You might think that it’s easy to measure maternal mortality. You might think because it’s one of the Millennium Development Goals that countries have agreed what data should be collected to calculate it. You might think that there is a single equation which computes a final undisputed figure. And then everyone can get started on the tricky business of working out how to reduce it.You’d be wrong.
You might think that it’s easy to measure maternal mortality. You might think because it’s one of the Millennium Development Goals that countries have agreed what data should be collected to calculate it. You might think that there is a single equation which computes a final undisputed figure. And then everyone can get started on the tricky business of working out how to reduce it.You’d be wrong.It’s far from straightforward to measure. There are difficulties at every stage of data collection, analysis and even publication, as I learnt last week.I was in Addis Ababa in Ethiopia, on my first trip with E4A-MamaYe. I’m a little in awe of the size of the brains in the room.We are here to learn about estimating maternal mortality from census data. Government statisticians and a few other people who love numbers have gathered at this workshop to receive training, share experiences and debate the best practices for moving forward.E4A-MamaYe is one of the sponsors along with USAID, MEASURE Evaluation, UNFPA, WHO, UNICEF and ACIPH. We are joined by 45 participants from Ethiopia, Tanzania, Ghana, Zambia, Rwanda, Liberia, Kenya, and Nepal.Closely listening to the expertsHaving reliable ways of calculating maternal mortality is enormously important. Without it, we cannot reliably take action to tackle it, reliably know if governments are successfully implementing maternal health policies or reliably allocate resources based on evidence.As Dr Kavita Singh Ongechi, Senior Technical Advisor for Maternal and Child Health for the MEASURE Evaluation project says:“The UN Declaration of Human Rights indicates that ‘Motherhood and childhood are entitled to special care and assistance", yet we know that millions of women do not have adequate access to maternal health services. A key step in addressing the problem is empowering governments to estimate maternal mortality as accurately as possible. With accurate estimates, a government can understand the magnitude of maternal mortality within their country and also within regions within their country. This information is essential for program planning and for developing appropriate policies and interventions.” Brains working hardThe experts leading the workshop were keen to promote open access to census data and encourage countries to share their data as well as tabulated results – something which countries like Ghana and Ethiopia are doing with their online microdata systems (see below).There’s a global movement toward opening up this data and allowing anyone with the skills to analyse it for themselves. This kind of transparency and access can only be a good thing for increasing knowledge and ultimately helping us to put an end to maternal deaths.We had been asked to bring along the latest census data from our own countries to analyse. Our Tanzanian participants were particularly excited as their data had not yet been used to calculate their maternal mortality ratio (MMR). We all waited eagerly for the preliminary results.We pondered about what the figures meant to each country. Better or worse than expected, trends, comparisons with other data sources.We then moved on to discuss how to improve the next round of censuses so that it could provide a better basis on which to measure maternal mortality. For example:Whether countries should include a cause of death question in the census? The conclusion was ‘probably not’ as it can give misleading results where deaths weren’t medically certified.Whether to recommend countries to follow up a census with a verbal autopsy as best practice. This would allow the stats people to separate out the women who died of causes relating to their pregnancy from those who died when pregnant from unrelated reasons. While everyone agreed that this was what we want to know, the debate was about whether it should be prioritised given the limited resources available.Here is a summary of the findings from some of the MamaYe countries:EthiopiaEthiopia conducted its third census in 2007. Their Central Statistics Agency has worked hard to make the entire dataset available on its website. This means their data are open and transparent. Anyone interested can interrogate their data and produce tables of indicators. Fascinating reading for data geeks.TanzaniaIn 2012 Tanzania conducted its fifth census since the Union of Tanganyika and Zanzibar in 1964. Some data from the 2012 census is available online, but analysis of the maternal mortality ratio from this data has not yet been published. This workshop was the first time that a revised maternal mortality ratio (MMR) had been estimated from this data source.Tanzania has had five Demographic and Health Surveys (DHS), the most recent in 2010. The National Bureau of Statistics is currently preparing for the next survey in 2014/2015, and plan to include the module that allows calculation of the MMR.GhanaGhana’s most recent census was in 2010. Ghana Statistical Service is in the process of making their entire census data available to users on the internet. A full report of the census can be read here including the MMR estimation.The Ghana Statistical Service is currently planning for the next DHS in 2014 and they will be including the maternal mortality module for the first time.