Poverty Reduction Strategy Papers and their contribution to health: An Analysis of Three Countries.

Poverty Reduction Strategy Papers (PRSPs) represent the World Bank and the International Monetary Fund's (IMF) most recent initiative for reducing the plight of the poor. This paper examines whether the PRSPs for Liberia, Afghanistan and Haiti follow World Bank guidance on health. The health data, analysis and strategy content of the three PRSPs are assessed with respect to the 'Health, Nutrition and Population' chapter of the World Bank's PRSP Sourcebook. This guidance states that PRSPs should include: health data on the poor and a clear analysis showing the determinants of ill health and pro-poor health strategies. Unfortunately, none of the PRSPs analysed comply with the guidance and, consequently, do not adequately portray the health situation within their countries. Thus health is not given a high priority in the PRSP process and is seemingly low on the agenda of both poor country governments and the International Financial Institutions (IFIs). If the situation for the world's poorest people is to improve, health and the right to health need to be promoted within PRSPs.

INTRODUCTION

The relationship between poverty and health has been widely reported. Moreover, extreme poverty is classified as a disease in itself (1). More than one billion people worldwide live in extreme poverty, which equates to an equivalent income of less than $1.25 per person per day at purchasing power parity. Reducing poverty is an international priority and is the focus of the Millennium Development Goal 1.

The World Bank and the International Monetary Fund (IMF) are United Nations organisations. They are International Financial Institutions (IFIs) that are involved in global development, providing countries with technical, operational, and financial assistance. Poverty Reduction Strategy Papers (PRSPs) represent their most recent initiative for reducing the plight of the poor.

This paper seeks to assess whether the health content of three PRSPs, Liberia, Afghanistan and Haiti, follow the guidance provided by the World Bank. Specifically, the PRSPs are examined with respect to Chapter 18 of the World Bank's Sourcebook entitled 'Health, Nutrition and Population' (from now on HNP Sourcebook) (2) concerning health data, data analysis, and health strategies.

A BACKGROUND TO PRSPS

PRSPs are documents written every three years by governments of less economically developed countries (LEDCs) and are intended to "describe a country's macroeconomic, structural and social policies and programs to promote growth and reduce poverty, as well as associated external financing needs" (3). PRSPs promote 'national ownership' and place poverty reduction at centre stage (4). Governments can produce an interim PRSP (iPRSP) before developing a full PRSP, enabling them to qualify for partial debt relief without having to wait until a full PRSP is prepared (5). The World Bank and IMF analyse a country's PRSP and write a Joint Staff Assessment Note (JSAN). Countries require positive JSAN feedback to receive debt relief (6). In addition to debt relief, PRSPs represent the main mechanism for LEDCs to receive both loans from the IMF and World Bank (6) and foreign aid (7).

There are fears that countries develop PRSPs with a rushed approach, which are biased in favour of donor wishes, in order to receive much needed financial aid (5, 6). In addition, even though the countries preparing PRSPs have different economic climates, resources, and governments, PRSPs have resulted in very similar economic policies. The homogenous nature of economic policies given in PRSPs, suggests that IFIs are influencing LEDC governments and thus that there is a lack of country-ownership (8).

Both iPRSPs and PRSPs have been criticised for their inadequate health content. For example, Laterveer et al. (9) found few iPRSPs gave disaggregated health data for income and most did not identify the major causes of illness amongst poor populations. In one study, six of twent-one PRSPs gave no disaggregated health data at all (10). This results in the PRSPs that do not discuss the inequalities in health between the rich and the poor and thus, this questions their poverty focus. Major causes of illness were identified but with scant evidence and analysis (10). In addition, there is a lack of analysis on why and how health systems are failing the poor (10). Niger and Mozambique, for example, provided disaggregated health data by district and yet the inequalities identified were not addressed in their health system strategies (10). Additionally, it is difficult to determine whether health strategies outlined in PRSPs are new and due to the PRSP process or are from existing health strategies (5, 10). It may be that governments simply 'copy and paste' existing health policies into PRSPs, possibly in an attempt to complete the PRSP as quickly as possible.

Whilst PRSPs seemingly approach health from a developmental perspective, as a tool for increasing human capital and economic growth, none address health as a human right (1, 6, 10). The right to health is enshrined in various human rights documents like Article 12 of the UN International Covenant on Economic, Social and Cultural Rights (ICESCR) (11) and within the constitution of the World Health Organisation (WHO). Furthermore, every State has an obligation to work towards a global realisation of this right.

METHODS

All PRSPs available on the World Bank website were included in the selection process. Countries which submitted PRSPs from 2005 onwards were first selected. It is clear that countries with multiple PRSPs tended to include much more health data in the first PRSP than the subsequent papers. Therefore, this analysis was confined to a country's first PRSP. Thirteen PRSPs met these inclusion criteria and, in order to facilitate a detailed analysis, three PRSPs were selected at random from different continents. Wider determinants of health such as water, sanitation, housing and education are, in practice, intractable from this discussion, however are addressed outside of the HNP Sourcebook. In addition, although the HNP Sourcebook discusses nutrition, PRSPs commonly discuss it separately from health. Therefore, to maintain consistency and clarity, these factors are not discussed at length.

THE WORLD BANK'S SOURCEBOOK

The World Bank defines the Sourcebook as a framework and not a blueprint, emphasising that it is not mandatory for countries to rigidly follow it (12). The HNP Sourcebook briefly discusses the relationship between poverty and health, subsequently addressing the key stages in policy design; diagnostics and analysis, government action, prioritisation, and monitoring and evaluation (2). A summary of the recommendations made by HNP Sourcebook is given in Figure 1.

Figure 1:

The main recommendations made by the HNP Sourcebook. Note: Examples of household factors are: household income and knowledge of health services. Examples of community factors are: cultural norms and community institutions. Examples of health service factors are: accessibility of services and human resources within services.

THREE PRSPS: AN ANALYSIS

All three countries in this analysis (Liberia, Afghanistan, and Haiti) face challenges overcoming damaged infrastructures and dilapidated public services. Thus, they face considerable difficulties in collecting health data and subsequently developing health policy.

Data

All three PRSPs contain only limited health data and disaggregated data is missing despite the HNP Sourcebook emphasising its importance. Consequently, the health situation within the countries, the extent of health inequalities, and the population groups in greatest need are not reported. For example, the PRSP for Liberia gave the infant mortality rate (IMR) as 72 deaths per 1000 live births. However, it did not break this down by population group. Prior to the PRSP, the Government of Liberia conducted a Demographic Health Survey (DHS) (13) which showed that the IMR was higher for poorer households, yet this evidence was not included in the PRSP. Similarly, in Afghanistan, maternal mortality rates have been shown to vary between districts and are fifteen times higher in Ragh (rural) than Kabul (urban) (14), but this information is not included in their PRSP. Without such data, it is difficult to understand the variations in health outcomes between population groups and one cannot determine the burden of disease amongst the poor. A checklist of the data included in the PRSPs compared with the recommendations made by the HNP Sourcebook is shown in Figure 2.

Figure 2:

Data checklist for PRSPs vs the HNP Sourcebook recommendations

Analysis

All three PRSPs give some aetiology of the health outcomes they provide data for, however, evidence for household and community factors is limited. This is a potential problem considering an understanding of community factors like social and cultural norms, is imperative for the design of many interventions, especially community interventions or education programs. For example, in some parts of Afghanistan, there are normalised gender inequalities such that a women must require consent from her husbands in order to visit a health facility (15). This cultural practice is which are critical to the health situation of women in this cournty, yet Afghanistan's PRSP does not mention this.

The most commonly cited health system limitations are physical accessibility and availability of human resources, the importance of which is emphasised in the HNP Sourcebook. However, all three PRSPs lacked the disaggregated data required to fully understand the barriers facing the poor.

Providing pregnant women with professional supervision at the time of delivery is important to maternal health and yet all PRSPs approached this with inadequate analysis. All three PRSPs included the percentage of births attended by a trained health worker, which was below 50% in all cases. However, there is no explanation why this is low (16-18). Furthermore, none of the PRSPs give disaggregated data by income, wealth, or region. The Liberian DHS provides disaggregated data on the percentage of births attended by a trained health worker and showed this to be lower for women who give birth at home, in rural areas and, in lower income groups (13). Once again, this data was not included in Liberia's PRSP (17) hindering potential links that could be made between health outcomes and specific limitations within the health sector.

In another example, pregnant women in Ragh, Afghanistan, are more likely to die during birth than women in Kabul (14). In addition, none of the women who died during pregnancy in Ragh had a health worker present during delivery (14). Using this information, it could be suggested that the health system is failing women in Ragh and additional information is needed to understand why. Maybe health facilities in Ragh are inaccessible to most women or perhaps pregnant women are being denied the opportunity to give birth in a health facility by their families. However, this information is missing in Afghanistan's PRSP and similar content is missing in the PRSPs of both Liberia and Haiti. All three PRSPs highlight maternal mortality as a concern and yet the gravity of maternal health is insufficiently analysed. A checklist of analyses included in the PRSPs has been given in Figure 3.

Figure 3:

Analysis checklist for PRSPs vs the recommendations made by the HNP Sourcebook

Health Strategies

Drawing links between the burden of disease within the countries and the health strategies they propose is difficult given the lack of data and analysis. The problem worsens in the absence of disaggregated data, making it hard to establish whether countries are targeting the poor or worse off population groups. All three PRSPs align their objectives with the Millennium Development Goals ― specifically 4, 5 and 6 (reducing maternal and child mortality and the prevalence of major diseases, respectively) and all aim to improve accessibility to and the quality of health systems (16-18). Although none of the PRSPs explicitly target the poor within their health strategies, they do target rural populations, notably by aiming to improve primary health care and to increase accessibility to a basic package of health services (BPHS) (16, 17). Haiti's PRSP does not mention implementing a BPHS (8). A BPHS could be of benefit in Haiti considering nearly half of the population lack access to basic healthcare and, consequently, 80% turn to traditional medicine (19).

All PRSPs are deficient in detailed links between analysis and health strategies. In all three countries, women in rural areas are less likely to receive medical attention during labour (13, 14, 20). Accessibility to facilities amongst pregnant women is questionable. This is noted in Afghanistan's PRSP (16) and Chatterjee (19) reports that some women in Haiti face a six hour or more journey to the nearest health facility. Because of such barriers, "many pregnant women die en route" (19). One approach to tackling the high maternal mortality in these countries could be to provide delivery care in women's homes. If performed by a skilled birth attendant and appropriate referral mechanisms are in place, this can achieve marked reductions in maternal mortality (21). If there were greater discussion within the PRSPs, that forged links between health outcomes and their aetiologies, a better understanding of the interventions required to improve the health of those most in need could be discerned. A checklist for health strategies is shown in Figure 4.

Figure 4:

Health strategy checklist for PRSPs vs recommendations made by the HNP Sourcebook

DISCUSSION

This paper seeks to assess whether the health contents of three PRSPS (Liberia, Afghanistan and Haiti) are aligned with the suggestions made in the HNP Sourcebook, in relation to health data, health data analysis and health strategies. It was found that the three PRSPs contain inadequate health data (most notably disaggregated data) and insufficient analysis needed to portray the health situation in each country. Consequently, it is difficult to identify the health strategies which target the poor. This may be seen by donors as a lack of government capacity, which could threaten the aid the countries receive, with disastrous effects for the health sector. It is understandable, given the recent history and current situation of these three countries, that it may not be feasible to collect comprehensive data. However, in some instances, official data from surveys, for example, was available and was not included in the PRSPs, representing a serious missed opportunity. This suggests that governments are not paying due attention to health, which could be due to a number of reasons; governments may take a rushed approach when preparing PRSPs in order to quickly benefit from debt relief or t health may not be a high priority on the country's domestic agenda. Governments may believe the IFIs do not consider health to be a priority for PRSPs and make conscious decisions to focus on other areas.

The feedback from the World Bank and IMF on the PRSPs of Liberia, Afghanistan, and Haiti does not comment on the absence of disaggregated health data and analysis of health outcomes or provide a critique of the health strategies proposed in the PRSPs (22-24). The World Bank and IMF have produced guidelines (25) for the staff responsible for writing the JSANs (referred to by the World Bank and IMF as 'the staffs'), which state that these should be "selective, and present a limited number of priority areas for strengthening". Therefore, it can be assumed the staffs did not consider the gaps in health that have been highlighted in this paper from the three PRSPs to be high priorities. It is possible, then, that the World Bank and IMF actually do not consider health to be a priority in PRSPs.

The HNP Sourcebook, along with PRSPs, does not recognise health as a human right. This further supports the hypothesis that health is not high on the IFIs agenda. Hammonds and Ooms (26) suggest that, in order to protect the right to health, LEDCs are obliged to raise sufficient funds for health. However, the World Bank and IMF recommend governments restrict the amount they spend on health in order to maintain macroeconomic stability. This can reduce the amount of money LEDCs ask for from donors and in turn reduces the amount of money they receive (26). As a result, the authors posit that the World Bank does not allow countries to raise enough money to uphold their obligations to protect the right to health. They argue that countries partial to the ICESCR - which make up over 75% of the votes on the International Development Association (IDA) board - should use their influence to compel the World Bank to align with the ICESCR and recognise health as a human right. Verhuel and Rowson (27) go further and argue that human rights should be embodied in policies prescribed in PRSPs and that countries should reject reforms proposed by the World Bank and IMF that result in the breach of these rights.

In addition, whilst analyses of health in PRSPs have been carried out, to my knowledge no such analysis or critique of the HNP Sourcebook exists. Thus, the guidelines provided by the World Bank and used by countries to prepare the health sections of PRSPs, have not been analysed in depth and accordingly revised.

Health and the right to health require further promotion in the PRSP process and the World Bank's agenda needs to be challenged. The WHO is responding to calls for assistance from LEDCs by developing a database to monitor health in PRSPs and by providing guidance to countries in the preparation of PRSPs (10). For this to be successful, a strong leadership is needed from the WHOin order to raise the priority of health at all levels.

CONCLUDING REMARKS

The three PRSPs analysed do not provide adequate health information as recommended by the HNP Sourcebook. Consequently, none of the PRSPs clearly portray the health situation in their country and thus PRSPs are not affording due attention to health. It appears that health is a low priority for the World Bank and needs further promotion through the PRSP process so that health is recognzed as a human right and is subsequently placed firmly in the centre of the development agenda. PRSPs represent an opportunity for LEDCs to bring their people out of poverty. However, if health continues to receive insufficient attention, the situation for the world's poorest people unlikely improve.