How many women in Zimbabwe die as a result of pregnancy or childbirth complications? Two conflicting figures appeared in the last week.
First, the Zimbabwe Mail reported that 525 women die for every 100,000 live births in the country. Then a member of Zimbabwe’s parliament claimed the figure was as high as 960 deaths for every 100,000 live births.
“Fifteen women die every day in this country while giving birth,” the MP, Thokozani Khupe, said while calling for International Women’s Day on 8 March to be declared a national holiday.
So who is right?
99% of deaths in developing countries
Maternal deaths are almost entirely preventable and are therefore seen as a “sentinel indicator” of the quality of a country’s healthcare sector. This is illustrated by the estimate that 99% of all maternal deaths occur in developing countries.
Maternal mortality figures relate to all deaths caused by any stage of pregnancy or by childbirth complications, up to six weeks after delivery.
The cause of death could be direct (such as a haemorrhage) or indirect (when pregnancy aggravates an existing condition such as kidney problems), but accidental deaths are excluded. It is usually expressed as a ratio: the number of maternal deaths per 100,000 live births. Live births are used instead of the number of pregnancies in a country, which are difficult to determine.
As a United Nations member state, Zimbabwe is required to reduce maternal deaths by two-thirds from 1990 figures by the end of this year. The UN considers a maternal mortality ratio of less than 100 as low, between 100 and 299 as moderately low, and high when it is 300 to 499. Five hundred or more is very high, with a ratio above 1,000 deaths per 100,000 live births considered extremely high.
How are maternal deaths measured?
It would seem to be a relatively straightforward task to measure maternal mortality, but in reality that is not the case. Ideally, you would analyse death certificates, but even in countries with well-functioning birth and death registration systems the number of maternal deaths are routinely undercounted.
This is because death certificates are not always complete and in some cases the person signing a death certificate may not be aware that the woman was pregnant or that her pregnancy contributed in some way to her death. In some instances health facilities have been known to try and conceal maternal mortalities because of political pressure to reduce the numbers.
Very few African countries have accurate registration systems for recording births and deaths, says Dr Edward Nicol, a senior scientist at South Africa’s Medical Research Council. He specialises in the monitoring and evaluation of maternal health programmes.
“What we do is to use an indirect method during a household survey or census, where researchers would ask a female respondent which of her adult sisters are still alive or what a deceased sister had died of,” he explained.
But this method suffers from what is known as “recall bias” and does not always identify maternal deaths. According to Nicol: “Because the question spans a period of several years, the sibling may not be able to recall the exact cause of her sister’s death. Or you may ask someone if her sister died of maternal causes, and she could say yes, but then her sister had actually died of AIDS and she didn’t want to tell you because of the stigma.”
The sample for a household survey also needs to be large in order to provide accurate data. Statistically speaking, maternal deaths are “relatively rare events”.
Data gathered at clinics and hospitals is considered the least reliable measure for maternal death figures. With an estimated 20% to 35% of births in Zimbabwe occurring outside healthcare facilities, records of maternal deaths in hospitals and clinics don’t provide an accurate picture on their own.
Spike in the 2000s
A 2011 Demographic and Health Survey published in Zimbabwe suggests that there were 960 deaths per 100,000 live births for the period of 2003 to 2010. The 2012 census recorded a maternal mortality ratio of 525 per 100,000 live births for the preceding twelve months.
Zimbabwe’s most recent household survey, carried out in June 2014, reported 614 deaths per 100,000 live births for the seven year-period before the survey and 581 deaths per 100,000 live births for the preceding five years.
These figures therefore represent historical data, said Professor Robert Pattinson, director of the maternal and infant healthcare strategies research unit at South Africa’s Medical Research Council, and do not necessarily reflect the current situation in Zimbabwe. They are also not directly comparable because of the different time spans covered.
The World Health Organisation advises countries to focus on trends over three to five years. The WHO and other global organisations have also developed statistical modelling systems to help countries track maternal mortality over time.
Nicol recommends the estimates of the Maternal Mortality Estimation Inter-agency Group (MMEIG) which is made up of UN agencies such as the WHO and the UN Population Fund. Their latest calculations show that Zimbabwe’s maternal mortality ratio has increased from an estimated 520 deaths in 1990 to a peak of 740 deaths in 2005, before dropping to 470 in 2013.
The spike in deaths in the 2000s could be partly attributed to the HIV pandemic in sub-Saharan Africa, said Nicol.
Another body that calculates global maternal mortality ratios, the Institute of Health Metrics and Evaluation at the University of Washington in Seattle, reflects a similar trend for Zimbabwe. Confusingly though, the Institute estimated a much lower 1990 maternal mortality ratio than the UN for most countries because of different input data and methods.
The Institute of Health Metrics and Evaluation put Zimbabwe’s maternal mortality ratio in 1990 at 185.8 deaths per 100,000 live births, rising to 840.9 in 2003 and then dipping to 520.7 in 2013.
According to the Institute’s estimates, Zimbabwe falls in the category of only 16 countries worldwide that have a maternal mortality ratio above 500.
Conclusion: The claims are outdated
Because most pregnancy and childbirth related deaths are preventable, organisations such as the UN put pressure on countries to drastically reduce them. But monitoring maternal deaths is a difficult task, even in countries with an accurate birth and death registration system.
Zimbabwe is classified as a country with incomplete birth and death records by the UN. Researchers therefore rely on censuses and surveys to estimate maternal deaths. The most recent household survey reported a ratio of 614 deaths per 100,000 live births for the period between 2007 and 2014, and a figure of 581 for 2009 to 2014.
This is within range of global organisations’ estimates. Statistical modelling by the UN shows 470 deaths per 100,000 live births in Zimbabwe (this could range from 270 and 790 because of statistical uncertainty levels) whereas the Institute of Health Metrics and Evaluation estimates the ratio to be 520.7 (which could fall between 313.5 and 786.2).
The claims made this week in Zimbabwe were therefore based on outdated data.
Edited by Julian Rademeyer
Previous report Mmusi Maimane’s SONA 2015 response fact-checked
© Copyright Africa Check 2015. You may reproduce this report or content from it for the purpose of reporting and/or discussing news and current events, subject to providing a credit to "Africa Check a non-partisan organisation which promotes accuracy in public debate and the media. Twitter @AfricaCheck and www.africacheck.org".