Estimating childbirth deaths in prehistory

The role – and fate – of women in prehistory is set for renewed focus and better understanding following the development of a seemingly robust method of calculating rates of maternal mortality in ancient times. Today, death before, during, or soon after childbirth is a significant cause of female mortality, especially in countries and communities without universal access to healthcare.

The World Health Organisation estimates that maternal mortality worldwide has dropped by 43% since the year 2000. It remains at 239 deaths per 100,000 people in developing countries. In the developed world, it hovers around 12.

Simple logic suggests that death during childbirth must have been common in ancient communities, in which healthcare, if it existed at all, was basic and lacking understanding of the causes of infection.

However, deriving any indicative estimates has been extremely challenging.

“Previous attempts to understand the maternal experience in past populations have largely focused on burials where females appear to have died during pregnancy, childbirth, or shortly after, though occurrences are relatively rare,” write researchers Clare McFadden and Marc Oxenham, of the School of Archaeology and Anthropology, Australian National University.

Given their rarity, ancient remains of women who died during childbirth tend to gather a lot of attention in anthropological circles. Oxenham, for instance, was co-author of a celebrated study on the remains of a young female and unborn full-term breech foetus found in Vietnam in 2011. The remains were dated to between 2100 and 1050BCE.

The study revealed the woman to have been in remarkably poor health, which may have contributed to her death, and that of her child. Whatever the full story, however, researchers in the field have always acknowledged that so few relevant burials have been discovered that they cannot collectively serve as reliable indicators of death rates.

Despite this difficulty, the fact remains, Oxenham and McFadden argue in the journal Current Anthropology, that knowing the maternal mortality rate of any ancient population is critical to understanding group dynamic and direction.

The number of deaths associated with childbirth, they say, may indicate differences in health and maternal care.

“Conversely,” they add, “maternal mortality may impact population dynamics and thus drive change in society and culture.”

To establish a working model to make a reasonable calculation of the mortality rate, therefore, the researchers turned to much more recent statistics. They focussed on known maternal death numbers, collated by the United Nations, for 76 countries for the years 1990 to 1995.

The choice of time period was important. The drop in death rates since 2000 is the result of a major World Health Organisation drive – a Millennium Development Goal. Figures before its instigation are more likely to be similar to those existing in various early populations.

To test the predictive power of their model, Oxenham and McFadden used data detailing all deaths among men and women aged between 20 and 24, and compared them to the known maternal mortality figures.

The age range was carefully determined, for a couple of reasons. First, women aged between 15 and 19, or above 30, face increased risks of death during childbirth and so, the researchers say, “have the potential to overestimate the average maternal mortality risk for a population if used in isolation”.

Second, the range inherently recognises that often exact age dates for ancient burials are impossible. Any given cohort of prehistoric female remains estimated to be between 20 and 24 will inevitably include outliers above and below that range.

After applying a number of limiting analyses to the raw difference between male and female deaths, the researchers reached a conclusion they determined to be the number of deaths associated with childbirth. Comparing the result with the known maternal mortality rates showed a clear and robust agreement.

Even better, the model also stood up if the age range was increased to between 20 and 30.

The model thus appears to be viable for estimating maternal mortality rates for any population in which age-at-death proportions can be reasonably estimated – meaning that a heretofore significant but invisible cohort of women who died young can now, in the abstract at least, be recognised.

Further research in the field, they note, might now provide “significant insights into the causes of maternal mortality, sophistication of maternal care, and female experience of pregnancy and childbirth in past populations”.