Alice Reid
It is not unreasonable to believe that childbirth in the past was terribly dangerous. This view is common among popular history blogs and even some academic articles. Several internet sources, when discussing maternal mortality, state that in medieval or early modern times, as many as one in three women died during their childbearing years.
If one in three women died in childbirth it would indeed be a terrifying prospect, but although childbirth did come with substantially higher risk than today, childbirth was never the most common cause of mortality among adult women.
Between 2020 and 2022 in the UK, around 13 women died during or following pregnancy out of every 100,000 pregnant women, equivalent to a risk of .013 percent or 1.3 per 10,000. This measure is known as the ‘maternal mortality ratio’.
How much higher was this risk in the past?
The graph below shows that in the middle of the 17th century, at its highest level, maternal mortality was 170 per 10,000 (or 1.7 percent of women). Across the main pre-industrial period, average maternal mortality was 120 per 10,000 or 1.2 percent. During the 19th and early 20th centuries it was around 50 per 10,000 or 0.5 percent (although as explained below, this is likely to be an underestimate).
Even as far back as the medieval period, evidence from elite women suggests that maternal mortality was no higher than 1.2 percent.
The risk of childbirth-associated death was therefore far greater in the past than it is today, but does this mean that childbirth was terribly dangerous? It can be very difficult to assess what a particular risk expressed in percentage terms actually means (particularly when women might experience repeated childbirths and therefore repeated risks), so it can be helpful to think about this in different ways. One strategy is to look at the lifetime risk of maternal mortality, and another compares the risk of maternal mortality to the risks of other forms of mortality faced by women of childbearing age.
Lifetime risk of maternal mortality
The lifetime risk of maternal mortality takes account of the fact that many women have several children, and so face the risk of maternal mortality more than once. It is not a simple calculation, as it needs to take account of the risk of surviving the last birth and the chance of another pregnancy as well as the risk of dying in childbirth. In the overall period between 1550 and 1800, the lifetime risk for a married woman was about 5.6 percent, or one in 18 married women dying (because not every woman married, the risk of any woman dying in childbirth will be a bit lower).
For reference, this is just a little lower than the lifetime risk of maternal mortality in Sub-Saharan Africa in the year 2000 (one maternal death per 16 women), but a lot higher than that for Europe in 2000 (one in 2,400 women). One in 18 married women dying in childbirth may sound quite high, but it is the same as saying that 94 percent of married women did not die in childbirth: the vast majority of women survived, despite repeated exposure to the risk.
Comparisons with deaths from other causes
Alongside maternal mortality, women of childbearing age also faced the dangers of accidents and diseases, which were also much higher in the past than they are today. Over the pre-industrial period as a whole, the risk of a woman dying in childbirth, or the 6-8 weeks following it, was similar to the chance of dying from a non-maternal cause over the course of a year. Because women did not face the risk of childbirth every year, only around one in ten deaths to married women between the ages of 15 and 49 would have been due to childbirth. Childbirth was therefore far from the dominant cause of death, and women had far more reason to fear dying from a different cause.
Nevertheless, it is likely that women did approach childbirth with trepidation. The risk of dying from other causes was dependent on uncontrollable, random, and often sudden factors such as the occurrence of an accident or contracting an illness.
Once pregnant, however, the risk of childbirth was both inescapable and highly concentrated in time. Although comparable to other mortality risk over the course of a year, during the birth and the following few weeks a woman faced around six times her normal risk of death. A fear of death during childbirth is therefore not incompatible with the fact that the vast majority of women survived childbirth and that childbirth was a relatively minor form of mortality among adult women.
Trends and social differentials
The risks of childbirth were not static across time, or between social groups. Maternal mortality rose a little from the late 16th century, peaked in the mid-17th century, underwent a sustained fall to the early 19th century and then plateaued until another sharp fall in the mid-20th century. This last fall has been uncontroversially attributed to the introduction of sulphonamides and antibiotics in the 1940s, which allowed safer caesarean section and infection control. The earlier trends are less easy to explain, with some arguing for the quality of obstetric care and others favouring deteriorations and improvements in underlying maternal health as the key explanatory factor.
Given the importance of both the quality of care at the birth and the underlying health of the mother, it is surprising that most privileged members of British society, the peerage, had higher maternal mortality than the general population until the early 20th century. The explanation lies in the fact that the elite were more likely to have been attended by doctors in childbirth, who were more likely to actively ‘assist’ in the delivery, often unnecessarily and always with the risk of introducing infection.
However, the parallel reduction in risk between the elite and the wider population until 1837 suggests that changes in obstetric care were not responsible for trends in maternal mortality in this period (obstetric care, midwifery training and birth outcomes will be returned to in a future blog). Maternal mortality trends run parallel to trends in general mortality until the mid-19th century, and this indicates that changes over this period were driven by changes in the underlying health of women. It was likely that changes in women’s health was affected more by changes in exposure to epidemic disease (which was likely to have affected everyone) than by changes in the quality or quantity of dietary intake (which was likely to have affected social groups differently).
Why is it tempting to think that death in childbirth must have been higher?
There are several reasons why it is so tempting to think that death in childbirth must have been terribly high in the past. Some of these have already been hinted at above. Firstly, although maternal mortality was not a major killer, the risk was concentrated into relatively short space of time and was inescapable once a woman became pregnant. There were few other foreseeable events which carried such clear and focused danger.
Secondly, the heightened risk of elite women has also been highlighted. As such women are much better documented, examples of their maternal mortality can make it seem more ubiquitous than it was. In addition, the fact that since the end of the 19th century there has been a clear social gradient in most forms of mortality makes it tempting to think that maternal mortality among non-elite women in the past would have been even higher. The emergence of the social gradient in mortality will be addressed in a future blog.
Finally, literature is full of maternal mortality: this useful plot device removes a character and places a surviving child into a challenging environment which provides a good opportunity for an interesting narrative. It has been used in relation to both heroes and villains, from Dickens (Oliver Twist’s mother) to J.K. Rowling (Lord Voldemort’s mother).
In the 19th century the ubiquity of the device was already so well recognised that in her satirical novel Northanger Abbey, Jane Austen wrote of her main character: “No one who had ever seen Catherine Morland in her infancy would have supposed her born to be a heroine. Her situation in life, the character of her father and mother… were all equally against her… Her mother was a woman of useful plain sense, with a good temper, and, what is more remarkable, with a good constitution. She had three sons before Catherine was born; and instead of dying in bringing the latter into the world, as anybody might expect, she still lived on.”
Why is maternal mortality so hard to measure?
The WHO defines a maternal death as the death of a woman while pregnant or within 42 days of the end of a pregnancy (usually a birth), from causes directly related to pregnancy or childbirth (such as post-partum haemorrhage or puerperal fever) – termed direct maternal deaths, or from a cause which could have been aggravated by the pregnancy (such as tuberculosis or heart disease) – termed indirect maternal deaths, but excluding accidental or incidental causes of death. This means that measurements of maternal mortality depend on knowing which women were pregnant or recently delivered, as well as knowing the cause of death. Many death certificates now have tick-boxes to indicate a current or recent pregnancy, but this was not the case in the past.
For historic periods, two main strategies have been used to generate maternal mortality ratios.
For the pre-civil registration period, many maternal deaths can be identified by linking the baptisms of children with the burials of their mothers. However, this misses maternal deaths among women whose children died before baptism, or were stillborn. It also misses women who died while pregnant, without delivering a child.
Pioneering work was carried out at Campop by Roger Schofield to inflate observed deaths depending on the average time to baptism, the likely rates of stillbirth, and the risk of dying un-delivered, and to deflate for accidental deaths. His inflation and deflation factors were informed by detailed analysis of maternal mortality in Sweden, which had a similar demographic profile but much better data, and were subjected to careful testing. The results were reassuringly similar to maternal mortality levels and trends in a range of other European locations. Maternal mortality in both the wider population until 1837 and in the peerage was calculated using this method, by Roger Schofield and Jim Oeppen respectively.
For the period from the introduction of cause of death registration in England and Wales in 1837, and Scotland in 1855, the number of deaths are published by age, sex, and cause. The maternal mortality ratios for the period from 1850 onwards in the graph above and for the UK will be dominated by the larger population in England and Wales, where maternal mortality included direct causes of maternal death only. One might expect this method to yield an accurate assessment of maternal mortality, but work at Campop indicates that the true maternal mortality ratio in the UK between 1850 and 1930 could be up to twice the reported figure.
Alice Reid and Eilidh Garrett used a rare set of Scottish civil registers for the second half of the 19th century to compare maternal deaths identified by linking a birth and death of the mother, with maternal deaths identified through the cause written on the death certificate. The exercise revealed that up to 30 percent of maternal deaths could not be identified as maternal through the cause alone. Instead of writing ‘post-partum haemorrhage’ or ‘puerperal peritonitis’ doctors might just write ‘haemorrhage’ or ‘peritonitis’.
English and Welsh maternal mortality is also underestimated because it only included direct maternal causes and not indirect causes. The Scottish exercise reveals that the latter could form up to 40 percent of all mortality.
Hence the real maternal mortality ratio for England and Wales could be up to twice that reported. However, making a blanket adjustment is not easy, as the Scottish exercise showed that underestimation depended on the proportion of causes of death certified by doctors and underlying mortality. Paradoxically, causes of death offered by a next–of–kin could be more accurately allocated to maternal or non-maternal causes as they were often simply ‘childbed’, and underestimation was therefore greater where there were more doctors.
Urban areas, where background mortality was higher, had considerably higher indirect maternal mortality. Therefore, the extent that the official maternal mortality needs to be inflated will depend on medical provision, practices of medical recording, and on background mortality and these will have changed over time as well as varied between town and countryside.
On the other hand, the UK maternal mortality ratio uses the numbers of live births as a denominator rather than the total number of birth events, as stillbirths were not recorded until 1927. This is likely to inflate the ratio and to some extent balances the underestimation already discussed.
A couple of lessons can be learnt from this: firstly, levels of reported maternal mortality can vary with the levels of medical services due to reporting practices, and this can complicate comparisons over space or across time. Secondly, using maternal mortality as an indicator of the coverage and quality of health services is likely to be flawed.
Further reading
Galley, Chris, and Alice Reid. “Sources and Methods Maternal Mortality.” Local Population Studies 93.1 (2014): 68-78.
Loudon, Irvine (2000). “Maternal mortality in the past and its relevance to developing countries today”, The American Journal of Clinical Nutrition, 72(1): 241S-246S, https://doi.org/10.1093/ajcn/72.1.241S
Podd R (2020). “Reconsidering maternal mortality in medieval England: aristocratic Englishwomen, c. 1236–1503”. Continuity and Change 35(2):115-137. doi:10.1017/S0268416020000156
Reid, Alice and Eilidh Garrett (2018). “Medical provision and urban-rural differences in maternal mortality in late nineteenth century Scotland.” Social Science & Medicine 201: 35-43.
Schofield, Roger (1986) ‘Did mothers really die? Three centuries of maternal mortality in “the world we have lost’’’, in L. Bonfield, R. Smith and K. Wrightson (eds) The World We Have Gained: Histories of Population and Social Structure (Oxford): 231–60
Smith, Richard and Jim Oeppen (2006). “Place and status as determinants of infant mortality in England c. 1550-1837″, in E. Garrett, C. Galley, N. Shelton and R. Woods (eds) Infant Mortality: A Continuing Social Problem (Ashgate): 53-78.
Wrigley, E. A., R. S. Davies, J. E. Oeppen and R. S. Schofield (1997) English Population History from Family Reconstitution (CUP). pp. 307–21.