SEARCH

SEARCH BY CITATION

Keywords:

  • Maternal;
  • mortality;
  • ratio;
  • measurements;
  • sub-saharan Africa;
  • MDG-5

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. What has happened since 1987 in Nairobi?
  5. Problems being encountered
  6. Moving forward
  7. Disclosure of interest
  8. Contribution to authorship
  9. Funding
  10. Acknowledgements
  11. References

Monitoring of maternal mortality levels in sub-Saharan Africa (SSA) to assess the achievements of safe motherhood programmes and for MDG-5 has been made difficult because of the lack of precise estimates of the maternal mortality ratio (MMR). Projections based on the slow rate of decline of the MMR indicate that MDG-5 may not be reached before the end of this century in this region. Measurements done using demographical and health surveys, statistical modelling and censuses are imprecise and do not allow trends in individual countries to be established. SSA countries should be encouraged to measure mortality levels from their own resources, using methods that produce precise estimates such as population-based surveys. Establishment of the trends will lead to country-specific program targets. The less frequent but more precise measurements can be afforded by SSA countries, as a case study from Zimbabwe shows.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. What has happened since 1987 in Nairobi?
  5. Problems being encountered
  6. Moving forward
  7. Disclosure of interest
  8. Contribution to authorship
  9. Funding
  10. Acknowledgements
  11. References

At a conference to commemorate the 20th anniversary of the Safe Motherhood Initiative in 2007, many reports were presented describing developments in maternal health around the world. Sub-Saharan Africa (SSA) stood out as the region which had experienced the least success. According to the World Health Organisation,1 the annual change in maternal mortality ratio (MMR) in the region was only −0.12%. Figure 1 shows a linear model based on the data from 1990 to 2005. If the estimated tiny reduction of MMR in SSA was to persist, then it will take a very long time to achieve Millennium Development Goal (MDG) number 5, which is a reduction of 75% from the 1990 level. Even though a projection that far into the future may be unreliable (the rate of change is unlikely to be constant), common sense would suggest that the goal is unlikely to be reached during this century. A similar projection of data from South Asian countries suggests that MDG-5 will be achieved there in the year 2076.2 It is not known when individual countries in SSA might reach the target because no precise estimates of the national MMR have been established over the 20-year period. To monitor change, a country would have needed at least two precise estimates during the 20 years. There is no formal definition of what ‘precise’ means but it has been suggested that the 95% confidence intervals around such an estimate should not exceed ±15%.3

image

Figure 1.  Decline of maternal mortality ratio estimates in sub-Saharan African countries.

Download figure to PowerPoint

What has happened since 1987 in Nairobi?

  1. Top of page
  2. Abstract
  3. Introduction
  4. What has happened since 1987 in Nairobi?
  5. Problems being encountered
  6. Moving forward
  7. Disclosure of interest
  8. Contribution to authorship
  9. Funding
  10. Acknowledgements
  11. References

The Safe Motherhood Initiative was launched in 1987 in Nairobi. Among the recommendations were the need to study levels and causes of maternal mortality, especially in developing countries. But the absence of vital registration in the SSA region makes it very difficult to make a direct estimation of the MMR. Although many maternal mortality studies have been conducted and published, the vast majority was about causes of deaths, and very few of them reported actual levels of maternal mortality. And if they did, the studies reported MMR estimates from institutions, whose populations were hardly representative of the country. Population-based surveys, which were needed because the majority of women delivered and died at home, could not be considered because of logistical issues. To identify births and maternal deaths for the sample which was required, enumerators would have had to reach a large number of rural homesteads. This was thought to present insurmountable challenges in most countries and be very expensive.

Within SSA countries, the political will demonstrated by signing declarations in international forums was not matched by advocacy and the introduction of effective interventions to reduce maternal mortality. Safe motherhood programmes were put in place but most of them were weak and did not have a strong monitoring and evaluation component. There was insufficient information about the magnitude of the problem to justify budgetary requests, and the problem became worse as the HIV/AIDS pandemic diverted the attention and resources of governments. At this time, national estimates of the MMR in SSA countries were being made available from demographical and health surveys, statistical modelling and censuses. These methods provided information about the magnitude of the problem, but they could not be used for monitoring due to the wide confidence intervals around the estimates.4–6 In other words, successive plots on the graph could not be joined. This problem is illustrated in Figure 2 which shows six plots from estimates done in the last 20 years in Zimbabwe. Three of them are from the World Health Organisation,5,7,8 which used statistical modelling, and the other three are from demographical and health surveys, which used ‘sisterhood methods’.9–11 None of the six plots can be joined, not even the three done by the same organisation. The WHO estimates were done using noncomparable methods each time, and the DHS reports (which did not publish confidence intervals for the estimates) specifically mention that the results should not be used for monitoring trends. An analysis of the other SSA countries would reveal the same pattern; multiple estimates in the last 20 years, none of which can be used to monitor progress in safe motherhood.

image

Figure 2.  Maternal mortality ratios for Zimbabwe from the World Health Organisation and Demographical and Health Surveys between 1994 and 2005.

Download figure to PowerPoint

There have been calls to measure maternal mortality with precision but they have not yet been heeded by national governments or other interested parties.12,13 The international community feared that poorly done population-based studies would set the cause of safe motherhood back by giving wrong estimates. The difficulties of obtaining precise estimates in SSA countries led some to question the need to retain the MMR as an indicator, and others to suggest that process indicators would be sufficient for monitoring. However, the relationship between the MMR and process indicators is complex, and the latter are only proxies for health outcomes. For example, commonly used indicators such as the presence of a skilled attendant at birth or antenatal care attendance are not necessarily good predictors of maternal outcomes.12,14

Problems being encountered

  1. Top of page
  2. Abstract
  3. Introduction
  4. What has happened since 1987 in Nairobi?
  5. Problems being encountered
  6. Moving forward
  7. Disclosure of interest
  8. Contribution to authorship
  9. Funding
  10. Acknowledgements
  11. References

The major problem now is that nobody knows what is happening to maternal mortality in SSA countries, except that it is known to be high. The change of −0.12% in MMR per year between 1990 and 2005 in the region included a mix of countries where the MMR is unchanging, decreasing or increasing. Countries are unable to learn from their own experiences as the overall trend is often not known. Neither can they learn from other countries. It would be instructive to know in which countries the MMR is decreasing, and to compare their programme content and implementation with those whose outcomes are less favourable, but this is currently not possible. Countries such as Malaysia and Sri Lanka have reduced maternal mortality significantly in the last 50 years, but their lessons for SSA may be limited, given the differing contexts.

The use of estimates from international organisations let national governments in SSA off the hook. Not only did they not feel under any pressure to obtain their own estimates, but they could question the figures provided and find reasons to disbelieve them. During this time, there was no capacity building by national governments for the skills required for monitoring and evaluation, with the result that there were no advocates for the type of study required to obtain precise estimates.

Moving forward

  1. Top of page
  2. Abstract
  3. Introduction
  4. What has happened since 1987 in Nairobi?
  5. Problems being encountered
  6. Moving forward
  7. Disclosure of interest
  8. Contribution to authorship
  9. Funding
  10. Acknowledgements
  11. References

Because of the length of time it will take to reduce maternal mortality to acceptable levels, it is reasonable to stop, take stock and, if necessary, go back to the beginning regarding the measurement of maternal mortality levels. The MMR is the most important indicator for women’s health in SSA, because of the high fertility rates in the region. A continent in which women are disempowered needs an affirmative indicator, such as the MMR, which not only assesses access to health but progress in education, employment and social status. It is the female development indicator in SSA.

A range of methods have been advocated for estimating the MMR in developing countries,13,15 but perhaps only a few are applicable to SSA at this time. Zimbabwe chose to do a population-based survey, combined with a reproductive age mortality study (RAMOS) in 2007. This was a cross sectional study of pregnancy outcomes in which 45 000 live births were recruited over a period of 1 year.16 The fieldwork in 2007–8 established that the logistical challenges can be overcome. The MMR estimate obtained has narrow confidence intervals, which will allow a change of 20% or more in the MMR to be detected in 10 years. Zimbabwe is planning to repeat the study in 10–15 years and will then be able, for the first time, to join two dots. The attributes which made this possible in Zimbabwe are present in many SSA countries, so replication can be attempted. It is of critical importance that other SSA countries should obtain estimates precisely enough to be able to join dots every 10–15 years.

Even though there are no ‘gold’ standards for the MMR in SSA, the ratios should be established as reliably and precisely as possible, because we have no choice. Wide confidence intervals will reduce the ability to detect change. If the MMR is underestimated at the first attempt, it will be more difficult to show improvement with subsequent estimates. Conversely, if it is overestimated at the first attempt, subsequent estimates may give a false sense of improvement, and it will be more difficult to detect an increase in the MMR. The MMR estimates should not be changed to include deaths of women up to 1 year after pregnancy, however desirable this may be. Doing so will put it out of reach of most SSA countries.

Figure 3 shows the essential steps required to monitor progress in safe motherhood programmes. In Zimbabwe, as in many SSA countries, the third step has been introduced before the first two. Goals and targets need to be individualised by country. Such differentiation will allow countries to identify their weaknesses. The role of the United Nations’ agencies and other organisations will be to advocate for precise estimates, assist the scientific process and respond to requests for capacity building. The World Health Organisation has already published a paper which sets out standards for data collection in maternal mortality studies,17 and with others they should build on this and propose guidelines for population-based surveys and other appropriate methodologies. An important standard could be that every study must be better in methodology and execution than the last one.

image

Figure 3.  Essential steps to determine progress in safe motherhood programmes.

Download figure to PowerPoint

It is, however, very important that countries must ‘own’ the estimates, so the international community should support but not drive the process. Estimates which a country obtains through its own initiative, using its own people and resources are very difficult to ignore. A study of the type carried out in Zimbabwe will cost a government less than USD 0.10 (10 US cents) per woman of reproductive age per year every 10 years. This is within the budgets of SSA countries. Studies can be made cheaper by learning the lessons from previous attempts such as the one from Zimbabwe. In the long term, women in SSA countries cannot be helped until national governments are empowered to measure maternal mortality levels and monitor trends mainly from their resources.

The other methods used to derive estimates up to now, such as statistical modelling, and ‘sisterhood methods’ should be phased out and the funds diverted to methods which give precise and comparable estimates. In SSA, the MMR is too important to be an ‘add-on’ to routine surveys, given the imprecision of the estimates. It remains to be seen whether census estimates, which are proposed for the next round in 2012, will reach the required precision.6 A simple test for an intended survey would be: Does the methodology allow the country to join two dots every 10 years? If the methodology of the survey or census fails this test then there will be nothing lost by waiting for a better opportunity. Only by admitting our gaps in knowledge are we going to focus on what needs to be performed.

The graph of the decline in MMR in Figure 1 shows that we are in for the long haul. We are still near enough to the beginning of the journey to set the correct course.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. What has happened since 1987 in Nairobi?
  5. Problems being encountered
  6. Moving forward
  7. Disclosure of interest
  8. Contribution to authorship
  9. Funding
  10. Acknowledgements
  11. References

The author would like to thank J Hussein, D Braunholtz, L D’Ambruoso and E Fottrell for assistance with Figure 1 and to J Hussein and Lennarth Nystrom for the comments they made on the first draft of the paper. Thanks to Maxwell Chirehwa for assistance with Figures 1 to 3.This work was partly supported by the Umea Centre for Global Health Research, with support from FAS, the Swedish Council for Working Life and Social Research (grant no. 2006-1512). The Department for International Development (UK) funded the Zimbabwe Maternal and Perinatal Mortality Study which is referred to in the article.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. What has happened since 1987 in Nairobi?
  5. Problems being encountered
  6. Moving forward
  7. Disclosure of interest
  8. Contribution to authorship
  9. Funding
  10. Acknowledgements
  11. References
  • 1
    Hill K., Thomas K, AbouZahr C, Walker N, Say L, Inoue M, et al. Estimates of maternal mortality worldwide between 1990 and 2005: an assessment of available data. Lancet 2007;370:13119.
  • 2
    Hussein J, Braunholtz D, D’Ambruoso L. Maternal health in the year 2076. Lancet 2008;371:2034.
  • 3
    Hill K, El Arifeen S, Koenig M, Al-Sabir A, Jamil K, Raggers H. How should we measure maternal mortality in the developing world? A comparison of household deaths and sibling history approaches. Bull World Health Organ 2006;84:17380.
  • 4
    Stanton C, Abderrahim N, Hill K. An assessment of DHS maternal mortality indicators. Stud Fam Plann 2000;31:11123.
  • 5
    WHO/UNICEF/UNFPA and The World Bank. Maternal Mortality in 2005. Estimates Developed by WHO, UNICEF, UNFPA and the World Bank. Geneva: World Health Organisation, 2007.
  • 6
    Stanton C, Hobcraft J, Hill K, Kodjogbe N, Mapeta WT, Munene F, et al. Every death counts: measurement of maternal mortality via a census. Bull World Health Organ 2001;79:65764.
  • 7
    WHO/UNICEF/UNFPA. Maternal Mortality in 2000: Estimates Developed by WHO, UNICEF, UNFPA. Geneva: World Health Organisation, 2004.
  • 8
    WHO/UNICEF/UNFPA. Maternal Mortality in 1995. Estimates Developed by WHO, UNICEF, UNFPA. Geneva: World Health Organisation, 2001.
  • 9
    Central Statistical Office (Zimbabwe) and Macro International Inc. Zimbabwe Demographic and Health Survey 1994. Calverton, Maryland: Central Statistical Office and Macro International Inc, 1995.
  • 10
    Central Statistical Office (Zimbabwe) and Macro International Inc. Zimbabwe Demographic and Health Survey 1999. Calverton, Maryland: Central Statistical Office and Macro International Inc, 2000.
  • 11
    Central Statistical Office (Zimbabwe) and Macro International Inc. Zimbabwe Demographic and Health Survey 2005-6. Calverton, Maryland: CSO and Macro International, 2007.
  • 12
    Graham W, Hussein J. The right to count. Lancet 2004;363:678.
  • 13
    Graham WJ, Ahmed S, Stanton C, Abou-Zahr C, Campbell OM. Measuring maternal mortality: an overview of opportunities and options for developing countries. BMC Med 2008;6:12.
  • 14
    Shah IH, Say L. Maternal mortality and maternity care from 1990 to 2005: uneven but important gains. Reprod Health Matters 2007;15:1727.
  • 15
    Immpact (Initiative for Maternal Mortality Programme Assessment). Immpact Web Toolkit version 2.0. 2009 [cited 27/07/2009; available from: http://www.immpact-international.org/index.php?id=67&top=6].
  • 16
    Ministry of Health and Child Welfare. Maternal and Perinatal Mortality Study 2007. Harare; Government of Zimbabwe; 2009.
  • 17
    Gulmezoglu AM, Say L, Betran AP, Villar J, Piaggio G. WHO systematic review of maternal mortality and morbidity: methodological issues and challenges. BMC Med Res Methodol 2004;4:16.