
105 |
65 |
61 |
Colombia |
Republic |
43 035 |
0.772 |
9 |
8 |
20 |
60 |
18 |
India |
Federal republic |
1 032 355 |
0.577 |
34 |
43 |
68 |
69 |
35 |
Mexico |
Federal republic |
99 420 |
0.796 |
10 |
9 |
25 |
59 |
10 |
Thailand |
Constitutional monarchy |
61 184 |
0.762 |
14 |
5 |
25 |
60 |
13 |
Zimbabwe |
Parliamentary democracy |
12 821 |
0.551 |
41 |
11 |
73 |
65 |
26 |
* out of 90 countries.
Source: All data except corruption perception index: UNDP, Human Development Report (2000). Corruption perception index: Transparency International (2000).
The project design was developed by the team leaders within a framework of close collaboration and with the support of the IDRC. Based on Yin's terminology, this is a multiple-case, multiple-level study. It is multiple-case, in that each country is considered a case in which an in-depth analysis of changes is conducted. It is multiple-level because it has, in each of the cases, three levels of observation and analysis (macro, meso, and micro) conducted at the national, regional, and local levels.
Together, the team leaders developed a methodology having one part in common and the other part specific to the case, taking into account the particular objectives and contexts encountered.
TABLE 3. IMPLEMENTATION OF MACROECONOMIC ADJUSTMENT POLICIES AND HEALTH SECTOR REFORMS IN THE MAPHEALTH STUDY COUNTRIES
The general principles of the country studies are that they are cross-sectional and retrospective (over the past 15 years) and that they use both quantitative and qualitative data based on multiple primary and secondary sources, including national and regional surveys (Table 4). The finest unit of observation is the household, and the units of analysis are the country, regions, households, and individuals.
TABLE 4. GENERIC STUDY DESIGN PROPOSED TO, AND SUBSEQUENTLY ADAPTED BY, PARTICIPATING MAPHEALTH STUDY COUNTRIES
Level |
Study, survey |
Data collection sources |
Instruments | |
|
|
Primary |
Secondary |
|
National |
Health system study (HSS) MAPHealth study |
Interviews, focus groups Interviews, focus groups |
Policy plans, reports, previous surveys, system data, statistics Policy plans, reports, previous surveys, system data, statistics |
Health systems kit MAPHealth kit |
Regional |
Community survey (CS) Health facilities survey (HFS) Household surveys (HHS) |
Interviews Interviews, observations Interviews |
Reports, previous studies, system data, statistics System data, statistics, reports Databases |
Community checklist Health facility questionnaires Household questionnaires |
The national-level studies are concerned with the health system and the implementation of MAPs. Each country team selected two or three states or regions based on analyses of the country situation and regional differences in the implementation of reforms. Within these regions, a limited number of administrative units, communities, and health districts were surveyed. Finally, community, household, and health facility surveys were conducted locally. The methodology was piloted in India and Mexico and refined based on the findings and experience of those teams.
A common framework was proposed for the different teams to facilitate the process of analysis and future comparative approaches. It was, in fact, useful most of all in developing a common vision of the mechanisms by which macroeconomic policies can influence health systems and, later, in the formulation of specific research hypotheses by the country teams. The mechanisms through which adjustment can drive changes in health systems are both complex and deeply tied to the characteristics of these MAPs (measures of implementation, duration, intensity, and conditions of their introduction). Following Peabody's (1996) suggestions, these mechanisms are seen as both direct and indirect:
— Expenditures on health constitute the main lever through which adjustment can directly influence the health system. Adjustment and the eventual sector reforms that accompany it can directly affect the total spending, the distribution, and the allocation of expenditures. The relative prices of imported health goods (equipment, medications, consumables) constitute a second important lever. Equally, the positioning, adjusting, and directing of the general organization of the health system can be profoundly remodeled by institutional reforms and measures looking to promote the private sector. These different measures will also influence the system of care and play on diverse health services, in terms of their availability, quality, price, financial autonomy, geographical and financial accessibility, etc.
— Similarly, the demand for health care can be modified by different mechanisms. Austerity measures directly influence numerous determinants of health and of the ability to pay: employment, disposable income, education, consumption, nutrition, etc. The engendered price-level modifications of the quality and availability care in turn affect individual and household perceptions of quality and price, their willingness to pay for health services, and the capacity to reach these services, all of which will subsequently modify the behaviour of people seeking care. The most vulnerable social groups—the poor, women, and children—are also particularly exposed to the risk of exclusion.
As part of the framework, some key intermediate variables were identified, with implementation of macro policies being the main variable of interest. We distinguished:
— three levels of effects: (i) of macroeconomic adjustment programs, including health sector reform, on the health system and communities; (ii) of previous changes on the supply of health services and on the demand for health care among susceptible households; and (iii) of all these modifications on the geographical and financial accessibility of the services, and consequently, on the utilization of health services (Figure 1).
— three types of effects: (i) on the supply side, i.e., on health policies (regulations, organization, financing) and services (availability, price, quality); (ii) on the demand side, i.e., on determinants of health (employment, income, nutrition, education) and on the demand for health care (needs, preferences, budget constraints); and (iii) on accessibility and utilization of health services, resulting from the interaction of the first two groups of factors (Figure 3).
Macro variables include both contextual variables (demographic, epidemiological, social structure, political dynamic, economic data) and variables related to MAPs (importance, implementation, impact, duration, etc.). Meso variables concern both the health-care system (organization, regulation, financing, and delivery of care and health services) and communities (social and economic organization, political and community environments, local resources, needs and health behaviours, inequities, and gender-specific issues). Micro variables include information on access, quality, and utilization of care, and on household capabilities and well-being.
To facilitate the analyses, the group developed a set of instruments that were then adapted by each country team (Table 4) and made available in three languages (French, English, and Spanish). The three levels of measurement are reflected in three general tools: macro- and meso-level instruments with a guide, health facilities questionnaire with supplements and guide, and household questionnaires.
The macro- and meso-level kits were organized into four components: economic indicators, analysis of structural adjustment, health system indicators, and a community checklist. The economic indicators component included grids for data on population, human development, macroeconomics, public expenditures, households, and employment. The section on the analysis of structural adjustment provided researchers with guidance on evaluating macroeconomic measures, suggestions on the presentation of structural adjustment policies, and a matrix for analyzing the impact of MAPs. The evaluation of the performance of the health-care system was centred on four qualities or attributes: effectiveness, efficiency, equity, and sustainability. The section on health-system indicators included grids for data on baseline indicators, health resources, service coverage, health expenditure, prices of care and services, and utilization. Teams were encouraged to consult national sources as well as statistical yearbooks and publications from international agencies and to augment this quantitative data with qualitative information from key informants. Finally, the community checklist provided a list of factors guiding researchers to describe the general background of the location, sociocultural characteristics of the community, and details of the local health-care system.
The tool for the evaluation of health centres consists of three questionnaires plus a guide. The principal questionnaire elicits data about the centre and the accessibility, quality, and utilization of services, as well as information on the impact of macro-economic policies and health reforms on the centre. The first appendix is a questionnaire to be administered to centres with hospital beds, with more in-depth questions on services offered, physical plant, equipment, drugs, financial data, cultural data, types of activities, reasons for hospitalization, quality indicators, and the impact of health reforms. The second appendix is a questionnaire to be completed by health-care personnel. It asks questions on the personnel profile, their degree of satisfaction with work conditions, and their opinions regarding the evolution of certain parameters with respect to the health centre and its environment.
The largest tool in the methodology is the set of four household questionnaires. The first collects information on household composition, income, assets, expenses, education, and utilization of health-care services. The second asks about community participation, perceived health status, accessibility of health-care services, willingness to pay, perceptions of the quality of health care, and health problems experienced during the previous four weeks. Questions on women's and children's health and on women's workload are found in the third, and the last one asks specific questions about health care utilization.
The case studies evolved rapidly into research centred on country-specific issues. As such, the analyses were carried out separately by each of the country teams and the scope of comparative analyses initially envisioned was necessarily substantially reduced, insofar as the pace at which the studies were carried out tended to vary considerably from one team to another.
As a general pattern, the analyses paid special attention to the effects of reforms on specific subgroups of the population, on both the supply (e.g. health practitioners) and the demand (e.g. women, poor) sides. Because of local concerns, almost all the teams focused the analyses on the implementation and the consequences of health sector reforms. Special attention was paid to structural adjustment or external shocks by the teams in Burkina Faso, Mexico, Thailand, and Zimbabwe. The teams in federal countries (India and Mexico) gave more attention to the possible mediating effects of state- or province-specific policies and regulations. In general, more attention was devoted to the analysis of household questionnaires (demand side) than to the health facility questionnaires, which could undoubtedly be further explored.
1. For a more complete inventory, the reader may consult Bourguignon and Morrisson (1992); Behrman (1995); Costello et al. (1994); Genberg (1995); Lundy (1996); Peabody (1996); Sahn et al. (1997); and White (1996).
2. Phrase borrowed from Mohr (1995).
3. The category of countries "exposed to reform" could be divided into several subcategories to distinguish between different levels of exposure, or periods during which adjustment occurred (see Chapter 2).
Behrman, J.R. 1995. Santé et croissance économique: théorie, réalité et politique. In Environnement macro-économique et santé. Avec études de cas dans les pays les plus démunis. Organisation mondiale de la Santé (World Health Organization), Geneva, Switzerland. pp. 17–50.
Bijlmakers, L.A.; Bassett, M.T.; Sanders, D.M. 1996. Health and structural adjustment in rural and urban Zimbabwe. Research report no. 101. Nordiska Afrikainstitutet, Uppsala, Sweden. 78 pp.
Bijlmakers, L.A.; Bassett, M.T.; Sanders, D.M. 1998. Socioeconomic stress, health and child nutritional status in Zimbabwe at a time of economic structural adjustment. Research report no. 105. Nordiska Afrikainstitutet, Uppsala, Sweden. 123 pp.
Bourguignon, F.; Morrisson, C. 1992. Adjustment and equity in developing countries: a new approach. Development Centre of the Organisation for Economic Co-operation and Development. OECD Publications and Information Centre, Washington, DC, USA. 111 pp.
Breman, A.; Shelton, C. 2001. Structural adjustment and health: a literature review of the debate, its role-players and presented empirical evidence. Commission on Macroeconomics and Health (CMH), Working Paper Series, no. WG6:6, World Health Organization, Cambridge, MA, USA. 53 pp.
Campbell, D.T.; Stanley, J.C. 1963. Experimental and quasi-experimental designs for research. Rand McNally, Chicago, IL, USA. 84 pp.
Chen, L.C. 1986. Health policy responses: an approach derived from the China and India experiences. In Bell, D.E.; Reich, M.R., ed., Health, nutrition, and economic crisis: approaches to policy in the Third World. Auburn House, Dover, MA, USA. pp. 279–305.
Cook, T.D.; Campbell; D.T. 1979. Quasi-experimentation: design and analysis issues for field settings. Rand McNally, Chicago, IL, USA. 405 pp.
Cornu, A.; Massamba, J.P.; Traissac, P.; Simondon, F.; Villeneuve, P.; Delpeuch, F. 1995. Nutritional change and economic crisis in an urban Congolese community. International Journal of Epidemiology, 24(1), 155–164.
Costello, A.; Watson, F.; Woodward, D. 1994. Human face or human facade? Adjustment and the health of mothers and children. Centre for International Child Health, Institute of Child Health, London, UK. 68 pp.
Garenne, M.G.; Gakusi, E. 2000. Health effects of structural adjustment programs in sub-Saharan Africa. Working paper, French Center for Population and Development Studies (CEPED), Paris, France. 26 pp.
Genberg, H. 1995. Ajustement macro-économique et secteur de la santé. In Environnement macro-économique et santé. Avec études de cas dans les pays les plus démunis. Organisation mondiale de la Santé (World Health Organization), Geneva, Switzerland.
Grosh, M.E. 1990. Social spending in Latin America: the story of the 1980s. World Bank discussion paper no. 106. World Bank, Washington, DC, USA. 163 pp.
Hanson, K. 1999. Implementing health sector reform in Africa: a review of eight country experiences. UNICEF, Division of Evaluation, Policy and Planning, New York, NY, USA. 92 pp.
Jayarajah, C.; Branson, W.; Sen, B. 1996. Social dimensions of adjustment: World Bank experience, 1980–93. Operations Evaluation Division, World Bank, Washington, DC, USA. 218 pp.
Kakwani, N. 1995. Structural adjustment and performance in living standards in developing countries. Development and Change, 26(3), 469–502.
Lachaud, J.P. 2001. Dynamique de pauvreté et inégalité de la mortalité des enfants au Burkina Faso. Documents de travail, no. 66, Centre d'Économie du Développement de l'Université Montesquieu Bordeaux IV. Pessac, France. 20 pp.
Lundy, P. 1996. Limitations of quantitative research in the study of structural adjustment. Social Science and Medicine, 42(3), 313–324.
Makinen, M.; Waters, H.; Rauch, M.; Almagambetova, N.; Bitran, R.; Gilson, L.; McIntyre, D.; Pannarunothai, S.; Prieto, A.L.; Ubilla, G.; Ram, S. 2000. Inequalities in health care use and expenditures: empirical data from eight developing countries and countries in transition. Bulletin of the World Health Organization, 78(1), 55–65.
Mohr, L.B. 1995. The theory of impact analysis: experiments and the elementary quasi-experiments. In Impact analysis for program evaluation. Sage Publications, Thousand Oaks, CA, USA. 311 pp.
Morrissey, O. 1997. Micro-macro linkages: an economist's perspective. Journal of International Development. 9(5), 755–760.
Mwabu, G. 1996. Health effects of market-based reforms in developing countries, UNU/WIDER Research Paper no. 120. World Institute for Development Economics Research, Helsinki, Finland. 53 pp.
Narayana D.; Navaneetham, K. 1998. Impact of structural adjustment programme and health sector reforms on reproductive health and family planning services in South Asia. Study sponsored by the United Nations Fund for Population Activities. Centre for Development Studies, Thiruvananthapuram, Kerala, India.
Newbrander, W.; Collins, D.; Gilson, L. 2000. Ensuring equal access to health services: user fee systems and the poor. Management Sciences for Health, Boston, MA, USA. 187 pp.
Noorbakhsh, F.; Paloni, A. 1998. Stuctural adjustment programmes and export supply response. Journal of International Development, 10, 555–573.
Peabody, J.W. 1996. Economic reform and health sector policy: lessons from structural adjustment programs. Social Science and Medicine, 43(5), 823–835.
Reed, D. 1999. Ajustement structurel, environnement et développement durable. Editions L'Harmattan, Paris, France. 377 pp.
Rossi, P.H.; Freeman, H.E. 1993. Evaluation: a systematic approach, 5th ed. Sage Publications, Newbury Park, CA, USA. pp. 215–260.
Sahn, D.E.; Dorosh, P.; Younger, S. 1997. Structural adjustment reconsidered: economic policy and poverty in Africa. Cambridge University Press, New York, NY, USA. 304 pp.
Shadish, W.R.; Cook T.D.; Campbell, D.T. 2001. Experimental and quasi-experimental designs for generalized causal inference. Houghton Mifflin Company, Boston, MA, USA. 623 pp.
Transparency International. 2000. TI corruption perceptions index 2000. http://www.transparency.org/policy_research/surveys_ indices/cpi/previous_cpi__1/2000, cited April 2006.
UNDP. 2000. Human development report 2000. United Nations Development Programme, New York, NY, USA, http://hdr.undp. org/reports/global/2000/en/, cited April 2006.
Van der Gaag, J.; Barham, T. 1998. Health and health expenditures in adjusting and non-adjusting countries. Social Science and Medicine, 46(8), 995–1009.
White, H. 1996. Adjustment in Africa. Development and Change, 27, 785–815.
WHO (World Health Organization). 2004. Maternal mortality in 2000: estimates developed by WHO, UNICEF and UNFPA. World Health Organization, Geneva, Switzerland. 39 pp.
World Bank. 2005. Health, Nutrition and Population Statistics (HNPStats), http://devdata.worldbank.org/hnpstats/
Yin, R.K. 1984. Case study research: design and methods. Sage Publications, Newbury Park, CA, USA. 160 pp.
Burkina Faso is country landlocked within West Africa. The majority of its 12 million inhabitants (83%) live in rural areas, and two-thirds of the urban population is concentrated in the two cities of Ouagadougou and Bobo-Dioulasso. The country is comprised of more than 60 ethnic groups, of which the Mossi are the most numerous. Muslims account for the majority of the population (52%); the rest are either animists (26%) or Christians (21%).
Burkina Faso is counted among the group of least advanced countries. In 2001, it ranked 159th out of 162 on the Human Development Index of the United Nations Development Programme (UNDP). Its per capita GDP in 2001 was 240 USD. Three out of four adults are illiterate. One out of every 10 households is headed by a woman. Nearly half the population live below the poverty line (45%) and 65% live on less than a dollar (USD) a day. Poverty is essentially rural (94% of the poor in 1998) and affects more than half the peasants. Rainfall is minimal and unevenly distributed, and the soil is poor. Nevertheless, the economy is mostly based on the primary sector (agriculture, livestock), which employs about 90% of the active population, according to the 1996 census. This sector, which accounts for 40% of GDP, is largely made up of small family enterprises using low-technology methods of cultivation and husbandry, except in the case of cotton farming. The informal sector is the country's second largest source of employment, in urban centres as well as in the countryside, in both the secondary sector and the service sector. Production units are generally small, diverse, and often headed by women.
Health needs are enormous and largely unsatisfied. Life expectancy is 46 years. Maternal mortality is exceedingly high, with estimates varying between 484 and 1 000 per 100 000 live births. Infant mortality is 104 per 1 000, and nearly one out of five children dies before the age of five. One out of every three children under the age of five is malnourished. Infectious diseases continue to be fearsome killers: diarrhea, malaria, measles, and acute respiratory infections being the principal causes of death among children (INSD 1998). The incidence of diseases for which there are vaccinations has diminished, but there continue to be cyclical epidemics of measles and meningitis. In 1986, 10 cases of AIDS were reported; by 2000, there were 1 951, a 20-fold increase. Problems related to water and sanitation continue unabated. Only 29% of the population have access to potable water and 30% of urban households have access to toilets.
Health policies have emulated major changes observed on the African continent. The 1980s were mainly characterized by implementation of primary care services, broader vaccination programs, and family planning policies, which had until then been prohibited under law dating back to colonial times. It was the era of major health campaigns including "commando vaccination" and "one village = one primary care post." Beginning in 1992, the country entered a period of reform inspired by the Bamako Initiative and by measures advocated by international development agencies.
The country has 53 health districts grouped within 11 health regions. Public health services are organized in a classic pyramidal structure. The base is made up of peripheral health structures—health and social development centres (centre de santé et de promotion sociale – CSPS), dispensaries, and maternity centres. The medical–surgical centres (centre médical avancé – CMA) are the first line of referral, followed by eight regional hospitals. At the summit are the two national hospitals of Bobo-Dioulasso and Ouagadougou. In addition to the public sector, there is the army's health services system, as well as a private health sector flourishing in urban centres, offering a patchwork of more or less qualified practitioners. As in other countries of this region, the private sector is largely made up of the private practice of professionals on the state's payroll, a practice that is technically illegal but tolerated.
Burkina Faso underwent two periods of adjustment. The first, implemented between 1983 and 1987, was carried out without the support of international financial institutions. This was an austerity program designed to bring public finances back into equilibrium after the upheavals caused by the advent of Thomas Sankara's revolutionary regime. Then, in 1991, to make the transition from a state-dominated to a market-based economy, the government decided to implement a new structural adjustment program (SAP), this time with the aid of the International Monetary Fund (IMF) and the World Bank. This assisted-adjustment program involved stabilization measures and a set of structural and sectoral reforms. The structural reforms consisted of a series of measures aimed at reducing public spending; reforming domestic and foreign taxation structures; easing restrictions on the labour market, businesses, and prices; privatizing and rehabilitating public enterprises; and reforming the banking sector. However, the most significant measure was adjustment of the exchange rate, under which the CFA franc, indexed to the French franc, was sharply devalued to one-half its value, effective 11 January 1994.1 These structural reforms were accompanied by sectoral reforms in agriculture, environment, transportation, human resources, and health. Health sector reform focused on decentralization and cost recovery and on the establishment of health districts; reforming the hospital sector; liberalizing health services with regard to the private sector; and ensuring the supply of essential medicines.
The methodology used in Burkina Faso conformed closely to that proposed for the MAPHealth project overall. The study covered the years 1983–2000 to include both periods of macroeconomic adjustment. The macro aspect of the study is based upon documents from national and international sources, a database of secondary socioeconomic data reconstructed by the study team, and interviews of 18 key informants working in upper management, health institutions, or development organizations. These inputs were complemented by a survey conducted in three regions selected as being representative of the diversity of environments and living conditions and therefore of the effects of the reforms in different geographical areas. The first was the city of Bobo-Dioulasso. The second-largest city in the country, with a population of about 400 000, Bobo-Dioulasso is relatively prosperous and serves as a major crossroads between Mali, the Ivory Coast, Ghana, and Burkina Faso. The second study site was the Bazega region, a rural truck-farming area about 50 kilometres from Ouagadougou, the capital. The third region was Nouna, in the west, a rural area devoted primarily to growing cotton. In Nouna and Bazega, the study covered both rural villages and towns that are considered, in Burkina Faso, to be semi-urban zones.
Altogether, the study covered 53 communities composed of either villages or districts (Table 1). The household survey included 1 604 families randomly selected from within the participating communities. In these households, 857 persons had presented with an illness during the four weeks preceding the interview. The health facilities study looked at 13 of the 32 existing primary care institutions (CSPS and maternity centres). There were 121 respondents to the staff questionnaire, and in each region, approximately 50 patients of each institution participated in exit interviews.
TABLE 1. DESIGN AND SAMPLES OF REGIONAL STUDIES, BURKINA FASO
|
Study sites |
TOTAL | ||
|
Bobo-Dioulasso |
Nouna |
Bazega |
|
Communities included |
18 |
21 |
14 |
53 |
Cities (urban) |
1 |
- |
- |
1 |
Small cities (semi-urban) |
- |
1 |
1 |
2 |
Villages (rural) |
- |
17 |
10 |
27 |
Samples |
|
|
|
|
Household surveys |
|
|
|
|
Households |
694 |
510 |
400 |
1 604 |
Individuals |
5 353 |
4 182 |
3 251 |
12 786 |
Persons with illness |
300 |
118 |
439 |
857 |
Institutional surveys |
|
|
|
|
Institutions |
7 |
3 |
3 |
13 |
Representatives |
38 |
38 |
45 |
121 |
Users |
43 |
43 |
59 |
145 |
Source: MAPHealth study (2001).
Editorial constraints require that we limit our examination of results in this chapter to the period covering the assisted adjustment and the health system reform. First we will look briefly at the evolution of the economic environment during the assisted adjustment period and of the implementation of decentralization and other elements of the health system reform. Then we will examine changes in resource allocation, access to treatment and primary care services, and utilization. The last section is devoted to an analysis of the economic burden of illness and the effects of illness on poverty.
The macroeconomic adjustment program was implemented gradually, within an initially difficult political context, because at the time of the discussions with the Bretton Woods organizations, the country was under an exceptional regime. The currency devaluation provoked a strong public reaction and forced the government to raise the guaranteed minimum wage and the salaries of civil servants by 5–10%. Implementation of certain measures, such as privatizing and liquidating public enterprises, was difficult. Overall, however, the adjustment program was carried out without major obstacles and Burkina Faso was considered by the IMF to have been a good adjuster.
Between 1992 and 1994, the economy was slowing down and by the time of the currency devaluation in 1994, the country was in recession. In 1995, the trend began to reverse itself, and the average rate of GDP growth from 1995 to 2000 was 4.4%. This growth can be attributed to significant expansion of the agriculture sector, increased consumption, and recovery in the mining, construction, and retail industries. After devaluation, foreign commerce benefited from market increases in cotton and in gold, and from the expansion of agricultural exports.
Poverty surveys show a small increase in the level of poverty between 1994 and 1998, going from 44.5% to 45.3% (Table 2). Poverty remained concentrated in rural regions, with little change, while there was greater impoverishment in the cities. There, 1 out of 10 households was poor in 1994; four years later, the figure was 1 out of 6. In 1994, less than 4% of poverty was concentrated in the cities; by 1998, it was 6.1%, a 60% increase. Overall, income in rural areas increased during the period in question, especially among those in the agricultural sector. However, persons on fixed incomes and particularly civil servants, artisans, and day workers—mostly concentrated in cities—saw their real incomes stagnate or decrease.
TABLE 2. EVOLUTION OF POVERTY IN BURKINA FASO, 1994–1998
|
Incidence of poverty (%) |
Contribution to total poverty (%) | ||||
|
1994 |
1998 |
Change (%) |
1994 |
1998 |
Change (%) |
Socioeconomic groups |
|
|
|
|
|
|
Public sector employees |
2.2 |
5.9 |
168 |
0.2 |
0.5 |
150 |
Private sector employees |
6.7 |
11.1 |
66 |
0.4 |
0.7 |
75 |
Craftsmen, merchants |
9.8 |
12.7 |
30 |
1.4 |
1.6 |
14 |
Other workers |
19.5 |
29.3 |
50 |
0.3 |
0.4 |
33 |
Production farmers |
50.1 |
42.4 |
-15 |
11.8 |
15.7 |
33 |
Subsistence farmers |
51.5 |
53.4 |
4 |
78.9 |
77.1 |
-2 |
Unemployed |
41.5 |
38.7 |
-7 |
7.1 |
4.0 |
-44 |
Environment |
|
|
|
|
|
|
Urban |
10.4 |
16.5 |
59 |
3.8 |
6.1 |
61 |
Rural |
51.0 |
51.0 |
0 |
96.2 |
93.9 |
-2 |
National |
44.5 |
45.3 |
2 |
100 |
100 |
0 |
Source: Ministère de l'économie et des finances, Burkina Faso (2000).
Over the course of the decade, prices remained relatively stable. Average inflation was 5.8%. Inflation went into two-digit levels only in 1994, when it peaked at 24.7%. After devaluation, prices of basic goods were generally contained. On the other hand, the prices of grain, rice, equipment, and manufactured imports were automatically affected by the parity changes of 1994.
The government considers that many of the SAP objectives were achieved. Public revenues have increased, from 11–12% of GDP at the beginning of the decade, to 13–14% at the end. However, controlling expenditures has been more of a challenge, especially when the government was obliged, in 1993, to back down from certain unpopular measures and satisfy public demands. The balance of payments remained in deficit, despite the positive influx of transfers in the form of grants and foreign subsidies, both public and private. Public investment was significant. Current expenditures, which were near 80% in the 1980s, decreased to less than 50% in the last three years of the 1990s, with public funds going primarily to benefit the water, agriculture, and public works sectors (Figure 1). If grants are taken into account, the deficit was on average no more than 4.8%, a level compatible with the macroeconomic objectives. Foreign debt grew considerably during the first half of the decade, going from 298 billion CFAF in 1990 to 770 billion CFAF in 1995. This was primarily due to increases in loans from multilateral organizations, mostly for structural adjustment credits. Since 1996, the country has not gone into arrears on its foreign debt, and debt servicing has decreased over the past five years, representing 2.26% of GDP and 17% of current revenues in 1999. In the end, the foreign trade deficit remains high, even though agricultural exports have very much benefited from the currency devaluation and price increases.
FIGURE 1. EVOLUTION OF PUBLIC FINANCES IN BURKINA FASO, 1991–1999, (CFAF BILLIONS).
Source: Ministère de la santé, Burkina Faso. Direction des études et de la planification.
To ensure decentralization, a number of laws and regulations were passed governing the functioning of the health districts and defining management structures and procedures. These gave public institutions a great deal of management autonomy and established the principle of cost recovery for operating expenses through setting fees for services, and particularly for essential drugs.
Decentralization began with the designation of 53 health districts in 1992. This was followed by the training of staff, construction and equipping of facilities, and the allocation—mainly through external partners—of financial resources to the districts. In 2000, six years after the beginning of decentralization, only two or three districts could be said to be operational, under the criteria set out in the reform.2 Country-wide, fewer than five districts had the requisite number of physicians and only 17 of the 53 district hospitals had the capacity to do cesarean sections (Bodart et al. 2001). In the same year, in the Dédougou region, only 34% of health facilities had a dispensary, a maternity centre, and a depot for essential generic drugs (EGD), and barely half met the minimal staffing norms (GTZ 2000).
In many respects, decentralization was not accompanied by real transfer of power to the districts. A good example of this is in the critical area of human resources management, where recruitment of personnel is almost exclusively under the authority of the central management level, which keeps firm control over the movement and allocation of staff. Even though action plans are prepared, in most cases, by the districts, planning is more often based on available resources than on the needs of the population. Because of inadequate management capability, action plans are poorly justified, insufficiently resourced, and rarely executed as planned (the exceptions being districts with outside support). Districts complain particularly about urgent activities being parachuted onto them by the Ministry without warning, which take over their staff and budgets and interfere with planned programs.
Another source of problems is the devolution of financial resources and their availability to the districts. Officially, the decentralization was to have been accompanied by budget transfers to the districts. In fact, since 1996, the national budget has contained a line item showing the assignment of such funds. However, only a part of that amount actually devolves to the districts, and in any case, the amounts transferred constitute only a small fraction of public expenditure on health. Operating budgets continue to be concentrated and mostly managed at the level of the Ministry. In 1998, the budget allocated to the districts represented 14% of the Ministry's operating costs, excluding salaries, which was hardly more than the budget allocated to the two national hospitals. The central government continues to be responsible for a portion of expenses that should be assumed by the districts. According to a World Bank analysis of the 1998 law on finances, 70% of equipment and supplies for the districts and 52% of their staff expenditures were being handled by the central authorities. As of 2000, the districts and the regional boards of health still only managed about half of their material expenses (Government of Burkina Faso 2001).
Because of this, districts must rely heavily on cost-recovery revenues. However, this situation can have negative effects, as district management teams are motivated to hoard these revenues, and in fact to increase them, to ensure coverage of operating expenses. A GTZ study showed that between 1995 and 1999, in two regions, the contribution of sales revenues of EGD depots to district management operating budgets multiplied 50-fold (GTZ 2000). The current situation also reinforces inequities between districts depending on whether they have financial support from external partners. Since 1998, at least 20 districts receive financing either directly from the State or from its partners and are thereby advantaged in relation to other districts. There are also inequalities within these advantaged districts; some financing may be largely symbolic, whereas in other cases the financing in fact exceeds the capacity of the district to absorb it. This was the case in Kongoussi, where between 1995 and 1997 only 65% of funds received were able to be used (Aubel and Sobgo 1998).
The private sector, which until then had been very constrained, was able to develop considerably during the 1990s. Legislation was revised and measures were introduced to promote and support the creation of laboratories, clinics, and private pharmacies. The result was a major expansion of the private sector, particularly in the cities. In 1996, nine out of 10 private health facilities and one out of four pharmacies were located in one or the other of the country's two largest cities. According to Ministry of Health statistics, between 1991 and 1996, the number of private health facilities went from 106 to 155, and the number of pharmacies from 40 to 69 (Bodart et al. 2001). There were 200 private pharmacies in operation in 2002, the vast majority (84%) being concentrated in the two cities of Ouagadougou and Bobo-Dioulasso (Ministère de la santé 2002). Nevertheless, there is general agreement that these figures are not very useful, as they do not take into account the number of undeclared providers, nor the widely practiced private services offered by public sector professionals.
The expansion has therefore been chaotic, with the State having left the sector to its own devices. In fact, neither the State nor the professional organizations exercise any form of effective regulation, and there seems to be no intention to control this activity, which led one of our key informants to confide that "actually, we are not engaged in private service, but rather in piracy." Laws regulating this sector are not applied, and there are no constraints on who may offer services on the market, nor on the quality or safety of their treatments. The fact that most establishments are not listed anywhere and that no one in authority has an up-to-date listing of providers in the private sector is a good indication of the extent to which the regulatory function has been abandoned and of the extreme disorganization of this sector.
The pharmaceutical reform has focused primarily on rationalizing the supply and distribution of generic drugs at affordable prices. In 1992, a central buying office for essential generic drugs (centrale d'achat en médicaments essentiels génériques – CAMEG) was created, and a policy for essential drugs (politique de médicaments essentiels – PME) was implemented in 1994 (Ministère de la santé 1996). At the district level, this reform was carried out through community measures: setting up EGD depots to supply the health centres, training staff for prescribing and management, and creating community-based management committees. These committees set fees for treatments and medications charged to patients and are also responsible for public education.
The 1994 currency devaluation occurred just as the PME was being implemented. This had an immediate impact on the cost of imported drugs, prompting the authorities to accelerate the setting up of EGD depots, even though it also meant hurrying the process of setting up the committees and training staff and even by-passing certain measures. Government referred to this as the acceleration phase of the Bamako Initiative (Ministère de la santé 1999). The result was that by 1997, 85% of health facilities had already built and equipped EGD depots. One study carried out in Nouna showed that these expediencies had the effect of improving not only utilization but also acceptance of EGDs among the population.
Nevertheless, by the end of the 1990s, implementation in the districts was still not complete. For example, in Dédougou, 62% of health facilities still did not have an EGD depot (Krause et al. 1999). Moreover, early studies tended to show that, as elsewhere in the subregion, rationalization measures had a limited impact on inappropriate prescribing and use of medicines (Krause et al. 1999). Of greater concern, as we shall discuss later in this chapter, were the prices for essential drugs being supplied that were actually not as affordable as expected.
The final element of health system reform was hospital reform, which focused on redefining the roles of each level in the hospital hierarchy, the governance of the institutions, and the improvement of operations. Interventions in this arena included efforts to upgrade the pool of equipment, strengthen human resources and the organization of care, and improve the quality of services. Hospital reform has been slower to get off the ground and so was not examined in depth in this study.
Public expenses in health have increased in both nominal and real value (Figure 2). The nominal budget was 7.7 billion CFAF in 1990 and reached 33 billion CFAF by 2000, representing a public expenditure of less than five dollars (USD) per capita. Setting aside the devaluation year, in which inflation went over 20%, real spending has been increasing steadily over the decade. The proportion of
FIGURE 2. PUBLIC SPENDING ON HEALTH IN BURKINA FASO, 1980–2000.
Source: MAPHealth study (2001).
health expenditure within overall public expenditure increased by nearly five percentage points between 1995 and 2000, demonstrating the particular attention given to this sector by government. These amounts include contributions from external partners, who played a key role in the expansion of public contributions. From 3.6 billion CFAF in 1986, their commitments grew to 27 billion CFAF in 1997. Between 1992 and 1998, their contributions increased at an average rate of 31% annually. This being said, the actual total of the contribution is most certainly underestimated, as the only figures available are those in the State budget and some agencies fund NGOs directly.
Analysis of public expenditure by major line item (Table 3) shows important changes, particularly in the allocation of resources between salaries and investments. Before the SAP, the budget was primarily devoted to salaries, which accounted for 51% of the health budget between 1986 and 1991. With decentralization and the inclusion of a portion of salary expenses under the heading of current transfers, it became difficult to calculate the salary component in public expenditures, but we can reasonably estimate that, from then on, it was about 40%. One of the most striking aspects of the reform was that the SAP, with the commitments of external partners, greatly benefited capital investments.
TABLE 3. BUDGET ALLOCATIONS (%) IN BURKINA FASO BY MAJOR LINE ITEM, 1990–1999
Item |
1990 |
1994 |
1997 |
2000 |
Staff |
54 |
37 |
18 |
19 |
Supplies and equipment |
12 |
12 |
7 |
12 |
Current transfers |
3 |
19 |
22 |
27 |
Capital investment |
31 |
32 |
53 |
42 |
Total |
100 |
100 |
100 |
100 |
Source: Ministère de la santé, Burkina Faso. Direction des études et de la planification. |
||||
Nevertheless, the allocation of resources at different levels of the hierarchy is still unbalanced (Table 4). Health system reform has not resulted in any notable reallocation of public hospital funding to the primary care level, nor from the centre out to the periphery. Government is spending more on health but without much better result, and budgets have not been redeployed to improve service to a population that is still 80% rural.
TABLE 4. DISTRIBUTION OF THE PUBLIC HEALTH BUDGET IN BURKINA FASO, 1998
Item |
% |
Central hospitals |
22 |
Central administration |
29 |
Regional hospitals (9) |
9 |
Regional administrations (11) |
1 |
Health districts (53) |
14 |
Other expenses (school of public health, national laboratory, training) |
25 |
Total |
100 |
Source: Bodart et al. (2001). |
In the absence of any nation-wide study on consumption, it is difficult to estimate the proportion of private spending on health within total expenditure. For the period 1997–2003, UN agencies reported a value of 29.3% (World Bank 2003). Total private spending, at about 1 600 CFAF per capita, would therefore have been less than half as much as total public expenditure. These values are highly improbable—in a subregion where it is generally estimated that private spending on health care exceeds public spending, Burkina Faso would be exceptional. Studies by Peters et al. (1999) led them to estimate a value of 61% for private spending on health in sub-Saharan Africa.
Because the MAPHealth household surveys do not represent the entire country, we cannot use them to produce aggregate national estimates. Nevertheless, even in the most remote areas, the results of our observations grossly exceed the values of the WHO and the World Bank; average per capita spending, in all 53 participating communities, was never less than 3 000 CFAF (Table 5). It therefore seems reasonable to assume that, despite the best efforts of government and its partners, private spending on health in Burkina Faso probably exceeds public spending, in both rural and urban areas.
TABLE 5. PRIVATE EXPENDITURE (IN CFAF) ON HEALTH PER CAPITA IN BURKINA FASO, MAPHEALTH STUDY ESTIMATE
Environment |
Average expenditure |
Standard deviation |
Minimum |
Maximum |
Urban areas |
12 699 |
25 326 |
10 809 |
14 589 |
Semi-urban areas |
12 818 |
46 504 |
7 307 |
18 328 |
Rural areas |
4 226 |
13 412 |
3 180 |
5 272 |
Source: MAPHealth study (2001). |
||||
As we have seen, partial cost recovery has become a generalized practice in public establishments. Unfortunately, there are no national statistics to show the size of the communities' contributions to the financing of health facilities' services. A 1995 study in Bazega and Gourda concluded that the average monthly revenue of a CSPS was 137 633 CFAF; but the disparities are significant, and a similar study in the region of Bobo-Dioulasso reported monthly revenues of 476 800 CFAF (Nacoulma et al. 1995).
At first sight, the few studies done on EGD sales suggest that cost recovery works well, from the perspectives of the depots and the health facilities. Between July 1997 and June 1998, the health region of Kaya as a whole had EGD revenues of more than 27 million CFAF, coming entirely from user payments (Chabot et al. 1999). In the province of Gnagna, the depots had, on average, annual revenues of 3 million CFAF each. Thus, each EGD depot recovers most of its operating costs (Condamine et al. 1999). It is important, however, to put these results in perspective, as the cost recovery is only partial. The results do not take into account salaries, technical assistance, supervision, depreciation of equipment, or of any other items; taking these into consideration, it becomes evident that the recovery applies only to a small part of the real costs of the facilities. Simulations done in Mali have shown that, of 26 health facilities with positive financial results, only two would remain positive if depreciation were factored in (Blaise et al. 1997).
In the case of Burkina Faso, the SAP entailed significant increases in health-care staff. Unlike other ministries, the Ministry of Health was able to embark upon an active recruitment process. Between the periods 1987–91 and 1992–1999, staffing in the public health sector increased 40%. Except for pharmacists (–25%), of whom many took advantage of the liberalization to move into the private sector, all other categories of public-service staff experienced significant increases in numbers: doctors (+37%), dental surgeons (+32%), midwives and state licensed nurses (+30%), childbirth aides (+112%), and field officers (+113%). The rate of increase in health-care staffing was greater than the rate of population growth, leading to an overall reduction in the population–health professionals ratio, as shown in Table 6.
TABLE 6. EVOLUTION OF DENSITY AND DISTRIBUTION OF PUBLIC SECTOR STAFF, BURKINA FASO
|
Ratio per 100 000 population |
Proportion of staff assigned to rural areas (%) |
|||||
Category |
1990 |
1999 |
Change 1990–99 (%) |
1988 |
1991 |
1998 |
Change 1988–98 (%) |
Doctors |
29.3 |
20.0 |
-32 |
55 |
51 |
48 |
-13 |
Pharmacists |
80.8 |
90.1 |
12 |
- |
- |
- |
- |
Dental surgeons |
484.8 |
327.1 |
-33 |
- |
- |
- |
- |
Midwives |
28.3 |
23.7 |
-17 |
42 |
38 |
47 |
12 |
State licensed nurses |
7.5 |
6.1 |
-19 |
74 |
68 |
59 |
-20 |
Registered nurses |
7.5 |
7.4 |
-1 |
76 |
69 |
73 |
-4 |
Childbirth aides |
15.3 |
11.1 |
-28 |
79 |
90 |
80 |
1 |
Auxiliary nurses |
28.8 |
11.1 |
-62 |
- |
- |
- |
- |
Source: Ministère de la santé, Burkina Faso. Direction des études et de la planification. |
|||||||
It is difficult to know whether this effort was sufficient to meet the demand for human resources. First, even though progress has been made, the overall level of resources remains modest, well below the norms set by the WHO of one doctor per 10 000 inhabitants, and one pharmacist and one midwife per 20 000 inhabitants. Also, and more importantly, these recruitments have not rectified the imbalance in allocation of personnel. The increase in human resources has not been uniform across the country. Some zones, such as the urban areas and the central and western regions of the country, are better endowed with health professionals than are the northern and eastern regions.
The percentage of physicians working in rural areas decreased between 1988 and 1998 (Table 6); this is partly due to some moving into the private sector (including NGOs), which is largely concentrated in urban centres. But this is only part of the story, because during this period the Ministry was actively recruiting physicians. Country-wide studies on workload undertaken in 1998 demonstrated that there was 10 times more health-care personnel in Bobo-Dioulasso than in the other cities (Ouedraogo 1999). More than half of all physicians (57%) and midwives are concentrated in Ouagadougou and Bobo-Dioulasso, where 16% of the population resides (Bodart et al. 2001). In the rural areas covered in the MAP Health study, there were twice as many inhabitants per nurse and four times as many inhabitants per midwife as in the urban centres (Table 7). This is not counting the fact that some of the personnel assigned to the interior never actually assume their positions and are incorrectly numbered among the resources working in rural areas. In semi-urban areas, the majority of specialized services in reference centres are provided by nurses. In Nouna, for example, one reference centre serves the entire population of the district and has only one physician.
TABLE 7. DENSITY OF PUBLIC SERVICE STAFF IN BURKINA FASO, IN THE REGIONS STUDIED, RATIO STAFF–POPULATION
|
Urban |
Semi-urban |
Rural |
Nurse |
6.856 |
2.048 |
12.499 |
Midwife |
21.057 |
11.473 |
87.483 |
Traditional birth attendant |
7.19 |
7.17 |
12.497 |
Source: Ministère de la santé, Burkina Faso. Direction des études et de la planification. |
|||
The reticence of health professionals to work in the interior can be partly explained by the appeal of a lucrative parallel private practice in the city and the constraints of rural life. The management of human resources and, in particular, the allocation of personnel remain mostly centralized at the Ministry level. It is generally believed that these decisions are largely influenced by political factors and vote-seeking considerations, resulting in excessive centralization and a lack of policies on human resources and career planning.
As a consequence of this indifferent management of human resources, motivation among health-care workers in the public service is low and absenteeism high. Nearly 80% of the workers we interviewed were unhappy with their schedules and workloads. Half of them said they were unmotivated. Fewer than half were satisfied with the training they had received on the job, and nearly 60% were dissatisfied with the way in which they were being supervised. Forty percent were content with their level of involvement in the management of the health centres. This was reflected in the arena of communications, as 39% of respondents felt they were being well informed on their centre's activities. The majority of workers were not satisfied with their salaries (79%) and bonuses (53%). More than half thought their lifestyles had deteriorated in recent years. Two-thirds felt that possibilities for advancement were unsatisfactory. The study of workload showed that, on average, 70 days per year per person were lost for a variety of reasons: training, participation in meetings, re-provisioning of equipment and medicines, etc. (Ouedraogo 1999). One study of seven health districts demonstrated 37% physician absenteeism (Bodart et al. 2001) and another reported 39% (GTZ 2000).
The health system reform did, in some measure, respond to a portion of overall human resources needs. However, the country unfortunately did not take full advantage of development opportunities, in that it did not adequately resolve major inefficiencies in resource use, nor the most glaring pre-existent inequities.
Investment has been largely focused on increasing the number of primary care facilities. Consequently, there has been a significant increase in coverage (Figure 3) such that 80% of the population now live within five kilometres of a public primary care facility.
FIGURE 3. SERVICE COVERAGE BY PRIMARY CARE INSTITUTIONS IN BURKINA FASO, 1980–2000
Source: Ministère de la santé, Burkina Faso. Direction des études et de la planification.
But even there, disparities continue. In the north and east, the catchment areas of health centres and CSPSs are larger than in other regions of the country. Urban areas are also better endowed in infrastructure than rural areas. Our household surveys showed an overall satisfaction with the availability and accessibility of primary care services. In the cities, 12% of respondents were dissatisfied with the time required to get to a health centre and 8% held negative opinions about the hours of service (Table 8). The level of dissatisfaction remained low in semi-urban areas, most likely due to the proximity of functional reference centres. The greatest dissatisfaction was in rural areas. Hospital accessibility was felt to be insufficient, owing to the much smaller number of public facilities, particularly in the countryside. The government's master plan calls for the creation of nine regional hospital centres and two national hospitals with an average catchment radius of 89 kilometres, which would involve significant travel time for users. For this reason, district medical centres were programmed to receive additional surgical satellites and the number of planned emergency facilities was increased. However, these facilities are not yet all operational.
TABLE 8. UNFAVOURABLE OPINIONS REGARDING PHYSICAL ACCESSIBILITY OF SERVICES, BURKINA FASO, (%)
|
Physical accessibility |
||
Institution |
Travel time |
Open hours |
Waiting time |
CSPS |
|||
Urban |
11.9 |
8.6 |
31.2 |
Semi-urban |
5.7 |
4.6 |
21.4 |
Rural |
20.1 |
7.3 |
10.2 |
Pharmacist |
|||
Urban |
8.0 |
3.1 |
9.2 |
Semi-urban |
9.5 |
7.8 |
11.4 |
Rural |
17.8 |
5.6 |
5.5 |
Public hospital |
|||
Urban |
18.7 |
12.4 |
49.9 |
Semi-urban |
14.6 |
11.6 |
24.9 |
Rural |
40.8 |
18.1 |
19.0 |
Source: MAPHealth study (2001). |
|||
Primary care facilities are expected to provide a standardized range of health care, preventive, and health promotion services, as defined in the minimum activities package. Visits to 14 of the health facilities in our sample demonstrated that service continuity is not constant, but rather is subject to the uncertainties of resource availability. Most health facilities have recurring shortages of vaccine stocks that disrupt their activities and affect vaccination coverage. The consequences are particularly serious for populations looked after through outreach programs; stock shortages at the time of wintering can make it impossible to reach them for another year. DTC shortages were greater in rural areas (60.5 days per year in the Nouna health district) than in urban areas (35 days on average for Bobo-Dioulasso).
Eleven of the 14 centres provided effective nutritional monitoring of children in their catchment area. During the previous year, two CSPSs had run out of oral contraceptives for a month, and one maternity centre ran out for two months. Two centres suffered serious shortages (30 days) of high-use medicines such as chloroquine, paracetamol, and salicylic acid. A study carried out by GTZ in Nouna found that 53% of the health facilities did not have an EGD depot, 67% offered at least one form of contraception, and 89% offered well-baby clinics daily.
It is generally agreed that significant progress has been made in the availability of basic services. Seventy percent of workers interviewed felt the availability of EGDs had improved in their centre, and 80% of user respondents thought the medicines were of good quality. Seventy percent of workers were satisfied with the quality of their equipment. Our own observations, furthermore, showed that the condition of buildings, the level of equipment, and the quality of assigned personnel were satisfactory in the majority of cases. However, these marks of progress do not conceal the enormous problems posed by shortfalls in consumable supplies that compromise the continuity of care and, more generally, by the shortcomings in service quality repeatedly pointed out by users, service providers, and health authorities.
The MAPHealth study did only a partial analysis of service quality. Nevertheless, as mentioned earlier, it did demonstrate the vulnerability of service activities to supply shortages. We have also discussed the lack of motivation and high levels of absenteeism among the staff, and we can assume these also have a significant effect on service quality. The study in the health facilities also revealed that more than half the appointed staff had not been trained to use the therapeutic guide provided to staff for guidance in diagnosis and treatment. Registers are not properly completed, and supervision is more in the nature of follow-up than training. There are no formal mechanisms for quality assurance, and neither public nor private service providers are subject to any quality control measures, whether from professional organizations or from the authorities. A study carried out in the Nouna region showed significant deficiencies in the quality of prescription, with frequent dosage errors and with contraindicated medicines being prescribed to one-third of pregnant women (Krause et al. 1999). A 1998 Ministry of Health opinion survey found that 66% of users were not satisfied with their reception in health service centres, particularly in cities (78%). Finally, our household survey showed that in the city of Bobo-Dioulasso, where the existence of different types of service allows for comparative analyses, private service providers are judged more favourably than the different types of public health facilities (Table 9). The private sector is preferred to public health centres for the five elements explored, and in particular for the technical quality of care, the quality of interpersonal relations, and accessibility.3 As for the public hospital, despite its status as the national hospital, it was less favourably judged on three of the five criteria. These findings are confirmed by an analysis of the maximum amount that the households of Bobo-Dioulasso are willing to pay for services from existing providers. On average, this maximum amount is 7 080 CFAF for services in public centres, 11 230 CFAF for private services, and 10 070 CFAF for the national hospital.
TABLE 9. PERCEIVED QUALITY OF SERVICE PROVIDERS IN URBAN AREAS, BURKINA FASO: MEAN AND STANDARD DEVIATION (N=600)
|
Public hospital |
CSPS |
Private provider |
Traditional provider |
Religious centre |
Technical quality |
2.69 |
2.66 |
2.81 |
2.75 |
2.81 |
Interpersonal quality |
2.33 |
2.55 |
2.91 |
2.9 |
2.8 |
Equipment and supplies |
2.77 |
2.48 |
2.69 |
2.49 |
2.55 |
Accessibility |
2.03 |
2.28 |
2.31 |
2.69 |
2.47 |
Cleanliness and hygiene of the facilities |
2.75 |
2.8 |
2.94 |
2.6 |
2.92 |
Total scores |
2.53 |
2.56 |
2.75 |
2.71 |
2.72 |
Source: MAPHealth study (2001). |
|||||
In brief, everything points to the fact that even if it did serve to extend and consolidate public services, the health-care reform did not resolve the most glaring deficiencies in quality of care. The 1998 observation by a mission of experts, to the effect that the activities of the health districts were not results-oriented in terms of service offered to the population (World Bank 1998), seems unfortunately to be still true.
As elsewhere, the liberalization of the private sector did not lead to a reduction in the price of services; between 1988 and 1999, fees for private consultations rose by 100–150% and for delivering a baby, 20–30% (Quinti et al. 1998). Costs for transferring patients for care outside the country have become so prohibitive that the government has greatly reduced the number of such medical evacuations. On the other hand, the currency devaluation in January 1994 had an immediate effect on prices. The costs of supplies and vaccines nearly doubled and the average price increase for drugs was 76% (Nyamba 1994).
What about drugs in the public sector, given that one of the central objectives of the reform was to guarantee access to essential drugs at affordable prices? As previously noted, the government's voluntarist policy had encouraged the rapid setting-up of a system of depots and EGD distribution covering most of the country. But high margins add to their costs, and Burkina Faso now has the reputation of being one of the countries in that region where drugs are the most expensive. Some drugs are even more expensive in the private sector than in the public one (Afogbe 1997).
In principle, the margins applied by wholesalers, warehouse distributors, and retailers are defined by the Ministry and the CAMEG. In practice, however, margins are higher. Although Ministry directives set the margins for warehouse distributors at 10%, they often go as high as 37% (Chabot et al. 1999). Our calculations estimate that a drug purchased for 100 CFAF by the CAMEG would, in the end, be sold to the patient for 172 CFAF, or nearly double. The population also finds EGDs expensive. In our study, two out of five respondents were dissatisfied with the cost of medicines in the health centres, and one out of two in hospitals (Table 10). Of course, there are somewhat more who find medicines sold in pharmacies to be too expensive, but the difference is modest, and pharmacists are nevertheless very much appreciated, especially in the countryside. Notably, it is mainly in cities that medicines sold in pharmacies and hospitals are considered too costly. It would be interesting to explore the source of these differences further. Perhaps they are related to the selective deterioration of living conditions among that part of the urban population who became impoverished under the adjustment period.
TABLE 10. UNFAVOURABLE OPINIONS REGARDING FINANCIAL ACCESSIBILITY OF SERVICES, BURKINA FASO, (%)
|
Financial accessibility |
||
Institution |
Visits |
Medicines |
Payment options |
CSPS |
|||
Urban |
25.7 |
37.7 |
53.6 |
Semi-urban |
24.8 |
42.1 |
50.5 |
Rural |
17.9 |
41.4 |
30.1 |
Pharmacist |
|||
Urban |
34.5 |
65.2 |
60.5 |
Semi-urban |
20.3 |
52.2 |
46.1 |
Rural |
17.7 |
44.9 |
31.0 |
Public hospital |
|||
Urban |
50.3 |
55.0 |
59.9 |
Semi-urban |
30.4 |
51.4 |
52.3 |
Rural |
22.0 |
40.7 |
27.7 |
Source: MAPHealth study (2001). |
|||
Medication represented 80% of the cost of an illness episode for those persons who visited a professional during their illness (Table 11). The consultations cost only one-sixth as much, which explains why respondents, whether in the city or the countryside, are more accepting of service fees than drug costs. In fact, a review of fees applied for services in the 14 health facilities reveals relatively modest prices. A consultation for an adult is, on average, billed at 120 CFAF, and for a child, 108 CFAF; the delivery of a baby, 808 CFAF; a prenatal consultation, 285 CFAF; and a family planning visit, 69 CFAF. In 12 of the 14 health facilities, vaccinations are free; two CSPSs charge 25 CFAF and 125 CFAF. Health workers and school children are exempted from these charges. Finally, people's perceptions of hospital costs are even more critical, particularly in Bobo-Dioulasso, where fees at the national hospital are considered excessive by more than half of respondents.
TABLE 11. AVERAGE EXPENDITURE PER ITEM DURING AN EPISODE OF ILLNESS, AMONG PERSONS HAVING CONSULTED A HEALTH-CARE PROFESSIONAL, BURKINA FASO, 2001
Item |
N = |
CFAF |
Visits |
236 |
1 054 |
Tests |
313 |
717 |
Hospitalizations |
49 |
1 607 |
Travel |
420 |
310 |
Meals |
420 |
165 |
Medicines |
420 |
6 529 |
Total cost of episode |
420 |
8 320 |
Source: MAPHealth study (2001). |
We asked households to indicate whether they had, in recent years, had difficulty meeting any of their basic consumer needs—food, clothing, schooling of children, health, etc. They were asked to respond to 12 questions, three of which related to health. On average, they had experienced difficulties with five out of 12 needs (40%). Seventy percent of families reported having had difficulties meeting expenses for food, and the same proportion said they had had, at one time or another, difficulties coping with health-care expenses. Multiple regressions (Table 12) showed that poor households, those with an uneducated head of household, and those that live in the countryside far from any health centre encounter many more difficulties in dealing with their health-care needs.
TABLE 12. FACTORS ASSOCIATED WITH THE OCCURRENCE OF DIFFICULTIES IN MEETING BASIC NEEDS. MULTIVARIATE MULTILEVEL REGRESSION: ESTIMATES OF PARAMETERS, STANDARD DEVIATION, AND PROBABILITIES
|
Model 1 (bin.) |
Model 2 (cont.) |
Model 3 (cont.) |
||||
|
Health care needs (yes–no) |
Other needs (cont.) |
All needs (cont.) |
||||
|
B |
SE |
OR |
B |
SE |
B |
SE |
Intercept |
0.92 |
0.19 |
|
2.77 |
0.21 |
4.56 |
0.30 |
Household size |
|||||||
5–9 |
0.33 |
0.14 |
1.40 |
0.54 |
0.16 |
0.88 |
0.22 |
10+ |
0.59 |
0.16 |
1.80 |
1.00 |
0.17 |
1.51 |
0.24 |
Education |
|||||||
None, or primary |
-0.31 |
0.16 |
0.73 |
-0.10 |
0.18 |
-0.25 |
0.25 |
High school or university |
-0.95 |
0.19 |
0.39 |
-0.57 |
0.22 |
-0.12 |
0.32 |
Independent study |
-0.04 |
0.16 |
0.96 |
-0.42 |
0.17 |
-0.57 |
0.24 |
Employment |
|||||||
Employed |
-0.27 |
0.30 |
0.77 |
-0.49 |
0.37 |
-0.64 |
0.52 |
Merchant, tradesman |
-0.82 |
0.20 |
0.44 |
-0.93 |
0.24 |
-1.42 |
0.34 |
Environment |
|||||||
Urban |
-0.06 |
0.18 |
0.94 |
0.87 |
0.18 |
0.85 |
0.27 |
Random effects |
Variance |
SE |
|
Variance |
SE |
Variance |
SE |
Level 1 (communities) |
0.19 |
0.07 |
|
0.10 |
0.06 |
0.35 |
0.15 |
Level 2 (households) |
1.00 |
0.00 |
|
6.09 |
0.20 |
12.16 |
0.36 |
Source: MAPHealth study (2001). |
|||||||
* bold: p<0 | |||||||
bin.: binomial dependent variable–logistical modeling | |||||||
cont.: continuous dependent variable–linear modeling | |||||||
B: beta; SE: standard error; OR: odds ratio | |||||||
The country is poor and living conditions are hard. One out of two heads of families (53%) reported having encountered new difficulties during the previous five years, and one out of three (35%) had encountered new difficulties during the previous year. When the question is specifically focused on new difficulties in meeting health expenses, these proportions become, respectively, 30% and 20%. Nine percent of households said they had not been able to provide care for a child during the previous five years, and just as many said they had had to delve into their savings to care for a family member (Table 13).
TABLE 13. NEW DIFFICULTIES ENCOUNTERED BY HOUSEHOLDS IN THE PREVIOUS FIVE YEARS, BURKINA FASO
|
Environment |
||
Proportion (%) of households who: |
Urban |
Semi-urban |
Rural |
Withdrew a son from school |
8.3 |
8.7 |
2.4 |
Withdrew a daughter from school |
5.3 |
6.5 |
1.9 |
Gave up land or cattle |
4.7 |
27.2 |
35.2 |
Reduced their assistance to relatives |
21.2 |
13.8 |
5.4 |
Sought help from relatives |
16.0 |
15.9 |
14.2 |
Sold agricultural stocks |
1.3 |
12.7 |
21.6 |
Pawned some of their belongings |
2.7 |
1.4 |
1.1 |
Borrowed from relatives |
11.2 |
11.2 |
6.9 |
Gave up caring for children |
8.7 |
10.1 |
9.8 |
Drew upon savings to pay for health care |
5.9 |
12.0 |
6.8 |
Gave up land to treat a sick person |
0.3 |
0.4 |
0.6 |
Source: MAPHealth study (2001). |
|||
The study showed that consumption of health care is strongly constrained by households' ability to pay. Seventy percent of those who had been ill in the previous four weeks believed they would have been better able to care for themselves if they had had more money, and 60% felt they had not received good care for the same reason. These constraints have an important influence on families' readiness to pay for health services and allow us to describe better the effects of price and fee structures on access to care. One-quarter of all survey respondents declared themselves unwilling to spend anything, on any service, provided in any setting at all. For 75% of households, the maximum amount they were willing to pay for a moderately serious health problem was 2 000 CFAF. The 75th percentile was poised at 5 000 CFAF for a serious illness in an adult, 5 000 CFAF for serious illness in a child, 4 500 CFAF for delivering a baby, and 1 250 CFAF for treatment of a malaria attack. In other words, 5 000 CFAF, or approximately 8 USD, was the maximum price that three-quarters of households were prepared to pay for treating one of their members. Given existing prices, this amount was modest and probably constraining. In any case, it was much inferior to the average cost incurred during one episode of illness or one visit that, as mentioned earlier, was more than 8 000 CFAF.
In this context, the cost of illness in general, and the fee structure for services in the public sector in particular, become sources of inequality and exclusion. Inequality is naturally expressed in the consumption of care and services, as is seen in the curve of health consumption as a function of income and of availability of primary care services (Figure 4). The demand is income elastic and, unfortunately, income elasticity is more pronounced at the point where there is already a confluence of opportunity deficits, i.e., in rural areas where poverty is concentrated and where primary care services are much less available. Inequality is also
FIGURE 4. HEALTH SPENDING IN BURKINA FASO AS A FUNCTION OF HOUSEHOLD INCOME (TOTAL CONSUMPTION PER ADULT EQUIVALENT) AND DISTANCE TO NEAREST HEALTH CENTRE
Legend:
d HC < 5km: distance to health centre less than 5km
d HC => 5km: distance to health centre 5km or more
Source: MAPHealth study (2001).
expressed as the families' capacity to overcome economic barriers to access. Among families in the lowest income quintile, fewer than a quarter of persons had gone to a health centre by the seventh day of illness. In families of intermediate income, 35% had gone for care, and among persons from the highest income stratum, more than half had visited a centre before the end of the first week of illness (Figure 5a). The poorer people are, the less they use public services, and usually later. Women use services less often and later than men (Figure 5b). Finally, for the two rural areas in the study, and likely for the rest of the country as well, living in the countryside and being far from a health centre are serious obstacles to accessing care (Figures 5c and 5d).
Although some simply postpone using the services, others are excluded. Among those who experienced an episode of illness during the preceding four weeks and did not see a professional, 35% stated that it was because they did not have the money. Moreover, two earlier studies have shown that utilization of public services decreased significantly after cost-recovery fees were instituted, which corroborates the findings of numerous other studies showing the regressive nature of such strategies (Yoder 1989; Haddad and Fournier 1995; Gilson 1997; Watkins 1997; Wood 1997). After these fees were implemented, utilization of health facilities for treatment in the Kaya region decreased by 17% (Sakho and Yonli 1997). Over a three-year period, utilization decreased by 15% in health centres that introduced the fees, while it increased by 30% in centres of other regions that had not (Ridde 2003).
Certainly, the reform has removed physical barriers to access, and essential services such as generic drugs are now more available than they were before. These results nevertheless conceal significant territorial inequities; improvement in resource availability has not removed existing obstacles. Economic barriers persist for those whose social condition, gender, or location of residence result in a concentration of opportunity deficits, i.e., those whom these reform policies were originally intended to serve.
FIGURE 5. LIVING CONDITIONS, OPPORTUNITY DEFICITS, AND INEQUITIES IN ACCESS TO CARE IN BURKINA FASO. CUMULATIVE PROBABILITIES OF USING PUBLIC HEALTH CARE SERVICES, AS A FUNCTION OF INCOME, GENDER, GEOGRAPHICAL LOCATION, AND AVAILABILITY OF HEALTH RESOURCES
Source: MAPHealth study (2001).
The government's specific objective was to create a new dynamic that would increase attendance at public primary care health facilities. Unfortunately, the results have been disappointing. None of the strategies adopted—investment in additional facilities, institutional changes, making EGDs more available and improving their use, and efforts to involve communities more in governance and management of facilities—has led to an increase in the level of utilization of public health facilities, which is still among the lowest in West Africa. For many people, access to public services remains a privilege (Figure 6). Given the Ministry's ineffectiveness in meeting the very real challenges of allocative deficiency, poor service quality, concentrations of opportunity deficits, and economic barriers for much of the population, the health system reform is, in this respect, a bitter failure.
FIGURE 6. UTILIZATION OF PRIMARY CARE PUBLIC SERVICES IN BURKINA FASO, 1984–2001
Source: Ministère de la santé, Burkina Faso. Direction des études et de la planification.
More than two-thirds of families stated that their health expenses had increased in the previous two years. Among the 802 persons who had been ill in the four weeks before being surveyed, one out of 20 had to sell food products to cover health expenses, one out of 15 had taken out a loan, one out of 15 had taken money from savings, and 14 had to pawn a household belonging. In cities, there is more recourse (9.9%) to taking out loans, whereas in rural and semi-urban areas, people are more likely to sell their food supplies (8% rural, 5.8% semi-urban) or take money from savings (2% urban, 11.6% semi-urban, 6.4% rural). These results are evidence of adaptive behaviors linked to the local economic context and available resources. Urban density and the more common use of money in transactions make recourse to loans easier in cities, whereas the same level of poverty among peasants is more likely to incite them to use savings or sell food stocks.
As with utilization, an analysis of households' consumption of health services sheds light on the overall performance of the system, as this consumption is also, to some extent, the product of many factors that can facilitate or constrain the capacity to receive care. In the prior 12 months, 73% of households had incurred health expenses for at least one member. In this country where families must devote most of their income to meeting their needs for food, spending on health care is constrained to a very modest level (Table 14). These statistics must be kept in context, remembering that health spending is elastic in relation to income and is largely constrained by the household's ability to pay.
TABLE 14. ALLOCATION OF ANNUAL CONSUMPTION OF HOUSEHOLDS IN THE STUDY, BURKINA FASO (IN CFAF)
|
Environment |
||
Item |
Urban |
Semi-urban |
Rural |
Health |
103 920 |
85 807 |
42 767 |
Food |
467 625 |
27 897 |
255 083 |
Education |
93 642 |
33 342 |
13 001 |
Total |
1 100 520 |
605 062 |
394 244 |
Source: MAPHealth study (2001). |
|||
Multivariate regressions help to refine these observations. The model presented in Table 15 aims to identify, simultaneously, factors associated with the incidence of health expenses and with the amounts paid out by the households concerned. It can be seen that income determines both the propensity to spend and the amounts. Looking at propensity to spend in very poor households, i.e., those whose income was less than half the poverty limit established in 1998 of 72 690 CFAF (110 USD) per adult equivalent, the results are gratifying. All other things being equal, their propensity to spend is twice as low as other poor households (those between the half-way mark and the poverty limit), four times less than those whose income is one or two times the poverty limit, and six times less than those families whose income is more than double that limit. At equal income levels, those with secure incomes use more services, most certainly because they are less exposed to seasonal risks. All things being equal, larger households consume more, and more often, than smaller ones, which may be due to having more needs. Households in remote villages, too far from health centres, will consume half as much; this would seem to be a function of poor geographical access and of the difficulties of satisfying health-care needs in that context. Finally, given equal individual capacity, it is in the city of Bobo-Dioulasso that spending is the lowest. To the extent that the model takes into account the effects of income and other individual characteristics, this finding probably illustrates a price effect: the diversity and proximity of service providers and drug retailers, in both the formal and informal sectors, would result in lower prices than are seen in smaller towns and villages.
The economic burden of health care is thus unevenly distributed. Furthermore, analysis of Lorenz curves and coefficients demonstrates that inequality in households' health consumption is more pronounced in cities, where the offer—public as well as private—is more dense and income more unevenly distributed. The Gini coefficient, which measures the extent of inequalities in households' out-of-pocket spending for health care, is 0.49
TABLE 15. FACTORS ASSOCIATED WITH HOUSEHOLDS' CONSUMPTION OF HEALTH CARE. MULTIVARIATE CONDITIONAL REGRESSIONS: CONDITIONAL PARAMETERS AND PROBABILITIES
|
Model 1 (binominal) Incidence of expenditure (yes–no) |
Model 2 (continuous) Amounts spent* |
|||
Fixed effects |
B |
SE |
OD |
B |
SE |
Intercept |
-0.68 |
0.23 |
0.51 |
-4 171 |
3 174 |
Revenue (Y) |
|||||
0.5PL<Y<1PL |
0.80 |
0.20 |
2.23 |
2 262 |
3 091 |
1PL<Y<2PL |
1.32 |
0.20 |
3.74 |
6 943 |
3 085 |
2PL<Y |
1.76 |
0.22 |
5.81 |
28 602 |
3 356 |
Household size |
|||||
1-4 |
|
|
|
4 271 |
2 014 |
5-9 |
0.78 |
0.15 |
2.18 |
|
|
10+ |
1.36 |
0.17 |
3.90 |
|
|
Profession of head of household |
|||||
Employed in private sector |
|
|
|
5 940 |
2 930 |
Employed in public sector |
|
|
|
15 820 |
3 533 |
Environment |
|||||
Small cities (semi-urban) |
|
|
|
8 504 |
2 516 |
Rural |
|
|
|
3 648 |
2 194 |
Distance from CSPS |
|||||
d > 5 km |
-0.54 |
0.19 |
0.58 |
|
|
Random effects |
Variance |
|
SE |
Variance |
SE |
Level 2 |
0.27 |
|
0.09 |
0.00 |
0.00 |
Level 1 |
1.00 |
|
0.00 |
1 161 334 |
41 004 770 |
Covariance |
0.00 |
|
0.00 |
6 699 |
844 |
Source: MAPHealth study (2001). |
|||||
* Health expenditure per adult equivalent |
|||||
B: beta; SE: standard error; OR: odds ratio; |
|||||
PL: poverty line All associations presented are statistically significant to within 5%. |
|||||
at Bobo-Dioulasso and 0.54 and 0.46 in the semi-urban areas of Nouna and Bazega, respectively. In rural areas, it is no more than 0.39 and 0.33. In the city, the inequality in health spending mirrors that of income, and the Kakwani index of redistribution is not significantly different from zero. However, in outlying villages, where primary care services are less accessible and relative prices higher, health expenditures are significantly more unevenly distributed than income (Figure 7). This is to a large extent due
FIGURE 7. COMPARISON OF CONCENTRATION OF INCOME AND OUT-OF-POCKET SPENDING FOR HEALTH IN BURKINA FASO IN VILLAGES MORE THAN 5 KM FROM A CSPS
Legend:
L(p) = Lorenz curve of income inequality (total consumption per equivalent adult)
C(p) = Concentration curve of out-of-pocket spending per equivalent adult
Source: MAPHealth study (2001).
to the financial barriers to access experienced by poorer households and the high income elasticity of demand.
What is the impact of illness and health-care spending on households? This has not been easy to evaluate, because it depends on the extent to which the illness has drawn upon the family's resources and on the loss of income that may be incurred in the short, medium, or long term. Unfortunately, the information available does not allow us to analyze the impact of illness on productivity and household outputs. We have nevertheless estimated some of these impacts, focusing on the monetary impoverishment of households. The evaluation rests on the hypothesis, reasonable in this context, that health expenses reduce the capacity to satisfy essential needs. The available income is reduced by the amount spent on health care. Because the poverty line is defined on the basis of satisfying basic nutritional needs, we can assume that: (i) poor households will become poorer by an amount corresponding to what they have spent on health; (ii) households that were originally not poor but whose residual incomes are below the poverty line, become newly poor; and (iii) monetary impoverishment4 attributable to direct health-care expenses can be measured as follows:
Δ(PI) = [PI(z)2 − PI(z)1]/PI(z)2
Δ(PI) is the impoverishment attributable to direct health-care expenses measured in terms of incidence, z is the poverty line, and PI(z)1 and PI(z)2 are the incidence levels of poverty before and after accounting for the deduction of health expenses.
Δ(PG) = PG(z)2 − PG(z)1
Δ(PG) is the impoverishment measured in terms of the intensity of poverty, z is the poverty line, and DI(z)1 and DI(z)2 are the poverty gaps before and after accounting for the deduction of health expenses from household income. Δ(PG) is also the poverty gap attributable to direct health-care expenses and has the advantage of being expressed in monetary units. It corresponds to the per capita minimum cost of a financial protection strategy in poor households (Ravaillon 1992). It can also be broken down to estimate the relative contribution of different opportunity deficits to overall impoverishment tied to health expenses. Finally, as measures of poverty are sensitive to the level of the poverty line, this parameter is used for sensitivity analysis.
The analysis demonstrates that illness is manifestly a factor of monetary impoverishment. Health expenses contribute to poverty, creating newly poor households and further impoverishing those already poor. They bring the greatest poverty to outlying rural regions where opportunity deficits are concentrated. In these areas, we find economic poverty, isolation, insufficient local solidarity mechanisms to help families in need, poor availability of public infrastructures, and relatively high health-care costs. Veritable poverty traps, where opportunity deficits converge, these areas are where the economic damages of illness are the greatest and where the need for financial protection of households against the risks of illness is most pronounced.
What strategies should the authorities implement to protect these households? Responses to this question are still to be found, as the issues are complex and the situations are specific. Nevertheless, our analysis suggests that part of the impoverishment observed was among families originally living above the poverty threshold. So, it is not enough to develop strategies that target only the poor.
The unpredictable nature of illness and its associated costs suggest the need for financial protection of households via interventions targeting barriers to access and opportunity deficits, rather than focusing on households that have been defined from the start as poor. A variety of strategies could be considered to overcome access barriers and protect families against the financial risks of illness. Some, like support to micro-insurance and mutual benefit societies, primarily target demand. Others target the supply side and might involve improving physical accessibility of public services or reducing prices. Our simulations suggest that targeting poverty traps by ensuring adequate geographical access to public health resources could reduce the poverty deficit attributable to households' direct health-care expenses by 20%. Another 20% of the marginal poverty deficit could be saved if the population were at least minimally covered by health insurance, which could be co-financed by the state and individuals, where the fees would be comparable to those already in place locally, on a smaller scale. The combination of approaches, one focused on supply (geographical access) and the other on demand (health insurance) could reduce impoverishment caused by illness by about half.
The Burkinabé reform was originally full of promise. It could certainly have become a model for others, as it brought together many favourable conditions, including the rare combination of a benevolent Minister of Health and an adjustment that aligned sectoral reform and expansion of public resources with a mandatory macroeconomic adjustment exercise. Burkina Faso's healthcare reform brought together several convincing strategies aimed at improving the efficiency of the health sector, rectifying existing inequalities of resource allocation, reinforcing the governances of the health districts, and, of course, removing barriers and providing access to services of quality.
However, the result has been disappointing, and, to paraphrase Bodart et al., the health-care system itself is a patient still in need of treatment. More has been spent, much more, but not better. Illness continues to impoverish the poor and even the less poor, access to services of good quality evades the majority of the population, regulation and governance leave much to be desired, community participation is reduced almost exclusively to fundraising, resource allocation is still very inefficient, national experts are leaving the country, attendance at publicly funded health centres is hopelessly low, and the overall system—decision-makers, contributors, managers, and those working in the field—continues to evade accountability.
Does this reform represent a lost opportunity? It may be so, as it is difficult to imagine those in charge being able to assemble, again, so many favourable conditions. Of course, they, and particularly the Minister of Health, have this responsibility. But responsibility lies also with development agencies, multilateral and bilateral, that have supported a reform that has not gone far enough in achieving a governance that is more efficient, equitable, and democratic.
1. The currency of Burkina Faso is the CFA franc BCEAO (Banque Centrale des États de l'Afrique de l'Ouest), currency code XOF, symbol CFAF. All references to this currency in this chapter are given as CFAF.
2. The typical district covers a population of 150 000–300 000 and is organized around a network of CSPSs within a radius of 10–12 km, with an appropriate level of administrative and community support.
3. For more detail on the composition, construction, and validation of the quality scores, the reader is invited to consult Haddad et al. (1998).
4. Of course it must be recognized that measuring poverty in this way, i.e., considering only direct, short-term monetary impacts, accounts only for the minimum impoverishment of the household. It does not take into account other losses of productivity and long-term effects on monetary and non-monetary elements of the household's well-being.
Afogbe, K. 1997. Rapport de l'étude sur l'utilisation des médicaments essentiels génériques (MEG). Ministère de la santé, Ouagadougou, Burkina Faso, BASP'96. 72 pp.
Aubel, J.; Sobgo, G. 1998. Étude de cas: un partenariat entre l'ONG Save the Children Pays-Bas et le district sanitaire de Kongoussi au Burkina Faso. Ministère de la santé, Ouagadougou, Burkina Faso. 38 pp.
Blaise, P.; Kegels, G.; Van Lerberghe, W.; Diakité, B.D.; Touré, G. 1997. Coûts et financement du système de santé de cercle au Mali. ITG Press, Antwerp, Belgium. 130 pp.
Bodart, C.; Servais, G.; Lamine, Y.M.; Schmidt-Ehry, B. 2001. The influence of health sector reform and external assistance in Burkina Faso. Health Policy and Planning, 16(1), 74–86.
Chabot, J.; Conombo, S.C.; Meuwissen, L.; Yonli, L. 1999. Rapport de l'évaluation finale du programme d'appui aux soins de santé primaires dans la région de Kaya (PASSPK). Ouagadougou, Ministère de la santé and Save the Children Pays Bas (SCPB). 86 pp.
Condamine, J.-L.; Artigues S.; Midol, S.; Buisset, K.; Burkina, K.B. 1999. Programme de soutien à la mise en place de l'Initiative de Bamako: analyse de la création de huit aires de santé, province de la Gnagna. Cahier Santé, 9(5), 305–311.
Gilson, L. 1997. The lessons of user fee experience in Africa. Health Policy and Planning, 12(4), 273–285.
Government of Burkina Faso. 2001. Rapport du conseil des ministres du 7 février 2001. Ouagadougou, Burkina Faso.
GTZ (Deutsche Gesellschaft für Technische Zusammenarbeit, GmbH). 2000. Étude sur la mise en place des médicaments essentiels et génériques dans les régions sanitaires de Gaoua et de Dédougou. GTZ, Dédougou, Burkina Faso.
Haddad, S.; Fournier, P. 1995. Quality, cost and utilization of health services in developing countries: a longitudinal study in Zaïre. Social Science and Medicine, 40(6), 743–753.
Haddad, S.; Fournier, P.; Potvin, L. 1998. Measuring community perceptions of the quality of primary health care services in developing countries. Validation of a 20-item scale. International Journal for Quality in Health Care, 10(2), 93–104.
INSD (Institut National de la Statistique et de la Démographie); Macro International Inc. 2000. Enquête démographique et de santé, Burkina Faso 1998–1999. Macro International Inc., Calverton, MD, USA.
Krause, G.; Borchert, M.; Benzler, J.; Heinmüller, R.; Kaba, I.; Savadogo, M.; Siho, N.; Diesfeld, H.J. 1999. Rationality of drug prescriptions in the rural health centres in Burkina Faso. Health Policy and Planning, 14(3), 291–298.
Ministère de la santé. 1996. Document cadre de politique pharmaceutique nationale. Ouagadougou, Burkina Faso. 17 pp.
_____ 1999. Revue de la mise en oeuvre de l'initiative de Bamako au Burkina Faso. Ouagadougou, Burkina Faso. 15 pp.
_____ 2002. Annuaire statistique 2002. Direction des études et de la planification. Ouagadougou, Burkina Faso. 165 pp.
Ministère de l'économie et des finances. 2000. Burkina Faso: cadre stratégique de lutte contre la pauvreté. Ouagadougou, Burkina Faso. 65 pp.
Nacoulma, D.; Petitjean, F.; Sanou, A.; Some, M. et al. 1995. Une expérience de mutuelle en zone rurale au Burkina Faso. Paper presented at the IVe Journées des Sciences de la Santé de Bobo-Dioulasso, 1995. 15 pp.
Nyamba, A. 1994. Un an après la dévaluation du Franc CFA: la réponse des consommateurs. Ouagadougou, Burkina Faso.
Ouédraogo, G.J. 1999. Étude sur la charge de travail dans les formations sanitaires. Rapport final. Ministère de la santé, Direction des études et de la planification. Ouagadougou, Burkina Faso.
Peters, D.H.; Kandola, K.; Elmendorf, A.E.; Chelleraj, G. 1999. Health expenditures, services, and outcomes in Africa: basic data and cross-national comparisons, 1990–1996. Report no. 19578. Health, Nutrition, and Population Series. World Bank, Washington, DC, USA. 55 pp.
Quinti G. et al. 1998. Etude de faisabilité d'un projet de santé au Burkina Faso dans le cadre du 8e FED (Fonds européen de développement). Rapport de faisabilité définitif. CERFE, Rome, Italy.
Ravaillon, M. 1992. Poverty comparisons: a guide to concepts and methods. Living standards measurement study. Report no. LSM88. World Bank, Washington, DC, USA. 138 pp.
Ridde, V. 2003. Fees-for-services, cost recovery, and equity in a district of Burkina Faso operating the Bamako Initiative. Bulletin of the World Health Organization (WHO): the International Journal of Public Health, 87(7), 532–538.
Sakho, M.; Yonli, L. 1997. Rapport de backstopping sur le recouvrement des coûts dans la région de Kaya. Ministère de la santé, Ouagadougou, Burkina Faso.
Watkins, K. 1997. Cost-recovery and equity in the health sector: issues for developing countries. Paper presented at World Institute for Development Economics Research (WIDER) Project on Provision and Financing of Public Goods in Developing Countries, 27 February 1997, London, UK. Oxfam, Oxford, UK, and Ireland Policy Department. 31 pp.
Wood, A. 1997. Annotated bibliography on user charges in developing countries. Discussion paper prepared for the ODA Research Work Programme "Improved care of diseases of childhood". Centre for International Child Health (CICH), Institute of Child Health, London, UK. 50 pp.
World Bank. 1998. Annexe de l'aide mémoire. Mission de supervision du PDSN (Plan de développement sanitaire national), 1–29 March 1998. Ouagadougou, Burkina Faso.
Yoder, R.A. 1989. Are people willing and able to pay for health services? Social Science and Medicine, 29(1), 35–42.
Health system reforms have been a regular occurrence in countries around the world for several decades. Many have been aimed primarily at controlling health expenditures and making health systems more efficient. Others have been driven by political agendas promoting universal coverage, solidarity, equity, and social participation. Most, however, have been both politically and economically motivated. During the 1990s, the pace of reforms accelerated. In some countries, this occurred in response to internal political changes, but in many others it was caused by external pressures from international organizations. A common thread has been the search for universality and efficiency, with most health reforms of the 1990s being market-oriented. Decentralization has also been a strategy of many reforms, being seen as effective in stimulating service delivery, better allocating resources according to needs, and involving communities in decisions on priorities, so as to reduce inequities.
In this chapter, we examine the equity effects of health sector reforms in Colombia resulting from macroeconomic adjustment policies introduced over the course of a 20-year period (1980–2000), focusing on changes introduced in the health-care system during the latter half of that period. We document increases in public social expenditure and particularly in health expenditure (public and private) as well the effects on equity in terms of financing, targeting of subsidies, health insurance coverage, and access to health services in an economy under profound recession.
This was a retrospective time series study with several points of observation, using secondary data from government sources in health and economics, as well as from several household surveys carried out by different institutions over almost two decades. Additionally, in two selected regions—one with the highest and one with the lowest level of development—specific data collection was done through a survey of health facilities (hospitals, health centres, and health posts1), individual interviews, and focus groups with executives at municipal levels, as well as through content analysis of documentation about national health and economic policies.
The results presented in this chapter are based on both government and cross-sectional secondary data. Macroeconomic analysis used secondary data from the Central Bank (Banco de la República), the National Planning Department (Departamento Nacional de Planeación–DNP) and the Statistics Department (Departamento Administrativo Nacional de Estadística–DANE). Cross-sectional data were taken from 1993 and 1997 Quality of Life Surveys carried out by DANE.
Located at the northwest extreme of South America, Colombia is a lower-middle-income country with a per capita GNI of 1 810 USD in 2002 (World Bank 2003a) and an estimated population of 44 583 577 in 2003, 49.4% males and 50.6% females. Population growth rate in 2003 was 2% (World Bank 2003b) and demographic density was 32.4 inhabitants per km2, with 71% of the country's inhabitants living in urban areas. Indigenous population represents 1.7% of all inhabitants and blacks, 1.6%. In general, the demographic indicators showed steady improvement between 1970–1975 and 1990–1995. On average, life expectancy has increased and both infant mortality and the overall fertility rate have decreased. However, there are still significant regional and socioeconomic differences.
Until 1997, economic growth of about 4.5% per year in the previous four decades, combined with a decrease in the population growth rate to 2.05% per year, led to substantial improvements in social conditions. Since 1997, with the slowdown of economic activity, economic growth (around zero between 1998 and 2002) has been smaller than population growth, with decreases in per capita income and social conditions. The Gini coefficient worsened in the last years and was.57 in 2000 (Ministerio de Salud et al. 2002).
Life expectancy at birth in 2002 was 71.8 years (World Bank 2003b), up more than 21% compared with 1965, when it was 59 years. Primary school enrolment was close to 90% in 1998, and the literacy rate was 97.2% (World Bank 2003b). Nonetheless, poverty remains a critical problem. It was estimated that, in 1992, more than six million Colombians, 19% of the population at that time, had incomes below a commonly accepted subsistence level, with three out of four of these living in rural areas.
During earlier decades, Colombia had made substantial progress in improving the living standards of its population, with poverty declining steadily from an estimated 50% in 1964 to 19% in 1992; but by 2002, poverty had increased again to levels near 30%.
Some 6.7% of children under age five are malnourished, 19% of the population have no access to health services, and about 30% lack access to safe water (1998). Sharp regional disparities in quality of life contribute to social disintegration in large areas of the country. Coverage for basic services at home increased significantly between 1985 and 1993, from 70.5% to 82% for water supply and from 59.4% to 69.9% for sewerage connections. Nevertheless, there are still 6 million persons without water supply and 10 million without sewerage. The situation is more critical in rural areas, where between 5 and 8 million people lack at least one of these services. The gaps are even greater when it comes to water quality. The ratio of the population living in abject poverty (presence of two or more indicators of unmet basic needs) in rural areas relative to those in the municipal seats increased from 2.2 to 5.0 between 1985 and 1993, reflecting improvement in urban areas that did not correspond to the situation observed in rural areas.
Estimated infant mortality in 2002 was 25.6 per 1 000 live births, and maternal mortality was 67.7 per 100 000 live births, with 93% of deliveries being institutional, 91.3% by a physician and 1.7% by nurses or auxiliary nurses (Ministerio de Salud et al. 2002). In 2001, immunization coverage in children under one year old was 77.6% for diphtheria, pertussis, and tetanus, 80.5% for polio, and 87.7% for prenatal tetanus vaccine.
The four greatest causes of mortality (1999) were cardiovascular diseases (121.1 per 100 000 inhabitants), homicides and suicides (67.5), cancer (64.8), and perinatal causes (19.4). Morbidity due to communicable diseases continues to be important. The incidence rate (per 100 000 inhabitants) for malaria was 766.5; for dengue, 180.1; for tuberculosis, 26.2; and for AIDS, 2.3.
Between 1980 and 2000, Colombia had two severe periods of adjustment, the first during the early 1980s with an IMF advisory and the second from 1999 to date. From 1980 to 1995, average GNP growth was 3.8% per year. Between 1995 and 2002, GNP stagnated; whereas growth averaged about 3% per year from 1995 to 1997, there was no economic growth between 1997 and 2002.
Four periods can be distinguished during these 20 years. The first corresponds to the debt crisis of the early 1980s, with a marked decrease in economic growth. The second, from 1985 to 1989 when the first, but incomplete, macroeconomic adjustment policies were implemented, was characterized by partial recuperation of economic activity. The third occurred in the early 1990s, with major structural reforms and rapid GNP growth after a semi-recession in 1991. The fourth, from 1995 to the present, has resulted in the greatest recession in Colombia in the 20th century.
Two main political reforms took place during this 20-year period. The first was the political, fiscal, and administrative decentralization of the State that started in the early 1980s, with substantial transference of autonomy, resources, and responsibilities to the municipal level. This decentralization had very important consequences for the health sector. The second was the formulation of a new Constitution (1991) replacing that of 1886. Under this new Constitution, the Central Bank was given autonomy from government action, transfers from central to local governments were increased (from 4.3% of GNP in 1990 to 7.7% in 1997), and, as a consequence, there was an unprecedented increase in public social expenditure (Figure 1). A labour reform law (Law 50, 1991) changed the labour regime to reduce rigidities and facilitate industrial restructuring. In June 1993, Congress passed legislation to further develop the ongoing decentralization in Colombia and, in December 1993, introduced profound social security reform in both the financial and health sectors.
There are important differences of opinion over the economic consequences of the reforms. There is consensus that the new Constitution increased social expenditure and introduced serious rigidities. It is also generally recognized that regional transfers increased local expenditure without diminishing central expenditure. The reforms did not create new income sources, hence contributing jointly with other internal and external factors to the deep fiscal crisis of the end of the century.
FIGURE 1. PUBLIC EXPENDITURE (AS % OF GDP), TOTAL AND CENTRAL GOVERNMENT, COLOMBIA, 1974–1994
Source: DNP and DANE. Estimates of the Social Mission.
The Colombian health system was reformed several times during the 20th century, in accordance with prevailing conceptions of the State and with the influence of international organizations (Quevedo 1990). In the second half of the 1970s, the National Health System (Sistema Nacional de Salud) was created, with a highly centralized structure headed by a powerful Ministry of Health. This system, which existed up until the most recent reforms, was three-tiered: (i) a private sector, serving mainly but not exclusively the affluent; (ii) a social security sector, which never reached beyond 23% coverage and which was mainly concentrated in the largest urban areas and among workers of the formal sector; and (iii) a public sector, with hospitals, health centres, and health posts serving the poor and uninsured (Diagram 1). Government accepted that 20% of the population was not reached by any sector. Traditionally, the national budget financed health centres, hospitals, and other public health programs. The private sector offered services through insurance or direct payment. The social security institutions financed themselves through their members' payments.
This health system was both inequitable and inefficient: inequitable because real access to health services privileged those with least need, and inefficient because Colombia was already spending 7.4% of its GNP on health with poor results, compared to other countries with similar or lower levels of spending, such as Costa Rica (Yepes 1990).
DIAGRAM 1. HEALTH SECTOR STRUCTURE IN COLOMBIA BEFORE 1993 REFORM
Since the late 1980s, there have been two main sets of reforms, decentralization and reforms to the social security system, both of which changed profoundly the functioning and configuration of the health-care system.
During the period 1986–1990, the health system was decentralized to the municipal levels as part of a general political, administrative, and financial decentralization of the State. Department governors and municipal mayors were elected by popular vote, and administrative competencies were transferred to the municipalities along with the financial resources required to support them. Through decentralization (Law 10, 1990) municipalities assumed major responsibility for the health of their communities, which they had not had before, and a significant number of new local health actors came into the sector (Yepes et al. 1999). Mayors became the heads of municipal health systems, with the duty of developing local health plans integrated into the municipal plans. Municipalities' responsibilities included defining local health policies, health planning, constructing and maintaining health facilities, and providing first-level health services to their communities either directly or by contracting with public or private providers, as well as by developing public health activities.
Decentralization entailed important power shifts among the various governments involved (national, departmental, municipal). The national level was freed of the actual provision of services, allowing it to concentrate on policy formulation and technical assistance. Municipalities assumed control over local health decisions, and departments' control over municipal decisions was reduced. Social participation in health was strengthened by regulating the Local Health Committees (Comités de Participación Comunitaria–COPACOS) and by including the community in boards of directors of public hospitals (one-third representation).
However, the decentralization process has not been easy. Many interests at the national and departmental levels have opposed it in one form or another with some success. In several departments, such as in Cundinamarca, it was systematically opposed and impeded by the departmental governments. This has led to an uneven process, such that decentralization is very advanced in some regions while almost non-existent in others. Recent changes in national legislation (Law 715, 2001) have created even more obstacles to the decentralization process. In fact, this law halted the process in municipalities that had not yet been decentralized by the time it was enacted.
Under the new Constitution of 1991, health and social security became rights. The social security reform introduced a concrete system to attain those rights based on five guiding principles: efficiency, universality, solidarity, comprehensiveness, and social participation. It significantly transformed the health sector and introduced new actors and practices into the system (Diagram 2).
DIAGRAM 2. HEALTH SECTOR STRUCTURE IN COLOMBIA AFTER 1993 REFORM
The health sector reform enacted by Law 100 in 1993 and implemented after January 1995 was extremely ambitious, aiming to integrate the social security and public health sectors and to create universal access to the system, while bringing market pressures to bear on institutions to improve their efficiency. In pursuit of universal access, Law 100 set the framework for a national health insurance system, with the goal of increasing insurance coverage from 20% to 100% by 2001 and ensuring entitlement to a basic package of services (Paquete obligatorio de servicios-POS) for all. By 2003, health insurance coverage was at 53% (up from 23% before the reform, although still far from 100%).2
Under the reform, a new General System of Social Security in Health (Sistema General de Seguridad Social en Salud-SGSSS) was created with two regimes: the contributive for those who can afford to pay and the subsidized for the poor. Anyone with a labour contract or an income above two minimum salaries (if working independently) must be affiliated to the contribituve regime and pay 12% of their declared income. The poor must apply to their municipality and undergo a means test. If they are under a certain level of income, they qualify and are authorized to become affiliated. The number of poor affiliates authorized in each municipality depends upon the resources available. Those with no insurance coverage, either because they do not have a labour contract or are not poor enough to be subsidized under currently available resources, receive care in public hospitals for as long as there are enough resources to universalize insurance.
New entities were also created. The contributive regime insurers (Entidad promotora de servicios-EPS) were wholly new to the system, as were the Compensation Fund and the National Health Social Security Council (Consejo Nacional de Seguridad Social en Salud). The Compensation Fund is a fiduciary account that manages all income from the EPSs, which must report their affiliates and premiums received every month. In this process, called compensacíon, they either compensate the Fund with the surplus collected, if the average premium collection exceeds the capitation, or are compensated by the Fund for any negative difference between the collected premiums and the corresponding capitation. The National Health Social Security Council is a collective body with representation from all system actors (government, employers, workers, professionals, pensioners, insurers, providers, and rural community). This Council makes the most important decisions in the day-to-day functioning of the system, such as defining the content of the POS and the amount of the capitation unit to be recognized to the insurers.
The Ministry of Health and the Health Superintendence (Superintendencia de Salud) were significantly restructured to adapt to the new circumstances, and a large body of regulations was enacted. Since the beginning of the reform, the Ministry of Health has put aside the direct provision of services and has undertaken a continuous process of adaptation to strengthen its normative and regulatory capacity. At the same time, the Health Superintendence was strengthened to increase its capacity to inspect and control public and private insurers and providers. However, it is generally acknowledged that it is overwhelmed by the magnitude of this task.
Public hospitals became autonomous entities called State Social Enterprises (Empresas Sociales del Estado-ESE) and were progressively moved from supply financing (historic state budgets) to demand financing (selling services in the market). Old institutions changed roles, as in the case of the Ministry of Health, now concentrating on sector stewardship while leaving service provision in the hands of the departments and municipalities and of public and private providers. The Health Superintendence became a major actor in the system, responsible for supervision and control. The private sector enhanced its participation as insurer and provider, and the public sector was forced to modernize and become efficient and competitive. Previous social security institutions were either closed or transformed into EPSs. Social participation was reinforced by the creation of consumer alliances.
With the Ministry of Health at the top and the Health Superintendence in charge of control, the system operates with two layers of competition, one at the level of insurers and another at the level of providers.
Insurers compete in both the contributive and the subsidized regimes within narrow limits. All must offer the same package of services set by the Ministry of Health, and all receive the same capitation unit (Unidad de pago por capitación-UPC). Competition rests on quality of services and profits depend on efficiency. Although, by law, all insurers must accept any person asking to be affiliated, to avoid cream skimming, in practice it nevertheless occurs by subtle means.
As of 2003, there were 35 insurers (EPSs) in the contributive regime, eight public and 27 private. In the subsidized regime, there were 33 insurers, nine of them serving the indigenous population.
Updated information on providers is very limited, but it is estimated (2003) that there are approximately 55 000 hospital beds in the country, 60% public and 40% private. All public and private hospitals, as well as health professionals, compete in the market to contract with insurers to provide health services to the insured population.
Between 1994 and 1998, human resources regulations were issued to address such matters as integration and organization of consultative commissions, the salary regime of territorial public health employees, salary increases for health employees in public institutions, and requirements and functions of ESEs. The regulation process was also reinforced. For example, the National Health Institute internal structure and functions were modified (it being assigned the control and vigilance of communicable diseases and public health laboratories). Regulations were also issued related to the new health and social security systems, such as the content of POSs that all insurers must offer to the affiliated population, the operational norms for EPSs, and regulations governing the functioning of ESEs, among others.
The public health component of the reforms had several aspects. Law 100 contemplated that promotion and prevention should be a part of the POS to be offered by insurers. It set a proportion of the UPC specifically for their financing. The law also stipulated that a Basic Care Plan (Plan de atención básica-PAB) including public health activities should be offered free of charge by the territorial authorities.
The functioning of institutions dedicated to extraction, processing, preservation, and transportation of total blood or its derivatives was also regulated, and both the Network and the Council of National Blood Banks were created.
Finally, measures were formulated to prevent people's displacement and to look after, protect, consolidate, and stabilize people who had been socioeconomically displaced by war in Colombia.
The system is financed by a mix of employer-employee contributions and general taxes. The contributive system is financed exclusively by 12% of the declared income contribution, two-thirds paid by the employer and one-third by the employee. Independent workers must pay 100% of the corresponding contribution.
The subsidized system is financed by general taxes collected by the national government and transferred to the municipalities. These resources are reinforced by departmental and municipal contributions plus the transfer of 8.33% of total collections from the contributive system.
There are solidarity mechanisms both within the contributive system and between it and the subsidized system. Within the contributive, paying a fixed percentage over declared income creates solidarity between those earning more and those earning less, and through the use of capitation, the sick are subsidized by the healthy. Between the contributive (richer and with employment) and the subsidized (poorer) systems there is a cross-solidarity, because 8.33% of the collected funds in the contributive system go to the subsidized. Finally, all taxpayers contribute to the financing of the subsidized insurance for the poor.
An in-depth analysis was carried out using two geographical, cultural, and socioeconomic regions deliberately selected to maximize socioeconomic differences. The richer of these (Antioquia) was a subregion of the Antioquia department, corresponding to the municipalities of the Aburra valley, an urban, mostly developed region. The other (Pacifico) was a subregion of the Pacific coast, mostly rural and underdeveloped, with small urban centres.
Analysis was carried out using the 1997 Quality of Life Survey. For our analysis, we defined the following general model to include all variables we thought might explain coverage:
Coverage = f (ability to pay3; income; characteristics4 of the head of household; individual characteristics). For the national level, the model included dummies for zone and strata.
Before proceeding to interpretation of the models, the following comments on the variables should be noted:
— Income and ability to pay presented the expected differences between regions, where Antioquia's mean was higher than the national average and Pacifico's was lower.
— Gender differences were not evident. Over 70% of individuals lived in households headed by a male.
— Age profiles also did not differ between the two regions. For more than 70% of individuals, their household head was aged between 25 and 59 years.
— Education level was higher in Antioquia. Household head education level had the same profile as the individual distribution. More than 70% of individuals lived in households in which the head worked. Regarding other characteristics, there was a higher proportion of pensioner household heads in Antioquia compared with Pacifico and with the national average.
— Differences in economic position were more evident. In Antioquia, household heads worked mainly as private workers or freelancers, whereas in Pacifico freelance, farm, and day workers were very prevalent.
— Antioquia's insurance coverage was 1.35 times that of Pacifico, although Pacifico's was near the national average (55% vs. 57%). The most important cause given for not being affiliated was lack of money, although in Antioquia, lack of employment was also significant.
Because the model gave contradictory results for ability to pay and household income, separate models were run with each of these. The results were statistically significant in all cases.
Logit models were used to estimate the probability of a household having the characteristic studied (health coverage) as a function of a group of explanatory variables.
Logistic regression models allow for modeling the probability of a yes/no outcome as a function of a set of independent variables (such as household income or ability to pay). We used Categorical Data Modeling (CATMOD) in SAS to estimate the effects of individual variables as well as to identify the most predictive model.
To interpret results, we used estimated equations to compute predictions of individual social security coverage, controlling for income or ability to pay (ATP), for persons with different individual and household characteristics, with the following results:
The probability of being covered was higher for those of working age (16–59), those with heads of household 25–59 years old, public workers, and pensioned household heads. There was also a positive relationship between the probability of being covered and the individual and household head level of education. There were no differences of coverage probability by gender, even by household head gender. The probability of being covered grew with household ATP. Regionally, the independent variables and coverage profiles were the same as at the national level, although there were differences between regions:
— The probabilities of being covered in Antioquia were over 90% for almost all the cases. The internal differences were not significant.
— By contrast, in Pacifico the differences in ATP were much more significant and affected the probability of being covered. For example, the probability of being covered with a female head of household grew from 55% to 83% as ATP increased, whereas in Antioquia this same probability decreased from 88% to 80%.
— In the Pacifico region the differences of coverage related to ATP were much more significant, and the standard deviation in all cases was much higher than in Antioquia and nationally.
The most important differences are seen in education level and in economic position of household head. The probability of being covered increased with increasing household income and with each additional educational level achieved by the head of the household. In terms of economic position, individuals with household heads who were public employees had a very high probability of being covered regardless of their income, whereas freelancers, employers, and domestic employees had 50% less probability of being covered. As was the case for ATP, household income levels had less effect on the probability of being covered in Antioquia than in Pacifico. For example, the probability of being covered with a household head with no education in Pacifico grew from 55% to 90% as household income increased, whereas in Antioquia it went from 80% to 94%.
The household quality of life survey used in this analysis measures health services quality through two variables: perception of timeliness of response when care is requested and the degree of satisfaction with the received services. Over 78% classified the service as good, although there was less satisfaction among those covered by the Institute of Social Security in the contributive regime. More than 78% answered that the service was timely although those covered by the Institute of Social Security in the contributive regime reported more service delay. The regional analysis presented similar tendencies in terms of quality and timeliness although the Pacifico region had lower service quality and longer delays in relation to Antioquia.
The survey also evaluated reasons given by the covered population for not using health services. The principal reason overall was "lack of money" followed by "the illness was not serious." In Pacifico, the principal reason given was "lack of money," and in Antioquia it was that "the illness was not serious."
The main positive results of the social security reform have been the increased resources, both public and private, devoted to health; increased coverage; improved equity in financing and insurance coverage; and people's satisfaction with the system.
Public expenditure on basic health activities increased from 0.53% of GNP in 1980 to 1.41% in 1996. Private health expenditure also rose after 1990 because of the prepaid medicine system, going from 2–2.5% of GDP before the 1990s to 3.7% in 1995 (Vargas and Sarmiento 1997). Meanwhile, per capita expenditure more than doubled in a period of eight years (1987–1995) going from 24.7 USD to 58.7 USD, while expenditures per person among the poor tripled, going from 58.5 USD to 192.4 USD.
The main impact of the reform was on insurance coverage. Before the mid-1990s, the rate of affiliation was around 23%; it reached 57% by 1997 but then decreased to 53% in 2003, as mentioned earlier. By comparison, in the 1980s the rate of coverage had increased very slowly, from 10.3% in 1980 to 11.9% in 1986. Health sector reform not only increased global coverage but made it more equitable across income levels, age, and education. A very important portion of the observed increase in coverage comes from the coverage extended to the worker's family and to the poorest sectors of the population.
In 1993, subsidies for people affiliated to the social security system constituted 1.21% of GNP, with lower-income families receiving 0.04% of GNP, quintiles 3–4 receiving 0.7% (60% of the subsidy), and the higher-income group receiving 0.27% of GNP (22.3% of the subsidy).
By 1997, subsidized regime subsidies had reached 3.2% of GNP, with 66% going to the poorest 40%, while those with higher income received only 2%. In the contributive regime, subsidies reached 0.63% of GNP, with 0.9% going to quintiles 2–4. The participation of the richest group was negative, i.e., they had to pay instead of being subsidized. Adding both systems, 62% of the subsidies went to the 40% poorest population, and the highest quintile had a negative subsidy of –5.3%.
As can be seen in Figure 2, health system reform reduced the subsidies of the richest segment of the population. Between 1993 and 1997, subsidies for the fourth quintile went from 0.61% to 0.51% of GNP, and for the fifth quintile, from 0.38% to –0.20%.
FIGURE 2. HEALTH SECTOR NET SUBSIDIES BY INCOME QUINTILES AND AFFILIATION REGIMES (AS % OF GDP), COLOMBIA, 1993 AND 1997
Source: CASEN Survey (2000). Quality of Life Surveys (1993, 1997)
Institutional changes entailed not only a significant increase in net health expenditure, from 2.57% to 3.82% of GNP, but also a significant effort of focalization, as the total increase of the expenditure was destined to the 40% poorest population. As Figure 3 shows, by 1997, health system subsidies represented 45% of households' total income for the first quintile, compared with 12% in 1993. Health subsidies implied a better income distribution, reflected in the reduction of the Gini coefficient from 0.535 to 0.52 in 1993, and from 0.565 to 0.523 in 1997, when the Gini was recalculated after subsidies.
In the following figures, the expansion of coverage for the family is evident. Figures 4 and 5 show coverage expansion by age groups and education level. Figure 6 shows the evolution of social security coverage by income deciles; it can be seen that social security coverage after 1994 is more equitable. In Figure 7, we can see that health system reform reduced the strong relation between
FIGURE 3. INCOME INCREMENT CAUSED BY HEALTH SECTOR SUBSIDIES, BY QUINTILES AND REGIMES, COLOMBIA, 1993 AND 1997.
Source: CASEN Survey (2000). Quality of Life Surveys (1993, 1997).
income and coverage. In 1980, coverage of the 10th income decile was 31 times higher than in the first decile; by 1986 it was just 13 times higher, and by 1997 that factor had been reduced to only 1.86.
The levels of expenditure and coverage increased dramatically, but even more important was the effect of health system
FIGURE 4. SOCIAL SECURITY COVERAGE BY AGE GROUPS, COLOMBIA, 1980, 1986, 1997, 2000
Sources: National Health Surveys 1980, 1986; Quality of Life Surveys 1993, 1997; Demography and Health Survey 2000.
FIGURE 5. SOCIAL SECURITY COVERAGE BY EDUCATION LEVEL, COLOMBIA, 1980, 1986, 1997, 2000
Sources: National Health Surveys 1980, 1986; Quality of Life Surveys 1993, 1997; Demography and Health Survey 2000.
FIGURE 6. SOCIAL SECURITY COVERAGE BY INCOME DECILES, COLOMBIA, 1980, 1986, 1997
Sources: National Health Surveys 1980, 1986; Quality of Life Surveys 1993, 1997.
FIGURE 7. LAG FROM EACH INCOME DECILE WITH RESPECT TO THE HIGHEST INCOME DECILE, COLOMBIA, 1980, 1986, 1997
Sources: National Health Surveys 1980, 1986; Quality of Life Surveys 1993, 1997.
reform on the gap between social sectors and geographical zones. Figure 8 shows the gap between rural and urban coverage. It is evident that although urban zone coverage in the 1980s was almost four times the rural coverage, by 1997 it had decreased to only 1.25 times.
FIGURE 8. SOCIAL SECURITY COVERAGE BY ZONE AND PUBLIC HEALTH EXPENDITURE (PRICES OF 1995), COLOMBIA
Sources: health expenditure – DNP, Misión social; health coverage – National Health Surveys 1980, 1986; Quality of Life Surveys 1993, 1997; Demography and Health Survey 2000.
In several national studies, population satisfaction with the health-care system has been rated high. Over 78% classified the treatment as good, and more than 78% also answered that the service was timely.5 Similar findings have been reported elsewhere.
Despite the many positive outcomes of the reforms, there were also negative aspects whose repercussions are still being felt. These include the detrimental consequences of inadequate transition measures and the high costs of monitoring and controlling the system.
Public health programs and activities were adversely affected, first by decentralization and then by social security reform. Decentralization transferred public health responsibilities to the municipalities without the required training, technical assistance, and supervision by the Ministry of Health or the Department Health Secretaries. Immunization levels decreased, and the epidemiologic surveillance system went into disarray, leading to a series of outbreaks of communicable diseases (dengue, measles, and yellow fever).
With social security reform, the Ministry of Health became immersed in organizational changes and the new insurance system, relegating public health responsibilities to second place. Although the reform contemplated delegating responsibilities for health promotion and disease prevention to the territorial levels (national, departmental, municipal), as well as to insurers and providers, the Ministry of Health was very slow in regulating them. It was only in 1996, with the PAB, that the Ministry regulated the public health responsibilities of the territorial levels and, only in 2000, the health promotion and disease prevention activities of insurers and providers.
Although the reform law specifically contemplated promotion, prevention, and public health in the PAB, it took several years to regulate them (Resolutions 412 and 3384 of 2000). Even today, there is no adequate information as to how public health responsibilities are being carried out by the different municipalities.
Before the reform, public insurers and public hospitals operated within protected environments. Public insurers had been working under monopoly conditions (the respective workers did not have any choice), which made them extremely inefficient. They also had very strong unions that, through the years, had attained very considerable and expensive labour conditions. Public hospitals were also inefficient because of high fixed costs, due to excessive personnel, and likewise had expensive labour conditions imposed by labour unions. The situation was further aggravated by the fact that both public insurers and public hospitals had been under politicized managements.
Although the reform was intended to correct this state of affairs by bringing competition to both insurers and hospitals, thereby forcing them to become efficient, it unfortunately failed to address the conditions of public institutions comprehensively. The reasons for this were both political (incapability or lack of decisiveness in dealing with labour unions) and economic (lack of funds to pay laid-off personnel).
In the case of public insurers, the reform did not provide adequate transition conditions to protect them from the very unequal competition of new insurers, who had very low fixed cost structures and, in many instances, practiced cream skimming against the public insurers. This combination of inadequate transition planning, lack of political will to make the required changes on time, and unequal competition led to serious trouble for both public insurers and hospitals. Several of them now face bankruptcy.
The model adopted by the Colombian reform is complex and cumbersome, difficult to understand, and lacking in transparency. Dealing with this complexity and lack of transparency requires multiple checks and controls, resulting in system costs that could otherwise have been avoided. Among these are:
— The poor must prove their condition to get into the subsidized regime. The State incurs the expense of applying a means test, which must be repeated periodically because the condition of poverty may change.
— The State must verify that the level of income declared by those in the contributive regime corresponds to reality. Otherwise, people tend to under-declare in order to lower their contribution. It has been estimated that evasion, which includes both under-declaring and not becoming affiliated at all, is around 35% (Econometría S.A. et al. 2002).
— The insurers, paid by capitation, must document the names of their affiliates every month. The State has to check that persons reported by the insurer are really their affiliates and do not appear in the lists of other insurers or in the subsidized regime.
— Persons who become unemployed and do not have means to pay their contributive insurance must move to the subsidized insurance. However, this transit from one regime to the other is not automatic, being subject to the application of the means test and the availability of openings.
One of the main achievements of the Colombian health system reform has been an overall increase in health insurance coverage, resulting in improved equity, with significantly lower urban–rural and socioeconomic differentials and better targeting of subsidies to the poor. The coverage increase is even more significant if we take into account the deep economic recession the country has been through since 1995. However, there is still much room for improvement, as the regional analysis demonstrates, given the persistence of a positive association of coverage with income and education and the more important effect of these variables in the poorer region.
After nine years of implementation, the reforms had not attained their target of universal coverage, and in fact the level of coverage was stalled at the same position for nearly seven years. This raises many questions about the feasibility of achieving universal coverage, the strategies employed thus far, and the conditions necessary to make it work.
There are also many questions to be explored related to the funding and costs associated with the current system. Is it possible to reduce the transaction costs associated with controlling and monitoring the system? How can evasion and misreporting be eliminated or significantly reduced? Are there more effective models for the mix of private–public insurers and providers? What are the best strategies to ensure their survival and healthy competition? Can the overall system be made more efficient?
Another issue that needs attention is the inequality between the packages of services offered in the contributive and subsidized systems. What conditions would be required to align these more closely and achieve more equity? Is this really feasible, politically and economically, and if so, what strategies would be most appropriate?
Available information, although scarce, shows evidence of improved equity in access to health services. There is not yet any information on equity in quality. Additional research would be required to look at the effects of the reform on access and equity in quality.
1. Health centres and health posts are both ambulatory facilities. Health centres vary in size and complexity but always have at least an MD. Health posts do not have a permanent MD and are staffed by auxiliary personnel.
2. By 1997 the official coverage reported was 57%. However, this figure later decreased due to corrections of double counting and to decreased contributive insurance because of higher unemployment after 1999.
3. Ability to pay was estimated using as a parameter the household's availability of domestic appliances.
4. Gender, age, education level, main occupational activity, whether in the formal or informal employment sector.
5. From our own analysis of DANE 1997 Quality of Life Survey.
Econometría S.A.; Ministerio de Salud Pública; Banco Interamericano de Desarrollo. 2002. Evaluación y reestructuración de los procesos, estrategias y organismos públicos y privados encargados de afiliación, pago y recaudo de aportes al Sistema General de Seguridad Social de Salud en Colombia, SGSSS. Technical report presented to the Ministerio de Salud Pública.
Ministerio de Salud; Instituto Nacional de Salud; OPS (Pan-American Health Organization). 2002. Situación de salud en Colombia: indicadores básicos 2002.
Quevedo, E. 1990. Análisis socio histórico. In Yepes, F.J., ed., La salud en Colombia. Estudio Sectorial de Salud. Departamento Nacional de Planeación, Ministerio de Salud. Editorial Presencia, Bogotá, Colombia. 105 pp.
Vargas, J.E.; Sarmiento, A. 1997. La descentralización de los servicios de salud en Colombia. Economic Commission for Latin America and the Caribbean. Public Policy Reforms Series, no. 51, Santiago, Chile. 82 pp.
World Bank. 2003a. World development indicators database, August 2003 (Atlas method). http://devdata.worldbank.org/data-query/
_____ 2003b. World development indicators database, August 2003. http://devdata.worldbank.org/data-query/, cited June 2004.
Yepes, F.J. 1990. La salud en Colombia. Estudio Sectorial de Salud. Departamento Nacional de Planeación, Ministerio de Salud. Editorial Presencia, Bogotá, Colombia. 1124 pp.
Yepes, F.J.; Sánchez, L.H.; Cantor, B. 1999. La descentralización de la salud: el caso de tres municipios colombianos. Investigaciones en Salud Pública. Documentos Técnicos, no. 14, May 1999. OPS/OMS (Pan-American Health Organization/World Health Organization), Washington, DC, USA. 21 pp.
This page intentionally left blank
India, a country of continental proportions, holding almost one-sixth of the world's population in about one-fiftieth of the world's geographical area, is one of the poorest in the world, with a poor record of improving the living conditions of its population. India accounts for more than half the total population of the 52 countries of the world where elementary deprivation is concentrated. The country reports an extremely low level of life expectancy (LE) at birth, low literacy levels, and a high infant mortality rate (IMR). However, within this large country of 28 states1 and a number of union territories, there are important variations in living conditions. At one end of the spectrum lies the state of Kerala, comparable to the developed countries in LE, IMR, and total fertility rate. At the other end are the states of Bihar, Madhya Pradesh, Orissa and Uttar Pradesh, where living conditions are not very different from those prevailing at the lower end of the 52 deprived countries.
After its independence from British rule in 1947, India followed a strategy of planned economic development, with a strong import substitution orientation, for the next four decades. The country has had a precarious foreign exchange reserve position since the 1960s. Despite the rising oil prices and import bill beginning in the mid-1970s, India did not face a balance of payments (BOP) crisis, owing to the Gulf boom and large worker remittances. However, in the 1980s, the complacent Indian economy moved into an import-dependent growth strategy. The rising import bill, slack in the foreign exchange inflow, and heavy external commercial borrowing in the late 1980s developed into a BOP crisis following the Gulf War in 1991. India was compelled to approach the IMF for a loan, and thus began the Adjustment era.
Since then, the overall thrust of macroeconomic adjustment policies (MAPs) has been toward less government and more private enterprise, the reasoning behind which has been that in many sectors private enterprise could do as well, if not better, than government, ushering in rapid economic growth. It has, however, been recognized that less government may not be the best idea in social sectors characterized by market failures and large externalities. This was especially so in India in the beginning of the 1990s, where:
India's social indicators at the start of the reforms in 1991 lagged behind the levels achieved in southeast Asia 20 years earlier, … The gap in social sector development needed to be closed, not only to improve the welfare of the poor … but also to create the preconditions for rapid economic growth. While the logic of economic reforms required a withdrawal of the state from the areas…, it also required an expansion of public sector support for social sector development (Ahluwalia 2002, p. 85).
The need for increased public spending on health was even greater. While total spending on health care in India in the 1990s was over 5% of GDP, public spending, at less than 1% of GDP, was one of the lowest in the world.
In India's economy—allegedly committed to 'socialism'—the share of public expenditure in total health expenditure is only around 15%, compared with 75% in Western Europe's 'market economies', rising to 84% in Thatcherism-ravaged Britain. In fact, the share of public expenditure in total health expenditure is lower in India than in any other major region of the world (Dreze and Sen 2002, pp. 203–4).
In the Indian Constitution, health is a state responsibility. During Adjustment, many state governments in India had recourse to Health Systems Development Project loans from the World Bank for carrying out health sector reforms (HSR), of which one of the key policies has been to raise public spending on health care from the abysmally low levels seen up to then.
In this chapter, we analyze three issues:
i) whether the health sector budget has been protected in India during the Adjustment years;
ii) whether states implementing HSR have enhanced public spending on health; and
iii) in one state with a high burden of communicable diseases, how disease control programs are affected when HSR does not succeed in raising health budgets.
The first two issues are discussed in the general context of health spending on the part of the central and state governments, and the third is presented in the specific context of Orissa, one of the poorest states in India, with a very high burden of communicable diseases.
The first issue is analyzed using data on public spending across the states over the period between 1980/81 and 1997/98, and the second, using data from 1994/95 to 2002/03.2 Such a division of time periods is called for, as the states began implementing HSR from 1995/96 (e.g. Andhra Pradesh from 1995/96, Karnataka and Punjab from 1997/98, West Bengal from 1998/99, etc.). Therefore, analysis of the first issue was done up to 1997/98, as after that point the effects of HSR would begin to complicate the results.
Our approach has been to compare states pre- and post-Adjustment, with and without HSR. We have concentrated our analysis on 18 states representing approximately 96% of the country's population. The data have been taken from the supplement on state finances of the Reserve Bank of India Bulletin. The proportions needed for the analysis of both issues have been computed using expenditure data at current prices. For the first issue, expenditure data have been converted to constant price series using the implicit deflator of the state domestic product. For the second, two constant-price series have been computed, one using the Wholesale Price Index and another using the Drug and Medicine Price Index.
The chapter is organized in seven sections, beginning with this Introduction. The following section provides, as background, a brief account of India's vast diversity in terms of health status and health systems development within the states and in the patterns of financing health expenditure. The third section briefly presents the salient features of MAPs and Health Systems Development Projects implemented in some of the Indian states since 1994. Government health-care spending in the overall context of government spending up to the mid-1990s is analyzed in the fourth section. The fifth section takes up the analysis of health-care spending by the states from 1994/95 to 2002/03, with the objective of seeing whether HSR has led to enhanced budget allocation to the health sector in the reforming states. The sixth section analyzes the impact of HSR on disease control programs under fiscal pressure in Orissa. The final section presents some conclusions and outstanding issues.
The 20th century has seen a dramatic decline in the crude death rate in all developing countries. In India, the crude death rate (per 1 000 persons) declined from 22.4 in 1931 to 10.2 in 1991. Although in itself the decline is quite significant, given that India has a much larger proportion of the younger age population than other developed countries, a crude death rate of 10.2 nevertheless suggests a poor record in bringing down mortality.
The aggregate mortality rate of a country of India's continental size could conceal marked differentials across the states. As is seen in Table 1, the crude death rate varies from 6.2 to 11.0. Kerala stands apart, reporting a crude death rate of 6.2, which is even lower than that of many developed countries. One group of eight states—Andhra Pradesh, Gujarat, Haryana, Karnataka, Maharastra, Punjab, Tamil Nadu, and West Bengal—report crude death rates within a narrow range of 7.3 to 8.3. Another group of four states—Bihar, Madhya Pradesh, Orissa, and Uttar Pradesh—report crude death rates of above 10.0. A further striking feature across these groups of states is the differentials in crude death rates between male and female. In Kerala, the crude death rate for females is 2.3 points lower than that for males. The group of eight states shows slightly lower crude death rates for females than for males. In the group of four states, however, crude death rates for females are significantly higher than those for males. Thus, not only do the levels of crude death rates vary across the groups of states, but also the differentials between male and female rates show entirely different patterns.
The IMR presents a pattern similar to that of overall death rates across the states of India (Table 1). Compared to the all-India average of 71 for 1997, the variation ranges from 12 for Kerala to 96 for Orissa. Karnataka, Maharastra, Punjab, Tamil Nadu, and West Bengal report IMRs within a narrow range of 47 to 55. At the upper end, the states of Orissa, Madhya Pradesh, Rajasthan, and Uttar Pradesh report IMRs over 85.
The pattern is similar in the death rate of the ages 0–4 group. Kerala reports an extremely low rate of 3.8 deaths per 1 000 population. Karnataka, Maharastra, Punjab, and Tamil Nadu report rates between 12.6 and 16.6, and Orissa, Madhya Pradesh,
TABLE 1. DEATH RATES OF THE YOUNG IN THE INDIAN STATES, 1971–1997
State |
Crude death rate |
Infant mortality rate |
Death rate 0–4 years |
Death rate 5–9 years |
||||
1971 |
1997 |
1971 |
1997 |
1971 |
1996 |
1971 |
1996 |
|
Andhra Pradesh |
14.6 |
8.3 |
106 |
63 |
44.9 |
17.8 |
4.7 |
1.6 |
Assam |
17.8 |
9.9 |
139 |
76 |
45.7 |
24.0 |
7.2 |
3.5 |
Bihar |
13.9* |
10.0 |
118* |
71 |
42.5* |
27.9 |
5.3* |
3.9 |
Gujarat |
16.4 |
7.6 |
144 |
62 |
64.9 |
23.3 |
4.1 |
1.3 |
Haryana |
9.9 |
8.0 |
72 |
68 |
32.5 |
23.4 |
3.0 |
1.9 |
Karnataka |
12.1 |
7.6 |
95 |
53 |
37.5 |
16.6 |
4.3 |
0.9 |
Kerala |
9.0 |
6.2 |
58 |
12 |
24.5 |
3.8 |
2.3 |
0.9 |
Madhya Pradesh |
15.6 |
11.0 |
135 |
94 |
49.8 |
33.5 |
5.0 |
3.0 |
Maharastra |
12.3 |
7.3 |
105 |
47 |
44.6 |
13.1 |
3.2 |
1.0 |
Orissa |
15.4 |
10.9 |
127 |
96 |
54.2 |
30.6 |
5.4 |
2.9 |
Punjab |
10.4 |
7.4 |
102 |
51 |
38.9 |
15.2 |
1.6 |
1.2 |
Rajasthan |
16.8 |
8.9 |
123 |
85 |
55.9 |
31.4 |
5.3 |
2.0 |
Tamil Nadu |
14.4 |
8.0 |
113 |
53 |
40.7 |
12.6 |
4.6 |
0.6 |
Uttar Pradesh |
20.1 |
10.3 |
167 |
85 |
83.7 |
31.4 |
5.4 |
3.0 |
West Bengal |
11.0* |
7.7 |
91* |
55 |
33.5* |
18.1 |
2.6* |
1.6 |
All India |
14.9 |
8.9 |
129 |
71 |
51.9 |
23.9 |
4.7 |
2.3 |
Source: Registrar General of India (1999). | ||||||||
Note 1: *1981 data; earliest data available within this source for these states. | ||||||||
Note 2: All-India total includes all states and union territories. | ||||||||
Rajasthan, and Uttar Pradesh report rates above 30.5. Thus, with regard to overall mortality, infant mortality, and child mortality, Kerala is far ahead of the rest of the states of India and is comparable to many developed nations, but four states report very high mortality rates that are closer to those in many poorer developing countries. The rest of the states are at various points along the spectrum in between.
The striking differences observed across the states with regard to infant and child mortality rates are not seen in mortality rates for ages above four years. For the ages 5–9 group, the death rate for Kerala is no lower than for some other states. Karnataka, Kerala, Maharastra, and Tamil Nadu report death rates below 1.0. Assam, Bihar, Madhya Pradesh, Orissa, and Uttar Pradesh report death rates above 3.0. The rest of the states report rates in between. The reduction further narrows when groups above age 55 are considered. Only Haryana and Punjab report death rates significantly lower than the all-India average, and Kerala reports death rates closer to the all-India average in those age groups (for detailed discussion, see Narayana 2001).
On the whole, the striking differences observed with regard to crude death rates across the Indian states are largely owing to differences in infant and child mortality rates rather than to differences in mortality rates at higher age groups. Thus, the unique achievement of Kerala and better performance of some other states can be attributed to having lowered the death rates of the younger (ages 0–4) population.
Turning now to the Indian health-care environment, we observe wide variations in the density of medical institutions, their composition, and ownership patterns. The most striking example of inequality in the distribution of medical institutions is the concentration of 45% of all institutions in only three states—Gujarat, Kerala, and Maharastra—accounting for just 18% of the Indian population (Table 2). Five states—Bihar, Madhya Pradesh, Rajasthan, Uttar Pradesh, and West Bengal—account for almost 50% of the Indian population but have only 25% of the total medical institutions. This inequality in distribution of medical institutions is reflected in the variation in population served per institution: over 25 000 per institution in Bihar, Madhya Pradesh, and West Bengal; fewer than 10 000 in Goa, Gujarat, Kerala, Maharastra, and Punjab; between 10 000 and 25 000 in the rest of the states. Behind the fairly low population per institution lies another type of inequality, that between rural and urban areas. In Goa, Gujarat, and Maharastra, over 50% of the institutions are located in urban areas, whereas such an imbalance is not to be seen in Kerala and Punjab.
The composition of medical institutions varies enormously across the states. In Andhra Pradesh, Assam, Gujarat, Kerala, Madhya Pradesh, and Maharastra, over 20% of the institutions are hospitals. In these states, where institutions are concentrated
TABLE 2. DISTRIBUTION OF MEDICAL INSTITUTIONS IN INDIA BY STATES, 1993
State |
TMI |
Population per institution |
Share of urban in TMI |
Share in TMI (%) |
Share of private in TMI (%) |
|
Hospitals |
PHC/CHC |
|||||
Andhra Pradesh |
3 495 |
19 029 |
34.65 |
53.20 |
38.03 |
55.05 |
Assam |
1 261 |
17 764 |
11.45 |
21.25 |
52.93 |
6.98 |
Bihar |
3 112 |
27 763 |
7.84 |
10.54 |
75.74 |
2.89 |
Goa |
769 |
1 560 |
50.72 |
14.82 |
3.38 |
90.38 |
Gujarat |
10 724 |
3 851 |
65.36 |
22.10 |
10.31 |
82.57 |
Haryana |
748 |
22 059 |
33.02 |
10.43 |
60.56 |
8.16 |
Himachal Pradesh |
500 |
10 400 |
11.80 |
11.40 |
49.80 |
1.80 |
Jammu & Kashmir |
1 009 |
7 631 |
2.87 |
6.64 |
32.90 |
-- |
Karnataka |
2 628 |
17 123 |
19.10 |
11.15 |
57.27 |
2.55 |
Kerala |
4 953 |
5 875 |
22.39 |
41.20 |
19.42 |
76.64 |
Madhya Pradesh |
1 992 |
33 233 |
12.25 |
30.45 |
68.93 |
-- |
Maharastra |
13 237 |
5 975 |
78.84 |
23.53 |
14.95 |
74.75 |
Orissa |
1 664 |
19 050 |
14.30 |
17.07 |
68.99 |
2.70 |
Punjab |
2 255 |
9 002 |
17.16 |
9.62 |
25.54 |
2.26 |
Rajasthan |
2 145 |
20 513 |
22.00 |
10.16 |
76.64 |
-- |
Tamil Nadu |
2 428 |
23 023 |
28.17 |
16.80 |
62.11 |
6.10 |
Uttar Pradesh |
6 470 |
21 499 |
16.75 |
11.36 |
61.59 |
3.52 |
West Bengal |
2 576 |
26 436 |
16.38 |
15.22 |
63.39 |
7.41 |
All India |
64 393 |
13 143 |
39.92 |
21.26 |
36.18 |
40.97 |
Source: Government of India, Ministry of Health and Family Welfare (1993). | ||||||
Note 1: TMI = Total number of medical institutions; includes hospitals, dispensaries and PHC/CHCs. | ||||||
Note 2: All-India total includes all states and union territories | ||||||
in urban areas and hospitals account for a larger share, primary health centres (PHCs) and community health centres (CHCs) make up a small proportion of the total institutions. In the rest of the states, over 60% of the institutions are PHCs and CHCs.
The distribution of hospital beds is as unequal as that of medical institutions. The three states of Gujarat, Kerala, and Maharastra account for 37% of the total beds, whereas the five states of Bihar, Madhya Pradesh, Rajasthan, Uttar Pradesh, and West Bengal account for 28% (Table 3). The inequality in the population per bed follows from the inequality of bed distribution across the states. However, there is one important difference in the distribution of beds, in that over 80% of the beds in the
TABLE 3. DISTRIBUTION OF HOSPITAL BEDS IN INDIA BY STATES, 1993
State |
Total beds |
Share of urban (%) |
Population per bed |
Share of private (%) |
Andhra Pradesh |
26 972 |
64.67 |
2 466 |
90.53 |
Assam |
12 703 |
68.63 |
1 763 |
16.01 |
Bihar |
29 186 |
89.66 |
2 960 |
29.19 |
Goa |
3 644 |
63.09 |
329 |
48.38 |
Gujarat |
69 359 |
86.81 |
595 |
61.52 |
Haryana |
7 430 |
92.45 |
2 221 |
30.04 |
Himachal Pradesh |
4 035 |
83.77 |
1 289 |
4.81 |
Jammu & Kashmir |
8 202 |
1.71 |
939 |
1.71 |
Karnataka |
38 838 |
90.78 |
1 159 |
25.76 |
Kerala |
77 362 |
42.87 |
376 |
63.56 |
Madhya Pradesh |
18 143 |
65.93 |
3 649 |
-- |
Maharastra |
80 542 |
87.01 |
980 |
47.85 |
Orissa |
14 629 |
76.06 |
2 167 |
9.37 |
Punjab |
20 142 |
64.36 |
1 008 |
18.84 |
Rajasthan |
20 605 |
94.90 |
2 135 |
-- |
Tamil Nadu |
49 058 |
91.09 |
1 139 |
21.33 |
Uttar Pradesh |
53 007 |
85.43 |
2 624 |
22.83 |
West Bengal |
54 767 |
86.33 |
1 243 |
12.62 |
All India |
621 376 |
78.26 |
1 362 |
35.66 |
Source: Government of India, Ministry of Health and Family Welfare (1993). | ||||
Note: All-India total includes all states and union territories. | ||||
states are located in urban areas. Only Kerala has about 60% of its hospital beds in rural areas.
States reporting a higher density of hospital beds, especially states with one bed for every 1 000 population or less, have a high proportion of beds in the private health-care sector—close to 50% in Goa and Maharastra, and over 60% in Gujarat and Kerala. Only Andhra Pradesh has a higher private share, combined with a lower density of beds.
The states present striking differences in population covered per institution. Low-density states owe their achievements to the efforts of the public sector in expanding PHCs and CHCs, providing a rural spread. With regard to the density of beds, the story is similar. The share of the private sector is low in the low-density states. However, unlike the medical institutions in the public sector, which are largely in the rural areas, the private-sector beds are largely in the urban areas.
The major Indian states have been categorized by the World Bank according to variations in mortality, fertility, immunization, and the capacity of the public health sector (Table 4). The differences in health transition status mean that while a state like Kerala, at one end, faces the burden of dealing with the high per-episode cost of health care, states like Madhya Pradesh and Orissa, at the other end, face the more urgent need to focus on pre-transition (i.e., predominantly communicable) diseases by implementing public control programs (World Bank 2001). Hence, states in each category face different challenges.
TABLE 4. CATEGORIZATION OF MAJOR INDIAN STATES BY CHARACTERISTICS INFLUENCING FUNDAMENTAL HEALTH SYSTEM CHOICES
Characteristics |
States |
% of India's population |
||||||
A. Middle to late transition, moderate to high capacity |
Kerala Tamil Nadu |
9.1 |
||||||
B. Early to middle transition, low to moderate capacity |
Andhra Pradesh Gujarat Haryana Karnataka Maharastra Punjab West Bengal |
39.1 |
||||||
C. Very early transition, very low to low capacity |
Madhya Pradesh Orissa Rajasthan Uttar Pradesh |
33.1 |
||||||
D. Special cases of instability, high to very high mortality, civil conflict, very poor governance |
Assam Bihar |
13.3 |
||||||
Source: World Bank (2001), Report no. 22304, Table 3.2. | ||||||||
Note: Major states (>15 million population) were ranked according to Infant Mortality, Child Mortality, Total Fertility, and Full Immunization Coverage rates. The estimates included the major states before bifurcation, so the state of Bihar includes Jharkhand, and Madhya Pradesh includes Chattisgarh, but Uttar Pradesh does not include Uttaranchal. | ||||||||
Provision of health care in India is a responsibility shared by state, central, and local governments. Health care delivery is actually a responsibility of state governments. The central government is responsible for health care in Union Territories without legislature. It is also responsible for developing and monitoring national standards and regulations; providing the link between state governments and international and bilateral agencies; and sponsoring numerous programs, through the provision of financial and other inputs, for implementation throughout the states. Within the overall ambit of national policies, there is scope for each state to formulate its own health policy and administer health programs in conformity with local conditions.
The contribution of the central and state governments to overall public spending on health may be seen in Table 5. The states mainly finance primary health-care facilities, hospitals, and insurance. Of the two departments at the centre, the Department of Health allocates over 45% of its budget to the central teaching hospitals and research institutions, about 15% to the Central Government Health Scheme benefiting its own employees, and about 35% to disease control programs. The Department
TABLE 5. CENTRE AND STATE SHARES IN DIFFERENT COMPONENTS OF THE GOVERNMENT HEALTH BUDGET IN INDIA, 1991/92
|
Centre (%) |
States (%) |
Hospitals |
3.1 |
96.9 |
Public health |
0.0 |
100.0 |
Primary care (disease control) |
99.7 |
0.3 |
Family welfare |
22.6 |
77.4 |
Insurance (CGHS, ESIS) |
18.2 |
81.8 |
Medical education & Other |
41.7 |
58.3 |
Administration & Other |
11.0 |
89.0 |
Capital investment |
49.7 |
50.3 |
Source: World Bank (1995), Report no. 13402-IN. | ||
Note: CGHS = Central government health scheme; ESIS = Employees' state insurance scheme. | ||
of Family Welfare allocates about 85% of its budget to family planning and 15% to maternal and child health and universal immunization. Local governments have no significant financial authority in India except in large cities.
The first and most important MAP targets are the amount and the distribution of government expenditure. Setting these targets consitutes a key policy objective as well as a policy tool in shaping other sectors of the economy. In the Indian context, in the initial years of Adjustment, one way in which reduced public expenditure affected state spending on health care was a decrease in direct spending on health care by the central government. Another was in loans advanced by the central government to finance health expenditure in the states, because in the Indian context the primary lender to the states is the central government.
Since 1992, the Government of India and the World Bank have been engaged in a dialogue on health sector development. The focus of that dialogue has been on addressing the most burdensome diseases cost-effectively and moving toward efficient and effective health systems. Development projects in seven states were begun at different times and were of different durations. Andhra Pradesh was the first to begin in 1994 and completed its project in 2002. Karnataka and Punjab began their projects a year later and were completed in 2004. West Bengal began in 1997 and finished in 2004. The projects in Maharastra and Orissa started in 1998; the former was concluded in 2005 and the latter in March 2006. Uttar Pradesh (including Uttaranchal) began in 2000 and will continue until December 2006. As mentioned, health is a state responsibility, and MAPs have not led to HSR in all the Indian states. We therefore differentiate states, in this discussion, as being with HSR and without HSR.
There is some uniformity in the development objectives of HSR across the seven states. The Health Systems Development Project seeks to develop strategic management capacity; strengthen performance, accountability, and efficiency; and build implementation capacity. Further, it seeks to improve clinical service quality by renovating and expanding district, subdistrict, and community hospitals and improving access to services. In all seven reforming states, around 15% of the total project cost is borne by the state governments.
All the project documents note the low levels of funding for secondary hospitals in the reforming states. This is attributed to the small share of overall public spending allotted to health, the limited portion of total health spending going to hospitals, and, within this, a skewed distribution of funds in favour of the tertiary hospitals. After analysis of the problems of the health sector, the governments of the reforming states have agreed—using terminology ranging from "assurances" to "commitments"—to several undertakings. These are: (i) to enhance the overall size of the health budget; (ii) to redress imbalances in public expenditure between secondary and tertiary care levels; (iii) to safeguard the operations and maintenance components of current expenditure allocations for the secondary health-care sector; (iv) to charge user fees for selected services; and (v) to address workforce issues.
Even though the reform is focused on community, area, and district hospitals—or first-referral hospitals—in all the reforming states, there is variation both in the number of hospitals designated for renovation and expansion and in the specific geographical areas and population groups targeted. In Andhra Pradesh, 21 district hospitals, 49 area hospitals, and 80 community hospitals are slated to be renovated and expanded, with an expected addition of around 5 000 beds. In Punjab, 13 district hospitals, 46 subdivisional hospitals, and 74 community hospitals will receive financing, with a relatively larger share of the investment targeted in the upper Bari Doab and rural Southern Malwa regions, where poverty is concentrated. The expected addition of beds in the secondary sector is only around 2 000. Although 15 district hospitals, 60 subdivisional hospitals, and 95 rural hospitals are to be renovated in West Bengal, no beds are to be added. However, 28 primary health centres and 8 subdistrict primary health centres will be upgraded in the Sunderban area, where the poorest in the state are concentrated. As Sunderban has a large number of widely distributed small islands, three floating medical units will also be set up. In Uttar Pradesh, 152 hospitals of various sizes are targeted for renovation, largely in the poorest three regions of the state: Bundel Khand, Uttar Khand, and eastern Uttar Pradesh. In none of the states is construction of new hospitals being financed; the focus is entirely on renovation, expansion, and addition of clinical specialties in existing rural hospitals.
There is considerable variation in the size and geographical focus of the various projects and in the population groups targeted. Nevertheless, the Health Systems Development Projects being implemented in the seven states are quite similar in their development objectives and key policy reforms. It is this similarity in project objectives, approaches, and components that allows us to consider the HSRs in all these states together as one HSR.
As in all other adjustment programs, the Indian program began with central government bringing down the gross fiscal deficit (GFD) as a percentage of GDP. A decrease of about two and a half percentage points was achieved by cutting total expenditure by over two percentage points in the initial burst in 1991/92. Since then, both the GFD and the total expenditure have come down slowly, reaching 4% and 14.14%, respectively, by 1996/97 (Table 6). The total expenditure of the central government consists of revenue expenditure and capital disbursements; cuts have fallen more heavily on the latter and the decline continues. The reduction in capital disbursements is more acute when viewed in the context of a persistent decline since the mid-1980s. Capital disbursements fall under two headings: capital expenditure by the central government, and loans and advances to state governments and union territories. The two had equal shares in total capital disbursements throughout the late 1980s. Then the year 1990/91 saw a sharp drop in capital expenditure followed by a steady decline, reaching 1.12% of GDP in 1996/97, with only marginal increase since then. Loans and advances had shown a slower decline, reaching 1.72% of GDP in 1996/97, then a pause in the following two years. Thus, the reduction in the central government's total expenditure has translated into cuts in loans advanced to state governments and in capital expenditure by the central government, with capital expenditure bearing it more heavily in the initial years.
TABLE 6. EXPENDITURE (% GDP) OF THE CENTRAL GOVERNMENT, INDIA, 1986/87–1998/99
|
Government expenditure |
Capital expenditure |
Revenue expenditure |
|||
Year |
Total |
DE |
Total |
SS |
Total |
SS |
1986/87 |
20.30 |
17.15 |
3.16 |
0.13 |
13.96 |
0.95 |
1987/88 |
19.23 |
15.86 |
2.79 |
0.08 |
13.86 |
1.02 |
1988/89 |
18.83 |
15.14 |
2.59 |
0.09 |
13.67 |
0.96 |
1989/90 |
19.25 |
15.35 |
2.58 |
0.07 |
14.06 |
0.87 |
1990/91 |
18.60 |
14.59 |
1.77 |
0.05 |
14.23 |
0.78 |
1991/92 |
16.62 |
12.31 |
1.36 |
0.04 |
13.36 |
0.83 |
1992/93 |
16.26 |
11.84 |
1.66 |
0.04 |
13.19 |
0.80 |
1993/94 |
16.94 |
12.36 |
1.65 |
0.04 |
13.51 |
0.89 |
1994/95 |
16.66 |
12.05 |
1.41 |
0.08 |
13.28 |
0.87 |
1995/96 |
15.97 |
11.41 |
1.42 |
0.05 |
12.74 |
0.96 |
1996/97 |
14.14 |
9.51 |
1.12 |
0.05 |
11.30 |
0.89 |
1997/98 |
14.86 |
10.21 |
1.30 |
0.04 |
11.58 |
0.92 |
1998/99 |
15.96 |
11.58 |
1.41 |
0.05 |
12.22 |
1.15 |
Source: Government of India, budget documents, various years. | ||||||
Note: DE = Discretionary expenditure; SS = Social services. | ||||||
The total capital expenditure of the central government may be viewed as a sum of capital expenditures on defense, social service sectors, and sectors providing economic services (Table 6). Capital expenditure on defense declined somewhat, mostly between 1993/94 and 1998/99. In contrast, after a declining trend seen earlier (1986/87 to 1991/92) in the social sectors, this area stagnated at around .05% of GDP during the Adjustment years. The largest cut was imposed on economic services, beginning in the late 1980s (2.45% of GDP in 1986/87) and continuing with greater intensity during the Adjustment years (0.73% of GDP in 1998/99). Thus, the cut in capital expenditure has largely been borne by economic services, of which the major headings include agriculture, industry, transport and communications, and irrigation and power.
Two headings of expenditure that had not experienced any systematic cuts before 1990/91 and were subjected to cuts during Adjustment are subsidies and grants-in-aid to states. Subsidies, which had increased to 2.29% of GDP by 1989/90, declined steadily to 1.16% of GDP by 1996/97. Grants-in-aid to states continued to increase until 1993/94, then fell by 1.23% of GDP by 1998/99. Social services show a reversal of the trend; expenditure increased mildly after 1993/94, with a significant increase in 1998/99. However, during that period the share allotted to health and family welfare decreased from 29.7% to 21.85%. Expenditure in economic services has been cut, going from about 4.4% of GDP in 1990/91 to 2.81% in 1998/99. Thus, overall, the sharp decrease in discretionary spending by the central government has fallen on subsidies, grants to state governments, and spending on economic services, whereas social sector spending by the central government has been protected during Adjustment.
Like all other adjustment programs, the Indian program has not been able to cut total expenditure to any significant extent since Adjustment began in 1991. As experienced the world over, the burden has fallen largely on economic services and not on defense or public services. In that respect, the Indian adjustment experience is no different from that elsewhere. However, the social sector has been protected. The distinctly Indian attribute of the constraint on spending is its translation into a contraction of capital disbursements and grants-in-aid to the states.
The states in India enjoy considerable autonomy and are not bound by the central government's undue concern for cutting the fiscal deficit under Adjustment. However, the central government is the source of financing for the states' deficits, and close to 50% of the states' debt is to the central government. The early 1980s witnessed unlimited overdrafts on the Reserve Bank of India (the central bank) by the states. Various control measures were imposed beginning in the mid-1980s, such as not allowing overdrafts for more than seven continuous working days. Since then, the gross fiscal deficit of all states has remained at around 3%.
The total expenditure of the state governments as a percentage of net state domestic product (NSDP) had shown significant increase throughout the 1980s. At the level of all-states, total expenditure increased from 20% to 22% between 1981/82 and 1987/88. Most of the individual states showed similar increases, with the exception of Bihar, Orissa, Tamil Nadu, and West Bengal, where there was no increase or only a marginal increase. Total expenditure dipped during the next four years in most of the states, except Bihar, Orissa, Tamil Nadu, and Uttar Pradesh. Of these exceptions, all but Uttar Pradesh had also shown no increase in total expenditure during most of the 1980s. The increase in total expenditure came in these states only between 1987/88 and 1991/92. The 1990s saw stagnation in total expenditure for almost all the states, except Bihar, Goa, Haryana, Himachal Pradesh, and Kerala.
A freeze on total expenditure often leads to pressure on capital expenditure, as various components of revenue expenditure, such as salaries and interest payments, are not amenable to contraction. The Indian states have had a similar experience since the late 1980s. Revenue expenditure as a proportion of NSDP had shown a significant increase of over three percentage points between 1981/82 and 1987/88 at the level of all-states. The exceptions were Bihar, Orissa, Tamil Nadu, and West Bengal, where the increase was of a smaller order and where there was no increase in total expenditure during this period. However, the period from 1987/88 to 1991/92 showed a reversal, with revenue expenditure coming down. Again Bihar, Orissa, Punjab, Tamil Nadu, Uttar Pradesh, and West Bengal were the exceptions. The 1990s saw increases in revenue expenditure in all the states except Gujarat, Maharashtra, and Tamil Nadu, where total expenditure stagnated.
As seen above, the 1980s were a period of significant increases in total expenditure. However, revenue expenditure was increasing in most of the states, and capital expenditure was decreasing in all the states except Jammu and Kashmir well into the 1990s. Between 1981/82 and 1997/98, there was a decline in capital expenditure of 2.5 percentage points at the level of all-states.
One component of total expenditure that drives revenue expenditure during Adjustment is higher interest payment on past debt. Already by the early 1980s, interest payments were over 1% of NSDP for all the states except Maharastra and rose steadily throughout the 1980s and 1990s. By the end of 1980s, they had reached close to 2% of NSDP and, by the end of 1990s, more than 3%. Meanwhile, discretionary expenditure at the level of all-states has shown a mild decline since 1987/88, except in Bihar, Goa, Himachal Pradesh, and Kerala. When total expenditure is frozen and interest payment increases, fewer resources are available for discretionary expenditure by the states.
The share of medical and public health (MPH), including water supply and sanitation (WSS), in total government expenditure increased in the 1980s. Although the total increase was only on the order of half a percentage point, almost all states showed an increase between 1981/82 and 1987/88. The next four years reversed this trend, and both MPH and WSS lost the gains of the 1980s. The exceptions were Bihar, Goa, and Himachal Pradesh, which showed gains in both these components during this period. The decline continued through the 1990s in
FIGURE 1A. TOTAL (REVENUE AND CAPITAL) EXPENDITURE ON MEDICAL AND PUBLIC HEALTH, WITH WATER SUPPLY AND SANITATION, AS % OF TOTAL EXPENDITURE, INDIA, 1991/92
Source: Reserve Bank of India, RBI Bulletin, various issues.
almost all the states, except Gujarat, Karnataka, Rajasthan, and Tamil Nadu (Figures 1a, 1b).
It is possible that while the share of a sector within total expenditure does not show much of an increase, the per capita spending does. Turning to per capita expenditure (in constant 1980/81 prices) on MPH, it may be seen that the 1980s were a period of significant increase in per capita expenditure in MPH, including WSS, in all the Indian states (Figures 2a, 2b). At the all-state level, the increase was about 40% in six years, with increases of comparable magnitude recorded by most of the major states. The next four years (1987/88 to 1991/92) showed no increase
FIGURE 1B. TOTAL (REVENUE AND CAPITAL) EXPENDITURE ON MEDICAL AND PUBLIC HEALTH, WITH WATER SUPPLY AND SANITATION, AS % OF TOTAL EXPENDITURE, INDIA, 1997/98
Source: Reserve Bank of India, RBI Bulletin, various issues.
in per capita expenditure at the all-state level, with most major states also showing no increase in expenditure. Tamil Nadu was an exception, in that it continued to increase. At the other end, Andhra Pradesh, Gujarat, Madhya Pradesh, and Maharastra experienced a decline in expenditure of about 15%. The initial years of Adjustment (1991/92 to 1995/96) showed a mild increase in per capita expenditure at the all-state level, although the pattern across states varied widely. All the southern states and Gujarat, Haryana, Himachal Pradesh, and Rajasthan showed increases, while Assam, Bihar, Punjab, and Uttar Pradesh showed decreases in per capita expenditure. Especially striking were the decreases
FIGURE 2A. PER CAPITA EXPENDITURE (IN INR) ON MEDICAL AND PUBLIC HEALTH, WITH WATER SUPPLY AND SANITATION, INDIA, 1991/92
Source: Reserve Bank of India, RBI Bulletin, various issues.
in Assam, Bihar, and Uttar Pradesh, where per capita expenditures are the lowest.
Per capita expenditure on MPH, with WSS, varies enormously across the states. In 1981/82, within an all-state average of 25.26 INR, expenditure varied from 13.82 INR in Bihar to 91.54 INR in Himachal Pradesh.3 At the lowest end, in the per capita expenditure class of below 25 INR, were the states of Andhra Pradesh, Assam, Bihar, Karnataka, Madhya Pradesh, Orissa, and Uttar Pradesh (Table 7). At the highest end, Himachal Pradesh and Jammu and Kashmir spent over 80 INR per capita. The 1980s saw a clear upward movement in almost all states. Only Bihar and Uttar Pradesh remained at the bottom class. Four more states were added to the top expenditure class of above 45 INR per capita. During the period 1987/88 to
FIGURE 2B. PER CAPITA EXPENDITURE (IN INR) ON MEDICAL AND PUBLIC HEALTH, WITH WATER SUPPLY AND SANITATION, INDIA, 1995/96
Source: Reserve Bank of India, RBI Bulletin, various issues.
1991/92, while Bihar remained at the bottom class of less than 25 INR and Andhra Pradesh dropped to that class, five states at the top class maintained their position.
Because the inclusion of water supply and sanitation creates a data series that is not amenable to the analysis of MPH, we also looked at MPH alone. No significant increase in per capita expenditure was seen between 1987/88 and 1991/92, and in most states the expenditure had stagnated. The exceptions were Bihar, Goa, Himachal Pradesh, Jammu and Kashmir, Rajasthan, and Tamil Nadu. This trend continued in most states throughout the Adjustment years, with fewer exceptions. Among the six exceptional states of the previous period, two—Bihar and Goa—dropped
TABLE 7. DISTRIBUTION OF INDIAN STATES BY PER CAPITA EXPENDITURE ON MEDICAL AND PUBLIC HEALTH (INCLUDING WATER SUPPLY AND SANITATION)
|
1981/82 |
1987/88 |
1991/92 |
1995/96 |
Below INR 25 |
Andhra Pradesh Assam Bihar Karnataka Madhya Pradesh Orissa Uttar Pradesh |
Bihar Uttar Pradesh |
Andhra Pradesh Bihar |
Bihar Uttar Pradesh |
INR 25–34 |
Gujarat Haryana Kerala Maharastra Punjab Tamil Nadu West Bengal |
Andhra Pradesh Assam Karnataka Madhya Pradesh Orissa West Bengal |
Assam Karnataka Madhya Pradesh Orissa Uttar Pradesh West Bengal |
Andhra Pradesh Assam Madhya Pradesh Orissa West Bengal |
INR 35–44 |
Rajasthan |
Gujarat Haryana Kerala Tamil Nadu |
Gujarat Haryana Kerala Maharastra Tamil Nadu |
Gujarat Kerala Karnataka Maharastra Punjab |
INR 45 and above |
Himachal Pradesh Jammu & Kashmir |
Goa Himachal Pradesh Jammu & Kashmir Maharastra Punjab Rajasthan |
Goa Himachal Pradesh Jammu Kashmir Punjab Rajasthan |
Goa Haryana Himachal Pradesh Jammu & Kashmir Rajasthan Tamil Nadu |
Source: Reserve Bank of India, RBI Bulletin, various issues. | ||||
Note: States in bold are those that made the most progress. | ||||
out, and Karnataka and Kerala came in. Overall, there was hardly any improvement in per capita expenditure on MPH after 1987/88 in the Indian states. Eleven of the 18 states analyzed were spending amounts in the mid-1990s that equaled the amount they had been spending in the late 1980s. Many of these poorer states were spending 30% less than the all-states average.
On the whole, Adjustment has led the central government to cut total expenditure to bring down the fiscal deficit. The cut has fallen more heavily on capital disbursements, i.e., capital expenditure and loans and advances to the states. Central government expenditure on the social sector has been left untouched. The total expenditure of the states, which increased in the 1980s, became stagnant in the 1990s following Adjustment. At the same time, interest payments rose, leaving less for discretionary spending.
The share of MPH, including WSS, within total spending of the states showed a mild increase in the 1980s, which was reversed in the late 1980s, with Adjustment further accentuating that downward trend. The increase in the share allotted to MPH in the 1980s translated into a significant increase in per capita expenditure, which then stagnated under Adjustment and the consequent contraction. The extremely low spenders, namely Andhra Pradesh, Assam, Madhya Pradesh, Orissa, Uttar Pradesh, and West Bengal, have remained at that level for a long time. It is therefore not surprising that many of them have availed themselves of the World Bank loan for health sector reforms.
HSRs undertaken by the states in India have sought to enhance health sector spending. However, projects funded by the World Bank in India have generally been characterized by 60% to 75% of total funding being earmarked for capital spending on upgrading hospital buildings and purchasing equipment. Hence, it is pertinent to analyze trends in capital spending during the HSR era in India.
Capital expenditure on MPH, excluding WSS, as a percentage of total capital expenditure has shown a declining trend in all the Indian states since the early 1980s, with few exceptions (Table 8). The exceptions were mostly smaller states with special status, such as Assam, Goa, Himachal Pradesh, and Jammu and Kashmir, with only Tamil Nadu and Uttar Pradesh being the real exceptions among the larger states. The picture in the mid-1990s, just before some of the states decided to go for the World Bank-funded Second Health Systems Development Project, stood as shown in the second column of Table 8. Nearly one-quarter of the states were spending less than 1% of total capital expenditure on the health-care sector, one-third were spending 1–2%, and another one-third, 3–5%. Only two states were spending more than 5% on the health-care sector.
The dominant trends in capital spending on health care during the next 10 years (HSR era) were either of the health sector share steadily falling or of it being maintained at an already low level. The distinction of maintaining capital spending at a relatively high level goes to only one state, Kerala. The states showing a sharp increase from extremely low levels of capital spending were all states that had gone for HSR, the exception being Himachal Pradesh, which increased its capital spending without World Bank funding. Andhra Pradesh and Punjab, despite implementing World Bank-funded HSR, do not show significant improvement in their status because in both states, resources are received by parastatals and their expenditures are not shown in government budgets.4 The increase in capital spending seen in Andhra Pradesh as of 1998/99 must also be interpreted with caution, as that state had managed to get another World Bank project by then, called the Andhra Pradesh Reconstruction Project, of which 24% of the loan amount—around Rupees 300 crores (approximately 66.7 million USD)—was earmarked for primary health centres. Discounting that amount, there has been no
TABLE 8. DISTRIBUTION OF INDIAN STATES BY CAPITAL EXPENDITURE ON MEDICAL AND PUBLIC HEALTH (AS % OF TOTAL EXPENDITURE)
|
1994/95 to 1996/97 |
1997/98 to 1999/00 |
2000/01 to 2002/03 |
Less than 1 percent |
Andhra Pradesh Bihar Gujarat Maharastra* Punjab West Bengal* |
Assam Bihar Maharastra* Punjab |
Gujarat Haryana Punjab |
1 to 3 percent |
Himachal Pradesh* Karnataka* Madhya Pradesh Orissa* Rajasthan |
Andhra Pradesh Gujarat Haryana Madhya Pradesh Orissa* Rajasthan |
Andhra Pradesh Assam Bihar Goa Jammu & Kashmir Madhya Pradesh Maharastra* Rajasthan Tamil Nadu Uttar Pradesh |
3 to 5 percent |
Assam Haryana Jammu & Kashmir Kerala Tamil Nadu |
Goa Himachal Pradesh* Jammu & Kashmir Kerala Uttar Pradesh West Bengal* |
Himachal Pradesh* Karnataka* Kerala Orissa* |
Above 5 percent |
Goa Uttar Pradesh |
Karnataka* Tamil Nadu |
West Bengal* |
Source: Reserve Bank of India, RBI Bulletin, various issues. | |||
Note 1: * States with HSR. | |||
Note 2: States in bold are those that made the most progress. | |||
capital spending from the regular budget by the Andhra Pradesh government on the health-care sector.
The steady decline in capital spending by a large number of states, combined with the boost in spending by the reforming states due to HSR, has resulted in the clustering of 10 of the 18 states into the 1–3% expenditure group at the end of the period. Of the five states in the above-3% expenditure category, three have risen to that level owing to the World Bank loan, and it is almost certain that once the project is implemented they will quickly fall into the below-1% expenditure group. This is already the case for Punjab and, if the 2002/03 data are any indication, for Karnataka as well.
We begin our analysis looking at the share of health-care sector revenue expenditure within discretionary expenditure (defined as total revenue expenditure minus interest payment). The non-reforming states, as a group, show a decline in the share of the health sector within total current spending after accounting for non-discretionary spending (Figure 3a). The share, which ranged from 6% to 10% in 1994/95, had fallen to between 4% and 8% by 2002/03. The two exceptions to this trend are Jammu and Kashmir and Haryana, the former still showing a share of over 8%, and the latter going from 2% in the mid-1990s to 5% by 2002/03. With the exception of West Bengal, the reforming states as a group, which were spending less than 8% of discretionary expenditure in the mid-1990s, have shown a mild decline in expenditure (Figure 3b). Nevertheless, their spending levels are still only around that of non-reforming states, reflecting their poorer record from the start. Thus, in a small way, HSR has helped moderate the fall in the share of health-care sector within total expenditure.
The above conclusion, however, requires an important amendment, because HSR has also introduced the policy of charging user fees, thereby raising the revenue receipts of the
FIGURE 3A. REVENUE EXPENDITURE ON MPH, AS % OF DISCRETIONARY EXPENDITURE: NON-REFORMING STATES, INDIA, 1994–2002
Source: World Bank (1995), Report no. 13402-IN.
Note: Remaining non-reforming states: Assam, Bihar, Goa, Gujarat, Himachal Pradesh, Kerala, Madhya Pradesh, Rajasthan, and Tamil Nadu.
government to some extent. Part of the collection—the proportion varies across the states—is retained at the level of the health facility and, to the extent that it is spent on equipment maintenance, drugs, and consumables, constitutes an addition to the health budget. The rest flows into the state treasury as a revenue receipt. The revenue expenditure on the health sector, minus the revenue receipts, is then the allocation from the total discretionary expenditure to the health sector. When this allocation is analyzed, the reforming states show a slightly sharper fall over the period. The share tends to lie between 5% and 7%, except for West Bengal, which reports closer to 8%. These proportions look no different from those reported for the non-reforming states, thus suggesting that HSR has not helped to maintain the share of health sector budgets at levels prevailing in the mid-1990s.
FIGURE 3B. REVENUE EXPENDITURE ON MPH, AS % OF DISCRETIONARY EXPENDITURE: REFORMING STATES, INDIA, 1994–2002
Source: World Bank (1995), Report no. 13402-IN.
Note: Remaining reforming states: Andra Pradesh, Karnataka, Maharastra, Orissa, Punjab, and Uttar Pradesh.
The share of the health-care budget within the government's total expenditure is one indicator of the priority accorded to health care by the state. Another is the per capita expenditure. To do a trend analysis on this aspect, we converted the expenditure at current prices to constant prices and built two series using the wholesale price index and the drug and medicine price index. The per capita expenditure in the mid-1990s ranged from 50 to 100 INR among the reforming states and 40 to100 INR among the non-reforming states (Figures 4a, 4b). After nine years, when some states had already completed the World Bank project and others were midway through it, the per capita expenditure of the reforming states had risen to between 75 and 120 INR, except for Punjab, where it had reached close to 180 INR, while the expenditure of the non-reforming states had risen to between
FIGURE 4A. PER CAPITA REVENUE EXPENDITURE ON MPH, 1993/94 PRICES, USING WHOLESALE PRICE INDEX (WPI): NON-REFORMING STATES, INDIA, 1994–2002
Source: World Bank (1995), Report no. 13402-IN.
Note: Remaining non-reforming states: Haryana, Gujarat, Kerala, Rajasthan, Tamil Nadu, and Assam.
60 and 140 INR, suggesting that the non-reforming states have done better than the reforming states. In both groups, there are two poorly-performing states pulling the group down—Orissa and Uttar Pradesh among the reforming states, and Bihar and Madhya Pradesh among the non-reforming states.
When deflated by the drug and medicine price index, the per capita expenditure shows a lower rate of increase than when deflated by the wholesale price index. The per capita expenditure of the reforming states showed a mild increase; the expenditure ranges between 45 and 80 INR toward the end of the period, with many states reporting closer to 80 INR. Among the reforming states, Punjab has shown the sharpest increase, spending about 120 INR as of 2002/03. The non-reforming states began within a range of expenditure of 60 to 90 INR in the mid-1990s and continued within that range at the end of the period, with two exceptions.
FIGURE 4B. PER CAPITA REVENUE EXPENDITURE ON MPH, 1993/94 PRICES, USING WHOLESALE PRICE INDEX (WPI): REFORMING STATES, INDIA, 1994–2002
Source: World Bank (1995), Report no. 13402-IN.
Note: Remaining reforming states: Andra Pradesh, Karnataka, Maharastra, and West Bengal.
These are Bihar and Madhya Pradesh, which were spending less than 60 INR in the mid-1990s and continue at the same level with hardly any increase. Among the reforming states as well there are two exceptions, Orissa and Uttar Pradesh, that continue to spend less than 60 INR per capita in constant prices.
Capital expenditure in the Indian health-care sector shows a general decline since the early 1980s. The World Bank loan for Health Sector Development boosted capital spending in states with HSR during the period of the project, and the general falling tendency was arrested. The share of MPH in total revenue spending in the states also shows a decline, but milder for the reforming states. Despite the availability of the World Bank loan for capital spending, the reforming states have not enhanced the share of the health-care sector in their total spending.
Unlike the declining tendency of the share of health sector capital and revenue spending, the per capita spending—whether deflated by wholesale price index or by the drug and medicine price index—shows a mild increasing tendency. However, it cannot be said that the performance of the reforming states is better than that of the non-reforming states.
Orissa is one of the poorest states in India, with a population of over 30 million. In terms of infant mortality and maternal mortality, Orissa ranks lowest among the Indian states, with levels comparable to, or just above, levels reported by Bangladesh, Djibouti, Nepal, or Yemen. In 1996, intestinal, infectious, and parasitic diseases caused one out of every five deaths in Orissa, and more than one in 10 deaths were caused by conditions originating in the perinatal period. Over 30% of outpatient visits and inpatient admissions to hospital are for intestinal, infectious, and parasitic diseases (Narayana 2003). Orissa also ranks highest in the prevalence of certain major diseases, such as malaria and tuberculosis, with no perceptible decline over the past 20 years (Gupta 2002).
In 1993, Orissa reported a number of cases of vaccine-preventable diseases. The number of cases of diphtheria and measles had declined somewhat by 1993, but there was no perceptible decline in whooping cough and polio, and cases of tetanus and tuberculosis had, in fact, increased significantly.
The communicable disease situation in Orissa may be summarized in terms of disability-adjusted life years (DALYs). Communicable diseases account for over 60% of total DALYs lost. Perinatal conditions, diarrheal diseases, tuberculosis, and tetanus are the major causes of DALYs lost for both males and females (Padhi and Mishra 2000). In terms of Omran's epidemiological transition model, Orissa may be categorized as falling in the very early transition model, as has been done by the World Bank (see Table 4).
Nowhere in the Orissa Health Services Development Project (OHSDP) appraisal document could we find a systematic treatment of disease control, nor even a mention of it. The project document is completely silent on the issue of integrating disease control programs into the development of the health system in the state. The background document had, however, emphasized that the basic package of health services should address communicable diseases: "A basic health care package should take into account these state level variations in epidemiology and burden of disease. The package of services would consist of communicable disease prevention and treatment; limited clinical services; essential and emergency obstetric and pediatric care" (World Bank 1997, pp. 5–6; italics added). In the transition from the background document to the project document, communicable disease prevention seems to have been left behind. This leads us to examine the situation in Orissa under HSR with regard to disease control programs.
Epidemiological surveillance has improved in Orissa following the setting up of a surveillance system in 1999, under which Department of Public Health staff regularly transmit information on 10 epidemic-prone diseases.5 The initiative was introduced in 12 cyclone-affected districts and was expanded to the entire state a year later. The available evidence suggests the system has been working well. One recent example was the detection of a measles outbreak that prompted the government to institute a mass measles campaign to reduce the number of susceptible children (Gupta 2002).
The introduction of the externally funded Revised National Tuberculosis Control Programme (RNTCP) in Orissa has virtually divided the state in two. The 10 districts where RNTCP has been introduced report sufficient quantities of drugs and chemicals and adequate numbers of trained staff. The program follows the directly observed treatment, short course (DOTS) strategy with regular program monitoring. In the rest of the districts, where the national TB control program is running, the regular treatment regimen of 18 months' duration is available. The national program, in which the state government is supposed to share the drug cost on an equal basis, has not seen any allocation for drugs by the state government since 1999, the year in which OHSDP was initiated. Many of the newly formed districts6 do not have TB centres, and the posts of retired staff have not been filled. Neither of the control programs—RNTCP or NTP (National Tuberculosis Program)—has succeeded in establishing links with private providers who treat TB patients. The net result is that only 40–60% of the expected number of cases are detected in the state (Gupta 2002).
The World Bank-assisted Enhanced Malaria Control Program, which operates in about half of the state's 314 administrative blocks, is one of the largest such programs in India. The program suffers from delays in smear collection and reporting; poor availability of chloroquine tablets in remote hamlets; lack of awareness of presumptive radical treatment (PRT) among the staff in most areas; shortages of slides, lancets, and stains; and delays in reading blood smears collected and in sending the information back to the community. Going by the department's own evaluation, the project, which has been running for three years, has shown no significant development (Gupta 2002).
In situations of fiscal pressure, projects such as OHSDP serve the purpose of levering in external assistance. It is well-known that such assistance entails a commitment to fund part of the expenditure from domestic resources. However, as domestic resources are scarce—which is the original reason for levering in external assistance—diverting funds to such projects leaves other activities, whether of national priority or not, poorly funded (Narayana 1999). The NTP falls exactly into this category. The commitment to "strengthen and thereafter maintain a Policy and Strategic Planning Unit with adequate powers, functions and staff and resources satisfactory to the Administrator" (World Bank 1998, p. 14) leads to diverting personnel from other programs. Existing programs do not receive adequate attention, and their management suffers. This could well be the case of the malaria program.
India is one of the poorest countries in the world where elementary deprivation is concentrated. The country reports low levels of life expectancy and literacy and high levels of both general and infant mortality. Its record in the sphere of human development is also one of the poorest. India reports one of the lowest levels of public spending on health care, and there is increasing recognition among policymakers that public spending on social sectors needs to be enhanced. The persisting high burden of communicable diseases is an added reason for increased public spending.
India, with 28 states, many of which are larger than many countries of the world, has a federal structure wherein health care is a responsibility of the state governments. The central government confines itself to funding medical research and national disease control programs implemented through the state departments of health. By the end of 1980s, public spending on health (capital and current) was in decline, at the level of both the central and many state governments.
Following a BOP crisis in 1991, India initiated macroeconomic adjustment policies. As the Indian Adjustment came after UNICEF's advocacy of "adjustment with a human face," and as the need to safeguard social sectors was recognized, public spending on social services has largely been protected. However, because Adjustment has largely been a concern of the central government, the protection of social spending has occurred at that level, while state governments have followed their own policy directions. Many state governments, under fiscal pressure since the late 1980s, have been cutting both capital and current spending on health care. Some have obtained health systems development loans from the World Bank with the ostensible purpose of reforming the health sector.
The Health Systems Development Project initiated in the seven states recognizes the need for enhanced public spending on health and identifies it as the foremost policy reform to be pursued. Nevertheless, such assurances and conditions have not succeeded in enhancing health sector budgets in states implementing HSR. Worse, HSR has not been able to arrest the decline in the share of health spending within total government spending. The project has been used, rather, by reforming states as a tool for leveraging external financing.
That this external financing comes with its own cost is demonstrated by the case of Orissa. The International Development Association's assistance for the OHSDP is conditional upon the state government contributing about 15% of the total project cost. The funds-starved state has diverted resources from the TB and malaria programs to meet this condition, jeopardizing the disease control programs. This cost can be extremely high, as is shown by the repeated failure of control programs and the fact that disease incidence in the state is barely declining.
The findings of this study raise two fundamental questions regarding HSR in India. First, how is it possible to ensure a certain minimum public spending on health care? Second, is the uniform strategy of focusing on the secondary health sector across all states, irrespective of their health status and health systems development, the right one to pursue?
The Indian system is especially complicated, as the larger tax resources are controlled by the central government but the major responsibility for health-care spending is bestowed on the states. Even though the Constitution has provided for the Planning Commission and the Finance Commission for devolving capital and current resources respectively to the states, there is nothing in the Constitution to ensure that a certain minimum percentage of the total resources is spent on health care. If Colombia and Thailand, under conditions of high or low economic growth, could ensure such spending following amendments to the Constitution, then a similar approach could well be a solution for India with its rapid economic growth.
States in search of external funding often submit the most promising proposals with built-in assurances—all of which look similarly attractive across the states—that they are hardly in a position to fulfill. Surely the dominant funder, the World Bank, with all the technical expertise at its command, bears some responsibility for incorporating the specificity of the state into the project so that disease control programs do not suffer and current spending on health is enhanced?
What may be a good health sector reform strategy for infrastructure-rich, developed southern states, or for Punjab, may not be a good strategy for infrastructure-poor eastern states such as Orissa or West Bengal. The relative roles of access, cost, and quality in the utilization of health-care services are different in these two settings, and this factor must necessarily be incorporated into the HSD project design.
1. Although there are 28 states in India after the formation of Chat-tisgarh, Jharkhand, and Uttaranchal, the discussion carried out here is in terms of the erstwhile Bihar, Madhya Pradesh, and Uttar Pradesh, of which these newly formed states were part. This is largely necessitated by the non-availability of comparable data for the newly-formed states.
2. The Indian financial year is from April to March.
3. Currency: Indian Rupee, INR. Currently, 1 USD = approximately 40 INR. One lakh = 100 000 rupees, and one crore = 100 lakhs, or 10 million rupees.
4. This is a major lacuna in the analysis of public expenditure in India, as many externally funded projects do not get included in government budgetary accounts, and there is no easy way to get information on the total spending on such projects.
5. On 29 October 1999, a severe cyclone hit coastal Orissa, causing severe loss of life and property. Contaminated water and poor availability of food and shelter led to the spread of epidemics, which necessitated a surveillance system.
6. The districts were divided into smaller ones in the late 1990s and the total number of districts increased from 13 to 30.
Ahluwalia, M.S. 2002. Economic reforms in India since 1991: has gradualism worked? Journal of Economic Perspectives, 16(3), 67–88.
Dreze, J.; Sen, A. 1995. India: economic development and social opportunity. Oxford University Press, Delhi, India. 306 pp.
_____ 2002. India: development and participation. Oxford University Press, Delhi, India. 540 pp.
Gupta, M. 2002. State health systems: Orissa. Working paper no. 89. Indian Council for Research on International Economic Relations, New Delhi, India. 37 pp.
Ministry of Finance. 1993. Economic survey 1992–93. Government of India, New Delhi, India. 218 pp. + 115 pp. appendices. Ministry of Health and Family Welfare. 1993. Health information of India. Government of India, New Delhi, India.
Narayana, D. 1999. Public expenditure reform without policy change: infrastructure investment and health care provision under fiscal squeeze in Kerala. In Mackintosh, M.; Roy, R., ed., Economic decentralization and public management reform. Edward Elgar, Cheltenham, UK. 296 pp.
Narayana, D. 2001. Macroeconomic adjustment policies, health sector reform and access to health care in India. Centre for Development Studies, Trivandrum, Kerala, India. 92 pp.
Narayana, D. 2003. Disease control programs and health sector reform in India: the non-intersecting paths? Paper presented at the International Workshop on Health Care Reform: The Missing Jigsaw, 26–28 February 2003, Phuket, Thailand. 11 pp.
Padhi, S.; Mishra, S. 2000. Premature mortality, health status and public health care facilities in Orissa. Nabakrushna Choudhury Centre for Development Studies, Bhubaneswar, Orissa, India. 210 pp.
Registrar General of India. 1999. Compendium of India's fertility and mortality indicators 1971–97. New Delhi, India.
World Bank. 1994. India – Andhra Pradesh first referral health system project. Staff appraisal report. 2 November 1994, Report no. 13402-IN. South Asia Country Department II. Population and Human Resources Operations Division, New Delhi, India. 259 pp.
_____ 1995. India – Policy and finance strategies for strengthening primary health care services. Sector report. 15 May 1995, Report no. 13042-IN. South Asia Country Department II, Population and Human Resources Division, New Delhi, India. 252 pp.
_____ 1996. India – Second state health systems development project. Staff appraisal report. 20 February 1996, Report no. 15106-IN. South Asia Country Department II. Population and Human Resources Operations Division, New Delhi, India. 276 pp.
_____ 1997. India – New directions in health sector development at the state level: an operational perspective. Sector report. 11 February 1997, Report no. 15753-IN. South Asia Country Department II. Population and Human Resources Division, New Delhi, India. 222 pp.
_____ 1998a. India – Orissa health systems development project. Project appraisal document. 4 May 1998, Report No. 17653-IN. Health, Nutrition and Population Sector Unit, South Asia Region, New Delhi, India. 68 pp.
_____ 1998b. India – Maharastra health systems development project. Project appraisal document. 5 November 1998, Report no. 18403-IN. Health, Nutrition and Population Sector Unit, South Asia Region, New Delhi, India. 73 pp.
_____ 2000. India – Uttar Pradesh health systems development project. Project appraisal document. 30 March 2000, Report no. 19570-IN. Health, Nutrition and Population Sector Unit, South Asia Region, New Delhi, India. 75 pp.
_____ 2001. India – Raising the sights: better health systems for India's poor. 28 May 2001, Report no. 22304. Health, Nutrition and Population Sector Unit, South Asia Region. Washington, DC, USA. 187 pp.
This page intentionally left blank
Kerala is well-known the world over for its achievements in reducing levels of both fertility and mortality. This small Indian state is often compared with Sri Lanka, Costa Rica, and China for having attained high measures of human development despite low levels of per capita income. The processes underlying such achievements are thought to be "public action promoting a range of social opportunities relating inter alia to elementary education, land reforms, role of women in society, and widespread and equitable provision of health care and other public services" (Dreze and Sen 1995, pp. 52–3). The rapid decline in mortality and fertility have been attributed to both health care and other factors. However, among the health-care factors, public health-care services alone have been credited with "breakthrough periods" in declining mortality for "taking advantage of the opportunity for reorganizing the health approach to pregnancy, birth and the very young" (Caldwell 1986, p. 182). There is no mention of the size and role of the private health-care sector.
The perspective wherein health-care services are largely seen as government entities is part of a dominant thinking that "is proving more and more difficult to sustain in the face of growing evidence that governments are often minority, or even minor providers of health care in many countries and that they are becoming more so" (Berman 1998, p. 1464). This is especially so in India, which has a highly pluralistic health-care system, and where the private sector is a major provider of health care. It is possible that "the government has never been a major source of care for ambulatory illness care for much of India, although it has been the major provider of personal preventive services and population based public health interventions" (Berman 1998, p. 1465). Kerala is no exception to this Indian phenomenon of a large proportion of health care being provided in the private sector. Increasingly, people in Kerala have been turning to private health-care services and paying out-of-pocket. Two questions arise in this context: why do people turn to private care in a situation of so-called equitable provision of health care, and what implication does recourse to private care have for household incomes? In this chapter, we argue that one of the main reasons people turn to private care is the perceived poor quality of the public service, and that the high utilization of private service is a heavy burden on the poorer segments of Kerala's population.
A study of Kerala's health-care experience is highly relevant for three reasons. First, Kerala is known the world over as a case of public-action driven, policy-assisted development, supposedly characterized by equitable provision of health and other public services. Second, high utilization of health care is generally attributed to provision of insurance, but Kerala shows high utilization in a situation of out-of-pocket expenditure, contrary to what would be expected. Third, income elasticity of health care demand is supposed to be high, but Kerala shows a pattern different from other areas, in that health expenditure is high and not related to income.
In this chapter, after a brief introduction, we begin by presenting some distinct characteristics of Kerala, as background. Then we summarize briefly the dominant explanations of the mortality reduction in Kerala. One explanation most often put forward for the gains in life expectancy in Kerala is the higher utilization of health-care services, and the equitable provision of public services is often taken as the major reason for this high utilization. What is not so evident is that Kerala has a large and growing private health-care sector spread over the rural and urban areas. We describe Kerala's health-care environment and show that the utilization of private services is very high. One reason for this is the perceived poor quality of public service. We then present measures of perceived quality of public and private services in two districts of Kerala. The utilization of private services without insurance means out-of-pocket spending by households. We offer an analysis of their health expenditure that shows how high a burden this is for the poorer segments of population in Kerala. Finally, we draw conclusions and raise concerns about future directions.
Kerala, situated at the southwestern corner of the subcontinent, is one of India's smaller states. It is a narrow strip of land, bounded by the Arabian sea on the west and high mountains on the east, and extending over 600 kilometres from north to south. Kerala accounts for a population of 32 million, within the overall Indian population of slightly over one billion. As per the 2001 census, Kerala's population density is 819 persons/km2. Kerala has the third-highest population density among the Indian states, after West Bengal and Bihar. The population is distributed over 14 districts, among which Alappuzha has the highest density, at 1 489 persons/km2, and Idukki the lowest, at 252 persons/km2.
The steep decline in death rate and fertility in Kerala over the last 30 years has resulted in a rapidly aging population. India has the second largest number of elderly persons (age 60 or older) after China. In 1981, the elderly represented 6.5% of the population; in 1991, 6.7%; and by 2016, the proportion is expected to reach around 9%. By comparison, the proportion of elderly in Kerala was 6.2% in 1971, 7.5% in 1981, and 10.6% in 1991. By 2001, it was estimated to be around 15% (Economic Review 2002).
Despite having achieved high levels of life expectancy and low mortality rates, in particular infant mortality, at fairly low levels of income, Kerala reports one of the highest levels of morbidity. The morbidity rate is twice the all-India average in the rural areas and over 50% higher than the Indian average in the urban areas. In 1995/96, the number (per 1 000 population) of persons reporting an illness during the previous 15 days in Kerala's rural areas was 118, compared to the all-India average of 55, and in the urban areas 88, compared to 54 at the all-India level. Acute morbidity in Kerala is 80 in rural areas and 61 in urban areas, compared to 42 and 41, respectively, in rural and urban India. Chronic morbidity rates in rural and urban Kerala are 38 and 27, respectively, compared to 13 and 14 in rural and urban areas of India. The rural morbidity rate is over 40% higher than the urban morbidity rate; such a differential is not to be found in any of the major Indian states (NSSO 1998).
Kerala has a unique position with regard to gender ratio, with 1 058 females per 1 000 males in 2001, whereas the Indian average is only 933. The ratio in the two richest states is low: 861 in Haryana and 874 in Punjab. Kerala reports high levels of literacy—94% among males and 88% among females in 2001—compared to the all-India average of 76% and 54%, respectively. The exceptionally high level of female literacy and the low gender gap in literacy in Kerala are noteworthy, as these generally lead to lower infant mortality rate (IMR) and fertility. Kerala reports comparatively low IMR (16 per 1 000 live births in the mid-1990s) and total fertility rate (TFR) (1.8 live births expected per woman in her reproductive years). In terms of IMR, a half-dozen Indian states currently are where Kerala was in the early 1970s, while the others are worse off, and the Indian average TFR is almost twice Kerala's.
Thus, Kerala is more developed than the other Indian states, whichever aspect of human development is considered. Literacy levels, especially among females, are exceptionally high and mortality reduction in the last 50 years is one of the most rapid among all the states. At the same time, the fertility rate has fallen almost below replacement level. What explains such remarkable gains in Kerala?
In the mid-1970s, a group of researchers at the Centre for Development Studies, Trivandrum, concluded that Kerala, "a relatively poor state," had made "fairly impressive advances in health and education." Their study postulated that the sharp decline in birth rates in Kerala resulted from "societal changes in attitude to family size resulting from longer life expectation, reduction in infant and child mortality and female education" (United Nations 1975, p. 153). The societal changes in attitude, in turn, had been spurred by substantial government investment in health and education.
Kerala has devoted the bulk of public funds to infrastructure and social services, particularly in education and health (World Development Report 2004). Investment in health is guided by three objectives: (i) a public medical institution in every panchayat (see below); (ii) government funded medical education, to ensure an adequate supply of medical personnel; and (iii) encouragement of traditional medical systems. Gradual implementation of this strategy led to most districts having one medical institution (allopathy) per 20 000 population by 1990. The expansion of facilities under the Indian system of medicine—which includes ayurveda, homeopathy, and others—has also progressed as per the accepted policy. The emphasis on publicly funded medical education has led to an increase in the share of medical education in the total health expenditure from around 10% in 1977/78 to over 17% by 1991/92. (For detailed treatment of these issues, see Narayana 1999.) The strategy has seen two changes in recent years, namely, decentralization in the early 1990s and permitting of private medical educational institutes in the early 2000s.
India has a long history of experimenting with decentralization. The 73rd and 74th amendments to the Indian Constitution adopted in 1993 introduced a third tier of government, called panchayat, below the union and states and required states to delegate some administrative functions and taxation powers to local bodies. These amendments constituted a watershed, in that they left no room for the successive state legislatures to change the local self-government institution (LSGI) according to their whims and fancies. They brought about a uniform three-tiered structure of local governance—district, block/taluk, and village—in the rural areas. Only for small states with populations under two million was the intermediate tier optional. The amendments also introduced the concept of township for smaller urban centres, that is, rural areas in the process of urbanization.
Among the states, Kerala was in the forefront of decentralization of powers. The large public health-care sector in Kerala was also affected by the decentralization. As per the government order dated 18 September 1995, following the Panchayatiraj Act, the primary health centres (PHCs) and government dispensaries were transferred to the village panchayats; block PHCs, community health centres (CHCs), taluk headquarters hospitals, and government hospitals, to block panchayats; and CHCs, government hospitals, and taluk headquarters hospitals in corporation and municipal areas, to the corporation councils and municipal councils. In 1999, the district hospitals were brought under the control of district panchayats. Similar transfers of institutions and staff have taken place in the cases of ayurveda and homeopathy as well. Thus, in Kerala, except for medical colleges and a few general hospitals, all health institutions are managed by the local governments.
Following decentralization, the personnel of these institutions are under the supervision and disciplinary authority of the local bodies during their tenure with them, but their service conditions remain undisturbed. Further, the government continues to pay salaries, allowances, and other dues to the employees and officers transferred to the local bodies from the central government. Thus, the new system envisages dual control over the staff (Narayana and Hari Kurup 2001).
The smooth functioning of the institutions is to be ensured by the constitution of certain committees. There are supposed to be standing committees for health and sanitation at the level of the municipal council and municipal corporation in the urban areas, and at the block panchayat and district panchayat in the rural areas. These standing committees are to deal with matters relating to public health and health service, sanitation, environment, dangerous and offensive trades, education, arts and culture, and sports. At the district panchayat level, a management committee must be constituted for every public health institution transferred to it by the government.
In Kerala, public medical institutions have been brought under the administrative control of the LSGIs, which receive intergovernmental transfers, the criteria being population size, area, and the proportion of SC/ST1 population. The taluk hospitals and district hospitals also have duly constituted hospital development committees (HDCs), which collect user charges on some services and carry out some regular maintenance, cleaning, and repair.
The achievements of Kerala in the field of mortality and fertility were attributed to widespread health coverage. Although it was recognized that coverage was not lower in many other states, what distinguished Kerala was the extremely high utilization ratio, thought to have been enhanced by better physical accessibility (Krishnan 1975). Panikar (1979) added a new dimension by postulating that the progress of literacy and education had not only increased the rural population's awareness of the need to use health facilities, but also made it aware of its rights to demand and use them. Nag (1983, 1989) marshaled evidence to show that a high degree of political awareness in rural Kerala contributed to easier accessibility and better utilization of health facilities. In a comparative exercise with West Bengal, Nag (1985) brought out the aspect of higher utilization of public health facilities in Kerala. In terms of number of health facilities and medical personnel, the differences between the two states were not significant, but utilization of medical facilities—both preventive and curative—was distinctly higher in Kerala. The higher utilization was then explained in terms of accessibility, better transportation, and better literacy, especially female literacy:
Higher female literacy in Kerala has contributed significantly toward higher utilization of maternal and child health services by the women of the state. They are also less inhibited regarding physical examinations and more open to traveling to the health centres without male escorts. Indirectly higher literacy has contributed to a higher level of political participation of the rural poor, which, in turn, has forced the state government in the post-independence era to cater to their health needs along with other needs (Nag 1985, p. 68).
The need and right