Alice Reid

Both my grandmothers lost children during or shortly after birth, and laid at least some of the blame on their care during that period. My maternal grandmother, a trained midwife, was worried about being allowed to go well beyond her due date less than two years after a previous caesarean birth. When she finally went into labour the doctor delayed his attendance because he was reluctant to leave his game of bridge, and the baby was stillborn. My paternal grandmother blamed a bombing raid for precipitating early labour, and her baby only lived three days. As a premature infant the baby would have been very vulnerable, but my grandmother felt she would have lived had the midwife not insisted on bathing her so frequently.  

These events – tragedies for both my grandmothers – took place during the early 1940s, and care before, during or after the birth may not have been to blame in both cases. Rates of infant and maternal mortality are far lower today, but there are still periodical concerns about standards of care in some maternity hospitals. This blog explores the development of birth attendance and its potential role in birth outcomes for both mothers and infants across the last few centuries in Britain.  

A previous blog outlined the trajectory of maternal mortality across history. The graph above shows maternal mortality ratios together with stillbirth rates and deaths of infants in the first week of life – all considered sensitive indicators of the quality of maternal care. All three series show long declines from around 1650 to 1800, little change between 1800 and 1940, followed by a period of sharp decline. This suggests that there were common influences on all three forms of mortality. This blog asks: how much did birth attendants and their practices affect these patterns? And if changes to midwifery practices did not drive the changes over time, to what extent did differential care affect maternal and infant outcomes in the past? 

Birth attendants across history 

Midwifery is an ancient profession, but this does not mean that midwives in Britain have had an easy ride, with struggles against men encroaching into the market, fights for professionalisation, and disputes over limits to their remit.  

It is likely that most ancient, medieval, and early modern midwives were women without formal training, but many – such as the English midwife Sarah Stone who practiced in the early 18th century – learned though apprenticeship to an experienced midwife. From around 1600 a growing number of ‘man-midwives’ or ‘accoucheurs’, often trained as barber-surgeons, also started to deliver births in the UK. Such men increased their credentials through qualifications, command of knowledge, and increasing access to one of medical science’s best kept secrets: the obstetric forceps.  

Forceps are a tool to hold an infant’s head and draw them out of the womb and are particularly helpful to assist difficult deliveries. In historic times they could make the difference between a stillbirth and a live birth, and they are still used in about 1 in 9 births in the UK today. Their reputed inventors, the Chamberlen brothers, kept them a secret for many years in order to corner the market on problematic births and to make their services attractive to wealthy families. Even when they were more widely known, the use of forceps was restricted to members of the barber–surgeon guild.  

Gradually a class of generalist doctors grew out of surgeons and apothecaries, and were licensed under the Apothecaries Act of 1815. For these self-employed ‘general practitioners’, as they became known, midwifery could be a key part of building up a clientele. But attendance at a birth could be a lengthy process, and it has been suggested that doctors had a tendency to quicken the delivery  through use of instruments such as forceps. But while forceps could be invaluable in a difficult delivery, they worsened outcomes for normal deliveries, carrying a risk of injury to mother and infant, and rendering the mother vulnerable to infection. 

The encroachment of men both into the practice of delivering infants and as the bearers of knowledge devalued the status of traditional female midwives, but it is unclear whether this caused a decline in the numbers of female midwives and the proportions of births delivered by them. Certainly female midwives did not disappear, and by the 19th century there were moves for their training and professionalisation. This period has been viewed a struggle between women and men for the control of the profession, and a key question was whether female midwives should be equal in status to doctors and allowed to use instruments, or whether they should be allowed only to manage uncomplicated deliveries.  

The 1902 Midwives Act eventually provided a resolution, enshrining midwives’ subservience to doctors. All new midwives had to undergo an approved training course (although established midwives were allowed to carry on without training if they could show themselves to be of good character). Midwives had to be formally certified and were bound by strict codes of practice, including summoning a doctor for a wide variety of issues at the birth, and they were not allowed to use instruments. Over time with regulation and the growth of subsidised local midwifery services, the involvement of doctors in routine deliveries dwindled, and the dominance of midwife delivery for normal labours – as is the case today – was established.  

Why did the training of midwives apparently have no effect on maternal mortality? 

In the graph above, maternal mortality does not seem to have been affected by the measures introduced in the 1902 midwives act. This is something of a conundrum, as in other countries (such as Sweden, Denmark and the Netherlands), the early and widespread introduction of training and professionalisation has been linked to improvements in birth outcomes.  

Work at Campop offers a solution to this paradox. Analysis of the birth outcomes (maternal mortality, stillbirths, and deaths in the first week of life) of trained and untrained midwives in Derbyshire shows that, in uncomplicated cases, trained midwives had better outcomes than untrained midwives, and midwives in general had better outcomes than doctors.  

Although this should have led to improving outcomes over time, there was a simultaneous increase in the number of calls for medical help, as midwives became more compliant with the new rules. Many births undoubtedly benefited from the presence of a doctor who, unlike midwives, was allowed to use instruments such as forceps to aid a difficult delivery. But the range of conditions which required a medical call was wide, and in some cases doctors may have interfered unnecessarily, increasing risks for both mothers and children and raising women’s vulnerability to infection. Doctors were not subject to the same strict cleanliness regimes as midwives and were in constant contact with a variety of infections in their other patients.  

Women delivered by Derbyshire doctors were 6-7 times more likely to develop puerperal fever than those delivered by midwives, and some of these women will have died. It is likely, then, that improving standards among midwives were counteracted by increasing levels of interference by doctors with decreasing experience in normal deliveries.  

The difference with Sweden, Denmark, and the Netherlands, was that midwives in those countries were allowed to take responsibility for complex births and trained in the use of instruments, so the effects of improvements in midwifery practice on mortality were not counter-balanced by the introduction of infection. In all countries, the development of antibiotics and sulphonamides in the 1940s was critical for birth outcomes, as it minimised the dangers of infection and made intervention in childbirth – including caesarean section as well as the use of instruments – much more possible. 

Geographical patterns in maternal mortality 

Another puzzle relating to maternal mortality is the geographic pattern within England and Wales before the second world war. As the map below shows, maternal mortality was considerably higher in the north and west of the country than in the south and east. The map shows data for 1914-16, but the same pattern had been visible since cause-specific mortality started to be routinely collected in 1837.  

Research at Campop has shown that a number of factors contributed to this pattern. Numbers of doctors and midwives per person were fairly similar across the country, but lower population density in the rural areas of the north and west meant that for many women there was no midwife within easy reach. In addition, in the north and west smaller percentages of midwives were trained, and the female population had lower levels of underlying health (specifically higher rates of tuberculosis), which has been shown to increase vulnerability to death during or following childbirth. 

This illustrates that quality of care is just one aspect, along with access to services and women’s underlying health, driving systematic differences in birth outcomes. 

Reassessing the decline in maternal mortality, 1650-1800 

The fact that both midwifery and underlying health were important drivers of geographic patterns suggests they could both also be reasons for the long decline in maternal mortality between 1650 and 1800 (although we should also note that influences on differences are not always drivers of change over time). 

The earlier blog on maternal mortality attributed higher maternal mortality in the peerage to the fact that peeresses were more likely (than the poor) to be tended by doctors, with higher rates of interference and infection. It also argued that improvements in midwifery practice were unlikely to have driven declines in maternal mortality over time. Instead, changes in the underlying health of women and – particularly the exposure to epidemic disease – were probably responsible for reductions in maternal mortality.  

It is certainly true that by the 18th century there were some highly skilled birth attendants, such as the man-midwife Dr Smellie and the highly experienced female midwife Mrs Stone, who both published volumes of case notes. Their management of labour included elements which would still be considered good practice today, and some authors have argued that if such practice spread gradually, it could have been responsible for the improvements in birth outcomes. However it is difficult to tell how accessible these teachings were, particularly for low-status female midwives. 

We can look at the evidence a different way: if obstetric practice was completely responsible for the decline in maternal mortality, then maternal mortality would have declined faster than non-maternal mortality. The mortality of fathers makes a good comparison here, and evidence shows that in fact these fell at the same rate. Whatever was causing maternal mortality to fall, also caused male mortality to fall, and rules out midwifery practices as a major influence on trends over time.

Instead, underlying health conditions, and particularly the infectious disease load, could explain both the rise in the 17th century and the subsequent fall. Bob Woods singled out smallpox as a particular potential culprit, but the geography of tuberculosis in England and Wales perhaps fits better with the geography of maternal mortality. 

Back to the present

There are still differences in birth outcomes between societal groups in Britain today. Black women are more than twice as likely to die during or following childbirth than white women, and those living in the most deprived areas are nearly twice as likely to die as those in the least. Underlying health status is still an important risk factor but cannot explain these differences. Access to care and the quality of care, including implicit bias, is today arguably the most important driver of inequalities in birth outcomes. 

How do we know?

See the previous blog for the measurement of maternal mortality from parish registers for the period 1538-1837. Parish registers provide estimates of mortality in the first week of life. After 1837 this information is derived from the reports and statistical reviews of the Registrar General, based on death registrations. These did not start to publish first week deaths until 1906. 

Stillbirths were not recorded in parish registers and must be estimated. The literature provides a number of different estimations and comparisons with countries with better data, and the estimation shown here is based on a combination of maternal mortality and mortality of infants in the first week of life. Although the civil recording of births and deaths started in 1837, stillbirths did not have to be recorded until 1937, which explains the gap in the stillbirth series. 

Birth outcomes in Derbyshire are derived from a rare set of records left by health visitors which included maternal mortality and stillbirths. They also included the name of the birth attendant, whose qualifications and status were obtained by looking them up in the Midwives Rolls or Medical Directory, as appropriate. 

The analysis of geographical patterns in maternal mortality used data provided by local Medical Officers of Health and published by the Local Government Board and a separate report for the Carnegie Trust. 

 Further reading 

• Løkke, A. ‘Mrs Stone and Dr Smellie: British eighteenth-century birth attendance and long-run levels and trends in maternal mortality discussed in a north European context’. Population Studies, 72:1 (2018), 123–136. 
• Reid, A., ‘Birth Attendants and Midwifery Practice in Early Twentieth-century Derbyshire’. Social History of Medicine, 25:2 (2012), 380–399. 
• Reid, A., ‘Mrs Killer and Dr Crook: Birth Attendants and Birth Outcomes in Early Twentieth-century Derbyshire’. Medical History, 56:4 (2012), 511-530. 
• Reid, A., ‘The geography of maternal mortality in early twentieth century England and Wales’. Annales de démographie historique, 139:1 (2020), 111-140. 
• Vousden, N., et al, ‘Impact of maternal risk factors on ethnic disparities in maternal mortality: a national population-based cohort study’. The Lancet Regional Health – Europe, 40 (2024), 100893. 
• Woods, R., and Galley, C., Mrs Stone & Dr Smellie: Eighteenth-Century Midwives and Their Patients (Liverpool University Press, 2014). 
• Woods, R., Death before Birth (Oxford University Press, 2009), pp. 189-237. 
• Wrigley, E. A., Davies, R. S., Oeppen, J. E., and Schofield, R. S., English Population History from Family Reconstitution (Cambridge University Press, 1997), pp. 307–22. 

Tags: birth attendants, childbirth, demography, doctors, infant mortality, maternal mortality, midwives, stillbirth

| This entry was posted on Thursday, March 27th, 2025 at 12:07 am and is filed under Uncategorized. You can follow any responses to this entry through the RSS 2.0 feed. You can skip to the end and leave a response. Pinging is currently not allowed.