IMMPACT is a global research initiative that aims to promote better health for mothers-to-be in developing countries. Their research focus is on evaluating safe motherhood intervention strategies. IMMPACT is based at the University of Aberdeen in Aberdeen, United Kingdom.
IMMPACT is an international research project tasked to clarify the complexity that characterises “what works” for maternal health in developing countries, IMMPACT (the Initiative for Maternal Mortality Programme Assessment at the University of Aberdeen) also read with great interest the recent Lancet paper that demonstrated the first ever cross- national reductions in maternal mortality (Hogan et al, 2010). Echoing Dr Sanghvi, we too hope that these estimates signal a turning point, whereby the prevailing lack of progress (and negativity) that surrounds progress in maternal health in resource-poor settings is beginning to change, and the efforts of many committed individuals and institutions are finally bearing fruit.
We are also, however, cautious about the findings, and their interpretation, and await careful comparison and analysis with other forthcoming (inter-agency sponsored) global estimates to develop our understanding of the reasons for differences between these and previous (WHO et al, 2007) figures. In the meantime we feel motivated to write a brief ‘blog’ articulating our concerns regarding these, and other, estimates of maternal mortality, and their use.
Specialising in the field of measurement of maternal mortality, the Immpact group has grappled with the inherent complexities and uncertainties of accurate measurement for over a decade, in many of the world’s poorest regions. Immpact is also acutely aware of the need for ‘good data’ to inform policy and planning, and to monitor progress towards wider health and development targets. It is in this sense that we make the following comments related to the recent Lancet article.
We note that the uncertainty intervals for the maternal mortality ratios (MMRs) published by Hogan et al don’t include uncertainty arising from the (likely considerable) uncertainties in estimates of adult female mortality, and in estimated fertilities, upon which the MMR estimates are based. Nevertheless, the uncertainty intervals in the Hogan et al estimates are quite wide, as should be expected given the relatively small number of deaths in most of the data-sets, and due to the fact that many countries have few or no data. Considerable further uncertainty also arises from the evident (but uncertain) biases in some or all of the different methods of measurement (e.g. sisterhood method, vital registration and so on) contributing data to the estimates. We feel that research into the causes, and sizes, of biases present in current methods will be critical in reducing the uncertainty in future estimates of maternal mortality.
Uncertainty in the measurement of maternal mortality is likely to be unavoidable, but uncertain information is better than no information. In terms of interpreting information, decision-makers need to be provided with the means to assess the (un)certainty of estimates. The development of good policy and planning decisions (i.e. those that will lead to improved outcomes) under uncertainty requires that the degree of uncertainty is known. Estimates without specified, and comprehensive, levels of uncertainty are not a good basis for decision-making, and are, in a sense, irresponsible in that they encourage readers to under-rate or even ignore uncertainty, and jump to unwarranted conclusions about differences or trends, quite possibly leading to poor decisions and potentially the sub-optimal allocation of scarce resources. Researchers have often not been meticulous in specifying uncertainty; there is a need for wider recognition of the importance of comprehensive descriptions of the uncertainty associated with estimates, and for the development of methods to make this feasible.
Hogan MC, Foreman KJ, Naghavi M, Ahn SY, Wang M, Makela SM et al. Maternal mortality for 181 countries, 1980-2008: a systematic analysis of progress towards Millennium Development Goal 5. The Lancet In Press, Corrected Proof.
World Health Organization, UNICEF, UNFPA, and the World Bank. Maternal mortality in 2005 Estimates developed by WHO, UNICEF and UNFPA. Geneva: World Health Organization, 2007