Maternal Mortality, Dublin, 1864–1902

Maternal Mortality, Dublin, 1864–1902 Summary Using civil registration as a prism this article examines the complicated interface between the authorities, medical professionals, women in childbirth and the aftermath of delivery. It argues that Irish maternal mortality was underestimated and provides a number of arguments in relation to the complex socio-medical environment to explain why. Our research shows how a combination of cavalier attitudes towards cause of death classification and liberal interpretations of William Farr’s Statistical Nosology, served to obscure the true extent of maternal mortality in Dublin City from 1864 to 1902. By offering a microhistory of maternal mortality reporting and registration this article problematises the merits of using civil registration data as a resource for the social history of medicine. maternal mortality, gender, Ireland, civil registration Introduction Late nineteenth-century Dublin City held the unenviable record of having one of the United Kingdom’s worst urban mortality rates. The average death rate of 29.5 per thousand in the ten years from 1890 to 1899 was far in excess of London, and ‘the thirty-three large towns of England and Wales’; it was so appalling that it prompted a commission of inquiry in 1900.1 Annual returns to parliament included analyses of age cohorts and cause of death, but, in an era that was characterized by worryingly high mortality in the 5–35 age groups (from tuberculosis and pulmonary diseases), deaths during pregnancy, in childbirth or in the post-natal context did not receive much focused political attention.2 Vincent De Brouwere found in his comparative analysis of the historiography of European maternal mortality that maternal death was too narrowly defined. In addition he argued that because historic maternal mortality rates (MMR) omit deaths caused indirectly by childbirth and late maternal deaths (occurring after 42 days), the official returns are inaccurate.3 Using late-nineteenth-century metropolitan Dublin as a case study, this research aims to examine if underreporting of vital events occurred and to demonstrate whether, as a phenomenon, ‘hidden maternal mortality’ was as inherent to Irish civil registers as, for instance, Rebecca Kippen found in Tasmania.4 Apart from Cormac Ó Gráda and Lindsey Earner-Byrne’s work, scholars of Ireland have largely ignored the issues of maternal and infant mortality.5 This is partly due to the fact that, despite various government commissions on the health of nations, maternal health only began to emerge as a discrete medical issue in Ireland in the late 1920s.6 Historically, maternal mortality has been defined by its relationship to live births, which is why, with few exceptions it is usually studied in conjunction with infant mortality. Irvine Loudon has argued that analysing infant and maternal mortality together can provide insightful contrasts.7 But, given the nascent state of Irish scholarship in the respective fields, we posit that more useful insights can be garnered by separating them.8 This article is primarily concerned with mortality reporting under the civil registration system, which was introduced fully in 1864. As distinct from Loudon’s methods, which relied on national statistics, a combination of macro- and microhistory approaches are used to problematise the data presented in official Irish mortality returns.9 For example, an outline of overall mortality in Dublin from 1864 to 1901 is given to contextualise maternal mortality over that period, and is followed by a micro-study of the death registers in three pre-census years to assess the accuracy of the returns.10 Using these techniques we identified peculiarities in the registration process and patterns in the localisation of maternal death. In the data we use here 74 per cent of the maternal deaths were certified; all occurred in hospital settings shortly after delivery (mostly within one week, and rarely above one month), which raises questions, as De Brouwere cautions, about those occurring outside institutions, those attributed to indirect (underlying causes) and late maternal mortality. We trace these peculiarities to Ireland’s unique set of political and socioeconomic circumstances, which made the implementation of census taking and civil registration (both crucial to effective governance) very challenging.11 There are several reasons for this that range in origin from class and creed to culture. Historically the Irish people had a fractious relationship with British administration in all its guises and therefore studies of official census and civil registration returns should acknowledge that these data are deficient. Religion, power, social class and gender are critical issues in the timeframe we examine: the majority of the population (roughly 75 per cent nationally, but higher again for Dublin) was Roman Catholic and poor but men of Protestant faiths (primarily Protestant Episcopalians and Presbyterians) were the main powerbrokers.12 That civil registration was placed under the auspices of a public medical system, which was grappling with professionalisation, specialisation and later modernisation, did not aid its progress.13 The act stipulated that deaths should be medically certified, but the legislation paid little heed to the importance of accurate cause of death reporting.14 Furthermore, the matters of uncertified deaths, unspecified cause of death and imprecise categorisation of cause combined to reduce the accuracy of the records. The remainder of this article is divided into four sections. In Section I we outline our source material and explain our methodology. We discuss the merits of using civil registration data for medical history research and contend that some of the problems arising from the data emanate from the architecture of the system itself. While there are instances of poor compliancy with civil registration across the United Kingdom, casual observance of the law was more pronounced in Ireland where mortality under-registration persisted into the 1990s.15 A series of studies conducted from the 1970s to the 1990s advocated that undertakers form part of the solution to the problem of underreporting in the West of Ireland.16 To provide an understanding of why the Irish were reluctant to engage with certain aspects of civil registration, Section II outlines the foundation of the system and contextualises its operation within broader Irish administration. The General Register Office (GRO) was crudely grafted onto the Irish Poor Law in 1864, which created scope for general mistrust and engendered a fear of surveillance. We contend that the pre-history of the poor law, and its strong associations with the Great Famine, naturally inhibited the success of the GRO: it entrenched the power of the former system and was a retrograde step. In Section III we use two GRO data types—aggregate annual returns to parliament and individual register entries—to isolate overall mortality figures from 1864 to 1913 in order to contextualise the MMR. Section IV endeavours to give a better understanding of the nature of maternal mortality for metropolitan Dublin by examining a variety of available data. In the first instance, we extracted entries of maternal deaths from the 1870, 1880 and 1890 registers and mapped the data to see if local factors such as professional maternity services had an impact on the rate. Unsurprisingly maternal deaths were localised to maternity hospitals but the geographical dispersion shows a curious absence of deaths in workhouse hospitals, which operated maternity services. The city also had several ad hoc domiciliary nursing services, so the clustering of the majority of deaths near maternal hospitals raised our suspicions about underreporting elsewhere. The history of hidden maternal deaths is methodologically difficult to unravel and this article aims to show potential avenues for future research. To that end our concluding section returns to the matter of cause of death catergorisation and traces how international discourses on disease classification only gradually caused an improvement in cause of maternal death reporting in Ireland. Overall we show that deaths reported by health care professionals are problematic up to 1902. I. Methodology: Problematising the Sources In this section we focus on registered maternal deaths and how they were defined and categorised to show how the published aggregate data are fundamentally inaccurate. For the purposes of the 1863 act the Registrar General supplied a reprint of William Farr’s Statistical Nosology of Causes of Death to all registrars and registered medical practitioners in Ireland to standardise civil-registration entries, and achieve ‘uniformity in the registry of the causes of death’.17 The GRO in England first published Farr’s Nosology in 1845, a few years after civil registration had commenced there. In the Irish edition, the preface emphasised that the process was not intended to be doctrinaire and the nosology was pitched as ‘suggestions … to medical practitioners’.18 Unlike their Scottish counterparts, who mounted a sustained albeit unsuccessful campaign against the adoption of the ‘flawed English schedule’, Irish medical practitioners did not respond so vociferously.19 Instead, with respect to maternal death, they later used Farr’s guidelines to make a distinction between death in but not of childbirth.20 Under the English system non-medical personnel, such as undertakers, could issue death certificates.21 Opposition from the Fellows of the Royal College of Physicians of Edinburgh gave rise to the inclusion of a system of fines for uncertified deaths under its civil registration law. Irish physicians benefited from the Scottish advances and similar clauses were included in the 1863 Irish act, but only in cases where the deceased had sought medical counsel prior to death.22 World Health Organisation (WHO) defines a death occurring ‘within 42 days of the termination of pregnancy’, irrespective of duration, site or outcome, as a direct maternal death.23 Historically it was largely understood to mean a death that occurred within six weeks of the commencement of labour, following a successful pregnancy. Farr’s Statistical Nosology listed 145 causes of death.24 It made some allowances for arcane medical terminology through the provision of ‘synonymes’, which could be used ‘at the discretion of the medical informants’.25 Obstetrics was a nascent field in the 1860s and even recognised experts such as the Scottish physician, Dr James Matthews Duncan, used loose terminology of ‘say four weeks’ post-partum to define a maternal death.26 Childbirth was increasingly ‘medicalised’ in the timeframe we examine, but deaths in women who did not go to full term could technically be classified under categories such as fever or other underlying health conditions.27 Despite the fact that doctors were encouraged to note ‘pregnancy’ where applicable in cause of death recording, it was not a legal requirement, so maternal morbidity was poorly captured in the Irish records. The same provisions formed part of the English registration act and troubled contributors to the British Medical Journal (BMJ).28 Another factor that obscures the full extent of maternal mortality and morbidity in the United Kingdom is that stillbirths were not registered. Ireland was in the grip of a tuberculosis epidemic in the late nineteenth century and Greta Jones draws attention to the physiological effects of childbearing in aggravating latent cases. She also speculates (although it is difficult to quantify precise mortality) that tuberculosis was ‘implicated in stillbirths and in perinatal deaths in children’.29 As early as 1892 a House of Commons select committee heard that ‘every country in Europe, except Russia, registers still-born children’.30 Despite further criticism at the Brussels Congress on Hygiene in 1903 regarding the absence of stillbirth registration in the United Kingdom, provision was not made until 1926 for England and Wales and 1938 for Scotland.31 Ireland’s lag is quite remarkable: a national system for registering stillbirths was not introduced until 1994.32 The absence of stillbirths from the historical record means that an accurate study of Irish fertility and antenatal health from GRO records over a longue durée is impossible. From the outset we recognised that most deaths were recorded because they occurred in maternity hospitals but that there was a strong possibility even these were underestimated. Dr McClintock, Master of the Rotunda Lying-In Hospital, north of the river Liffey (founded in 1745) between 1854 and 1861, summarised how doctors understood their duties at a meeting of the Dublin Obstetrical Society in 1869.33 … if the cause of death is known to be puerperal fever—or anything pertaining thereto—quite a panic is created in the neighbourhood, and both doctor and nurse come in for more than their fair share of blame. Hence for their own sake, as well as the charitable motive of not alarming all the pregnant women in the community, the death is imputed to any other cause rather than the dreaded puerperal. … The defect lies in our system of registration—not in those who supply the returns. Practitioners make very proper distinction between dying in childbirth, and dying of childbirth. When a woman happens to die in childbed of some intercurrent disease—as phthisis, pneumonia, dysentery, apoplexy, albuminuria, bronchitis, morbis cordis, &c—this alone is returned and rightly so—to the registrar as cause of death. Consequently all these deaths have no place in the registration reports of deaths in childbed. (Emphasis in original)34 This was penned in response to an open letter written by the former Master of the Rotunda, Dr Evory Kennedy, on the structural failures of the hospital to cope with puerperal fever outbreaks.35 McClintock infers that even lying-in hospitals, could manipulate their cause of death statistics if an underlying disease was present, which points to a major weakness in these data. Given the way in which maternal death was defined and classified, a central issue for our research was whether or not lying-in hospitals might be guilty of manipulating their statistics, by attributing maternal deaths to other causes. Therefore, where extant, we examined their records to determine the accuracy of the published data. The two principal maternity hospitals in Dublin were the Rotunda and the Coombe Lying-In Hospital (founded in 1826), in the south west of the city.36 Both hospitals were charitable institutions and catered primarily for Dublin’s poor. Maternal mortality statistics from the Rotunda and the Coombe were published from 1858 in the Board of Superintendence of Dublin Hospitals (BSDH) annual reports.37 We used these in tandem with the Rotunda hospital in-patient registers (registers for the Coombe hospital pre-1914 are not extant) and the GRO death entries to enable us to ascertain how these records related to one another and to see if there were any discrepancies. As it was clear that we were dealing with deficient data we employed a second process in our examination of the GRO registers, where we identified other possible direct maternal deaths of women aged between 14 and 54 and used birth information and other sources to test our suspicions. Apart from being labour intensive, we encountered many difficulties. First, maternal mortality cases are not clear-cut. Obviously childbirth-related diseases such as ‘puerperal fever’ present no ambiguity, but other conditions such as peritonitis (a potential outcome in the final stages of puerperal fever) or phthisis do.38 The second difficulty is that the verification of a suspected pregnancy case depends on whether a live birth ensued and if it was subsequently registered. In one instance we identified an additional maternal death in 1880 caused by ‘unavoidable haemorrhage and pleurisy’ because of the accidental entry of a stillbirth in the GRO birth record that was subsequently struck out by the registrar when the error was realised.39 Uncertified deaths It is important to stress that death notification underwent an iterative process of translation, which increased the potential for inaccuracy and allowed scope for an interpretation of the cause of each death. The duty of recording maternal deaths occurring in hospital settings fell to the medical attendant, who was responsible for categorising the precise cause of death.40 While Irish law was clear on the registration procedure, the method followed by registrars is difficult to determine. A GRO death entry required information on whether it was certified or not, but there was no requirement to record who certified it, thereby reducing accountability. The Annual Reports of the Registrar General (ARRG) began to acknowledge the problem of uncertified deaths in 1906, when it was estimated that 24.4 per cent of all registered deaths were uncertified, and in rural areas that percentage was much higher.41 It was not until 1961 that the proportion of uncertified deaths nationally fell below 5 per cent.42 Uncertified deaths occurring in domestic settings are even more challenging because descriptions of causes of death were filtered through ordinary individuals, often family members with no medical expertise, or through midwives (qualified and unqualified) with varying degrees of medico-legal knowledge. Efforts to standardise disease classification were frustrated by uncertified deaths on which Thomas Wrigley Grimshaw (Registrar General, 1879–1900) commented that the cause of death ‘is a mere statement … it may all be incorrect’.43 In uncertified cases, the circumstances of a death were coloured in the first instance by vernacular accounts of cause of death, which in turn was complicated by the way in which registrars moderated the information they received. It was then returned to central administration for transcription and collation, and each stage created multiple opportunities for error or inaccuracy. An example of how the process of moderating cause of death data can distort the eventual returns is shown in Stephan Curtis’ analysis of Sundsvall in Sweden. His study of 213 direct maternal deaths among women aged 15–45 notes a curious absence of deaths from haemorrhage and suggests that they might have been classified under ‘difficult birth’ in the parish registers.44 Deficiencies in the data troubled nineteenth-century observers who invariably voiced their concerns in medical journals. For instance, Dr Francis Vacher, medical officer of health for Birkenhead, in an 1887 address to the Sanitary Institute of Great Britain, bemoaned the poor quality of cause of death data in the English mortality returns.45 If the death registration system was found to be flawed in England then post-famine Ireland was an even greater challenge, which Section II explains in greater detail. II. The Origins of Irish Civil Registration Edward Higgs has argued that the English GRO is erroneously perceived in binary terms, as a secular vital-registration agency on one hand, or as an organ of the public health movement on the other. He reminds us that Thomas Cromwell’s original 1538 parish registration act was a secular endeavour that aimed to regulate property rights.46 No such property regulations were in place in Ireland and civil registration most certainly became an important instrument of the public health movement. The process of introducing full civil registration was delayed primarily for socio-political reasons, but it was also hampered by religious differences, which had posed problems for the English authorities in Ireland for centuries. For example, there was strong resistance to the early Irish censuses in 1813–15 (which was incomplete) and again in 1821. Commenting on prevailing attitudes during the 1821 census, James Doyle, Roman Catholic bishop of Kildare and Leighlin, observed that ‘The Catholics have ever been unwilling to make known their numbers to any agent of the Government. Having too often experienced from it what they deemed treachery or injustice …’47 In spite of the difficulties, E. Margaret Crawford describes the 1821 census as ‘a great improvement on its predecessor’, not least because it was the first all-Ireland estimate of Irish population to see completion.48 Thereafter, national censuses were held decennially until 1911. Civil registration commenced in the United Kingdom in 1837, with the passage of two statutes in 1836, which applied to England and Wales only.49 In Ireland a first step towards civil registration was taken with the passage of the 1844 marriage registration act which stipulated that from 1 April 1845 all non-Catholic Irish marriages were to be registered by the civil authorities.50 Following a number of failed attempts, Scotland introduced full civil registration in 1855 leaving Ireland as the only part of the United Kingdom without a system of birth and death registration.51 Under the 1844 marriage registration act the Lord Lieutenant of Ireland was authorised to establish the GRO in Dublin and appoint a Registrar General of Marriages.52 The act provided for the division of the country into ‘districts’, with a register office and a district registrar, established in each.53 Initially 130 civil registration districts were created. These were coterminous with the Poor Law Unions (PLU) but employed a discrete staff. William Donnelly was appointed Registrar General and he proposed that the administration of civil registration and the poor law be kept separate; the Lord Lieutenant concurred.54 Gradually the GRO came to absorb other functions; it became responsible for carrying out the decennial census in 1851 until 1911, staffing was remarkably consistent and of excellent calibre.55 With respect to Irish civil registration five bills were brought before the House of Commons in 1846, 1859, 1860, 1861 and 1862, but all failed to pass.56 The 1846 proposal, introduced by Morgan John O’Connell and Benjamin Chapman was structurally unsound, whereas the Bills introduced between 1859 and 1862 foundered because of political differences between the parties as to whether the registration process should be tethered to the policing or the Poor Law systems; the Conservatives favoured the former, and the Liberals the latter. Financial constraints led to a Liberal victory, as the poor law option was cheaper.57 In 1863 the matter was finally resolved with the passage of two acts, one to register births and deaths, and the second to register Catholic marriages. The Act for the Registration of Births and Deaths in Ireland came into force on 1 January 1864. William Donnelly was assigned the position of Registrar General of Births and Deaths, which he held until his resignation in 1876.58 Like the 1844 act, the registration process for births and deaths was based on the PLUs, with each Union being deemed a ‘Superintendent Registrar’s District’ (SRD).59 Each SRD had a register office funded out of PLU funds, and a Superintendent Registrar, who was usually the Clerk of the Union. Each SRD was subdivided into ‘Registrar’s Districts’ (RD), with a registrar responsible for registering births and deaths.60 The Poor Law dispensary system had been established under the 1851 Medical Charities Act, and provided the poor with free medical care.61 Cassells argues that the 1851 act conferred to the Poor Law Commission a level of power over public health and medical affairs that was ‘unprecedented in Ireland and unparalleled in the United Kingdom’.62 It created 723 Dispensary Districts from the then 163 PLUs, each was under the control of a dispensary doctor who was answerable to a management committee.63 When registration commenced, dispensary doctors became responsible for registering births and deaths in their areas.64 Payment for duties such as civil registration and successful vaccinations were undoubtedly a welcome addition to the basic salary of a dispensary doctor, which was determined annually by the Board of Guardians and paid out of the poor rate.65 Civil registration placed a great deal of responsibility on individuals to develop a medico-legal awareness. In the case of a birth, the law tasked parents with informing the registrar. If they were unable to do so, responsibility then devolved to the birth attendant or occupiers of the same dwelling. Births were to be notified within twenty-one days.66 Similar to the procedure for a birth, a death was to be reported to the registrar by somebody present at the event or occupiers of the same dwelling. In normal circumstances deaths were to be notified within seven days, and an informant could be summoned to sign the register within the next fourteen days.67 To incentivise prompt, accurate reporting, a complicated system of fines was introduced for registrars and informants alike.68 Geary argues that despite establishment claims of Ireland being better served medically than any other country, the poor did not share the ‘enthusiasm for the workhouse hospitals’.69 It seems they were also reluctant to engage with other services associated with the PLU system, Thomas Wrigley Grimshaw in 1879 conceded that 9 per cent of deaths in Dublin in the previous decade ‘escaped registration’.70 It is to the matter of Dublin city’s MMR that our attention now turns. III. Dublin mortality rates in context Given the emphasis that the 1863 act placed on administrative boundaries, this section begins by tracing how it was applied to the Dublin metropolitan area. It proceeds by outlining national mortality rates between 1864 and 1913, in order to contextualise our micro-study of maternal mortality in the chosen sample pre-census years of 1870, 1880 and 1890. County Dublin was covered by seven SRDs, and two of these, Dublin North and Dublin South, encompassed all of Dublin city, and its rural hinterland (Figure 1). Dublin North SRD was divided into nine RDs, three (and later four), of which were in the city. Dublin South SRD comprised ten RDs, including four city divisions.71 Figure 1 View largeDownload slide Superintendent registrars’ districts in County Dublin Source: Author’s own. Figure 1 View largeDownload slide Superintendent registrars’ districts in County Dublin Source: Author’s own. Civil registration data for large urban centres present specific research challenges: the birth and death rates of RDs containing public institutions, such as maternity hospitals and workhouses, were artificially inflated by non-residents and therefore do not reflect natural population change within the RD. Ireland did not have laws of settlement, which would have caused demographic readjustment with the official return of strangers to their parishes of origin.72 The ARRGs report on the number of births and deaths per year at various administrative levels, including to the level of RD. The pattern for deaths in Dublin varied from the national picture (Figures 2 and 3). While between 1871 and 1911 the national population declined at each successive census, Dublin’s population increased. The increase can be partially explained by an expansion in the city’s borough boundaries by the 1898 Local Government (Ireland) Act but that change had no impact on the city’s RDs.73 In the thirty-four years between 1864 and 1901 a mean of 7,470 deaths per year were recorded in the two urban SRDs (Figure 3). Figure 2 View largeDownload slide Number of deaths recorded in Ireland, 1864–1901 Note: Horizontal line shows the average number of deaths per year (89,582) in Ireland, 1864–1901. Source: Vaughan and Fitzpatrick, Irish Historical Statistics, Population, 247–8. Figure 2 View largeDownload slide Number of deaths recorded in Ireland, 1864–1901 Note: Horizontal line shows the average number of deaths per year (89,582) in Ireland, 1864–1901. Source: Vaughan and Fitzpatrick, Irish Historical Statistics, Population, 247–8. Figure 3 View largeDownload slide Number of deaths recorded in Dublin city, 1864–1901 Note: Horizontal line shows the average number of deaths per year (7,470) in Dublin city, 1864–1901. Source: ARRGs, 1864–1901 (see footnote 76 for complete references). Figure 3 View largeDownload slide Number of deaths recorded in Dublin city, 1864–1901 Note: Horizontal line shows the average number of deaths per year (7,470) in Dublin city, 1864–1901. Source: ARRGs, 1864–1901 (see footnote 76 for complete references). A notable peak in deaths occurred in the three-year period 1878–1880, and in the latter two years the national death total exceeded 100,000, the only occasions when recorded deaths in a calendar year breached that benchmark. The 105,089 deaths registered in 1879 was the highest number of deaths recorded in the period we examine.74 Although these years coincided with the most serious agricultural crisis since the 1840s, the ARRG enthusiastically observed that the increase was not solely due to increased mortality; instead it cited increased engagement with civil registration and better accounting for burials under the Public Health (Ireland) Acts, 1878–1879.75 Similarly the outstanding feature of Dublin city’s death records (Figure 3) is the mortality peak in 1879 and 1880. We discuss the ramifications for maternal mortality further in the next section but in both years the mortality level exceeded the city’s mean level in the period 1864–1901 (7,470 deaths) by between 25 and 30 per cent. In two other years (1887 and 1899) the number of deaths recorded exceeded the mean level (1868–1901) by more than 10 per cent. The 9,653 deaths recorded in the city’s RDs in 1880 formed a peak in mortality in Dublin, which was not exceeded within our timeframe.76 The ARRGs only reported on deaths that were recorded by the various registrars, but, since not all deaths were registered, the published figures were deficient. The GRO was not blind to this and indeed noted that the number of burials in Dublin’s cemeteries and graveyards exceeded the number of registered deaths of city residents by almost 10 per cent during the first seven years of registration (see Table 1).77 Table 1 Dublin city burials and death registrations, 1864–1870   1864  1865  1866  1867  1868  1869  1870  1864–1870  Burials  7,116  7,536  8,338  8,493  7,248  7,237  7,352  53,320  Reg. deaths  6,260  6,959  7,571  7,374  6,804  6,557  6,625  48,150  Excess burials  856  577  767  1,119  444  680  727  5,170  % discrepancy  12.0%  7.7%  9.2%  13.2%  6.1%  9.4%  9.9%  9.7%    1864  1865  1866  1867  1868  1869  1870  1864–1870  Burials  7,116  7,536  8,338  8,493  7,248  7,237  7,352  53,320  Reg. deaths  6,260  6,959  7,571  7,374  6,804  6,557  6,625  48,150  Excess burials  856  577  767  1,119  444  680  727  5,170  % discrepancy  12.0%  7.7%  9.2%  13.2%  6.1%  9.4%  9.9%  9.7%  Source: Census of Ireland, 1871, pt ii, vital statistics, vol. ii, report and tables relating to deaths, ci. Conformity with death registration law appears to have been particularly bad during two of these seven years, 1864 and 1867, and the suggested reasons are illustrative. Understandably, deficiencies were evident for 1864, the first year of full civil registration, but the significant divergence between burials and registered deaths in 1867 was attributed to an outbreak of cholera in the city, which raises the prospect that the death registration figure was even more inaccurate during epidemics.78 Grimshaw estimated that the under-reporting of deaths in the municipal area of Dublin was 9 per cent in the previous decade so from 1879 he used burial returns to ‘correct’ the registered statistics.79 Maternal deaths pose a range of other research issues in relation to the definition, classification and certification, which Section IV examines in finer detail. IV. Maternal Mortality in Dublin, 1870, 1880 and 1890 Unlike their rural contemporaries, women in Dublin had a range of birthing options available to them; these included qualified, assisted delivery in domestic locations, lying-in hospitals and charitable institutions. The Rotunda, one of Europe’s oldest and largest lying-in centres, offered free medical assistance, as did the Coombe, so they were primarily patronised by the poor. There was also a proliferation of domiciliary midwifery schemes operating in the late nineteenth century. As Table 6 shows, domiciliary services such as the Rotunda’s were used extensively. Further to this a network of unqualified midwives or handywomen operated in the city.80 Unfortunately, the ARRG did not publish maternal mortality statistics to RD level. From the outset of civil registration the GRO had divided Ireland into eight regional ‘Divisions’, for reporting purposes, and in the ARRG the lowest level for maternal mortality figures were to ‘Division’ level.81 Since Dublin lay within the Eastern Division, which spanned all of three and parts of seven other counties, that reporting level is of little use in examining maternal mortality in the city.82 (These higher-level divisions were germane to English, Welsh and Scottish registers too.) For such reasons we examined each individual GRO death register entry in the North and South Dublin city’s RDs in order to determine the extent of recorded maternal mortality levels for three pre-census years—1870, by which time the returns were increasingly accurate, 1880, a year of general high mortality, and 1890, when the city’s mortality rates were reportedly in decline.83 The Rotunda registers survive so we used them to provide a benchmark of reporting accuracy from lying-in hospitals.84 As several inconsistencies emerged in how, or indeed if, deaths were registered we examined the BSDH annual reports to cross reference registered deaths with those located in the lying-in hospitals. Established by statute in 1856, the BSDH reported on a variety of medical matters in the eleven Dublin hospitals under its remit, including the Rotunda and the Coombe.85 The information reported to the BSDH by the two maternity hospitals varied over time, and occasionally one or other failed to report. For most years the information available in the annual reports included details on the childbirth-related deaths occurring during the previous year, and often date and cause of death were related. Detailed information on the diseases prevalent in the hospitals is provided too, and for six years for the Rotunda and one year for the Coombe the names of the deceased women were recorded in the reports. A complicating factor, however, is that the BSDH reports span a period from 1 April to 31 March the following year, making it difficult to match figures in the reports with maternal deaths recorded by registrars in a calendar year. Take the following case as an example. In 1869 as part of the rebuttal of Evory Kennedy’s damning report on puerperal fever in hospitals, the Coombe returned a detailed table of maternal mortality (from January 1861 until December 1868) to the Dublin Obstetric Society.86 In that timeframe 54 women died, eight from puerperal fever, but more problematically for our purposes 14 deaths were from peritonitis, which is categorised in non-gender specific terms in the ARRGs until 1902 and were probably not therefore returned as maternal. When the list was compared to the BDSH annual reports only 43 of the 54 deaths could be accounted for because the reporting periods are out of synchrony. As the admission registers for the Rotunda hospital are extant, we examined these in great detail. What is evident from official statistics emerging from the Rotunda is that mortality rates declined year on year. An unprecedented 80 childbirth-related deaths occurred in the hospital in 1861–1862, but above 30 deaths in a year were reported on seven other occasions; all of these occurred by the mid 1870s. In the quarter century after 1880–1881 the number of reported maternity deaths in what it termed the ‘labour ward’ did not exceed 17 during any year, but we suspect some cases were hidden because terminally ill maternity patients could be transferred to the on-site auxiliary hospital (which only ‘treated diseases peculiar to women’), or even be discharged from the hospital.87 The auxiliary hospital was administratively separate to the maternity section and, from 1876, it dealt exclusively with chronic illness and ‘gynaecology cases’.88 Of the 26 deaths occurring in the auxiliary hospital in 1880, for example, two were associated with abortion (defined as the expulsion of the fetus before seven months89), three with peritonitis and another five with septicaemia, diseases which are commonly associated with pregnancy or childbirth.90 But these deaths were not necessarily recorded as maternal deaths, thus falsely improving the official statistics for the maternity hospital. Puerperal or childbirth fever was ‘dimly’ recognised as infectious from the seventeenth century. But in the 1840s advances were made in America with the publication in 1843 of Oliver Wendell Holmes’ essay entitled ‘The Contagiousness of Puerperal Fever’ in the New England Quarterly Journal of Medicine and in Europe by Ignaz Semmelweis. His work at the Allgemeine Krankenhaus in Vienna showed that puerperal fever among women delivered by doctors was much higher than those delivered by midwives. He concluded that poor hand hygiene among doctors was the cause and introduced a rigorous hand-washing policy.91 His results were stark but emphatically rejected, as miasma theory still held sway. It was not until the 1870s that Joseph Lister’s anti-septic surgical practices were applied in maternity hospitals.92 Further to advances in medical science, Louis Pasteur’s discoveries in microbiology provided irrefutable evidence to debunk the arguments of the anti-contagionists.93 Under the Mastership of Lombe Atthill (1875–1882) the Rotunda became a strong proponent of ‘germ theory’ by encouraging the use of carbolic soap and hand washing using carbolic solution.94 Arthur V. Macan, who replaced Atthill, had travelled and received further training in Europe 1870s and he intensified the antiseptic strategies. In his first year of office 1,090 women were admitted and only six died, two of these from causes other than septicaemia. It was later noted that ‘This excellent result had been obtained by careful antiseptic precautions’.95 Improvements were less evident in the Coombe, where the worst year for maternal deaths was 1888–1889, when 16 deaths were returned, out of 430 cases.96 Our comparative analysis of the records supports the fact that the larger lying-in hospitals were usually careful in reporting mortality statistics to the BSDH. But a comparison between the 1880 BSDH figures for the Rotunda with the GRO entries for that year shows some anomalies. The hospital reported 22 maternal deaths occurring in its labour wards in 1880 (Table 2). No names were published in the BSDH, but the dates of admittance, delivery and death are available. By comparing the dates of delivery and death with the maternal-mortality cases identified in the GRO entries, it is possible to identify most of these patients in the Rotunda’s registers. A closer examination of the GRO registers shows that 36 maternal deaths occurred in the Rotunda; 14 more than the number reported by the hospital for its labour wards and our examination of registered births shows that each of these additional deaths was linked to a pregnancy. Table 2 Maternal mortality cases reported by the Rotunda Hospital, 1880 Month  GRO entries  BSDH maternal mortality totals     (additional possible cases)  Labour wards  Auxiliary hospital  January  2 (1)  2 (2 identified)  Data not available in usable form.  February  0 (0)  0  March  2 (0)  1 (1 identified)  April  7 (0)  5 (5 identified)  2 (2 identified)  May  7 (0)  4 (4 identified)  5 (3 identified)  June  8 (0)  2 (1 identified)  3 (2 identified)  July  2 (0)  2 (2 identified)  0  August  2 (0)  2 (2 identified)  0  September  2 (0)  0  2 (2 identified)  October  1 (1)  2 (2 identified)  0  November  0 (0)  0  0  December  3 (1)  2 (1 identified)  2 (2 identified)  1880 total  36 (3)  22 (20 identified)  14 (11 identified)  Month  GRO entries  BSDH maternal mortality totals     (additional possible cases)  Labour wards  Auxiliary hospital  January  2 (1)  2 (2 identified)  Data not available in usable form.  February  0 (0)  0  March  2 (0)  1 (1 identified)  April  7 (0)  5 (5 identified)  2 (2 identified)  May  7 (0)  4 (4 identified)  5 (3 identified)  June  8 (0)  2 (1 identified)  3 (2 identified)  July  2 (0)  2 (2 identified)  0  August  2 (0)  2 (2 identified)  0  September  2 (0)  0  2 (2 identified)  October  1 (1)  2 (2 identified)  0  November  0 (0)  0  0  December  3 (1)  2 (1 identified)  2 (2 identified)  1880 total  36 (3)  22 (20 identified)  14 (11 identified)  Source: NAI/Rotunda Hospital Register NAI/PRIV1263/2/20. In 1870 61 cases of maternal deaths were recorded in the city, but using our age-range criteria (14–54) we identified another eight cases of possible maternal mortality. With 7,154 births being recorded in the city that year, this equates to an MMR of between 853 and 964 per 100,000 births.97 The city’s MMR exceeded the reported national rate, which was 687 per 100,000 births that year.98 For 1880, a vastly different set of circumstances prevailed; 113 definite and a further 80 possible cases of maternal mortality were identified in Dublin city. With the GRO reporting 8,400 births in the city, the MMR ranged from 1,345 to 2,298 per 100,000 births, well ahead of the national rate of 695 per 100,000 births.99 Using the BSDH reports and GRO birth records, we were able to confirm that nine of the 80 possible maternal mortality cases corresponded to a registered birth, thus increasing the minimum MMR in Dublin to 1,452 per 100,000 births that year and providing further evidence of ‘hidden maternal mortality’. By 1890, with the crisis of the early 1880s over, and with a third maternity hospital functioning in Holles Street from 1884 (although it welcomed patients of all creeds ‘its management was exclusively Catholic’), 39 maternal mortality cases were recorded.100 We found another 15 suspect cases, equating to a MMR of between 526 and 728 per 100,000 births, out of the 7,416 births recorded in the city that year.101 This compared favourably with the national rate, of 657 per 100,000 births.102 The RD specific figures are presented in Tables 3, 4 and 5. Table 3 Maternal mortality registrations in Dublin, by RD, 1870 RD  Jan  Feb  Mar  Apr  May  June  July  Aug  Sept  Oct  Nov  Dec  Total  North City, 1      1  1    1  2        1  1  7  North City, 2  3    2    2  3  1    1      13  25  North City, 3        1          1    3  2  7  South City, 1  1  1    2  1  1  1            7  South City, 2  1            1      2      4  South City, 3              1    1      4  6  South City, 4    2  2            1        5  Dublin City  5  3  5  4  3  5  6    4  2  4  20  61  RD  Jan  Feb  Mar  Apr  May  June  July  Aug  Sept  Oct  Nov  Dec  Total  North City, 1      1  1    1  2        1  1  7  North City, 2  3    2    2  3  1    1      13  25  North City, 3        1          1    3  2  7  South City, 1  1  1    2  1  1  1            7  South City, 2  1            1      2      4  South City, 3              1    1      4  6  South City, 4    2  2            1        5  Dublin City  5  3  5  4  3  5  6    4  2  4  20  61  Table 4 Maternal mortality registrations in Dublin, by RD, 1880 RD  Jan  Feb  Mar  Apr  May  June  July  Aug  Sept  Oct  Nov  Dec  Total  North City, 1, E.  1        1  1              3  North City, 1, W.  5  2    5  2  1  1  2    1    1  20  North City, 2,  2  5  2  3  7  9  3  2    1  2  2  38  North City, 3  1        1  1    1          4  South City, 1    2      2  1    1      1    7  South City, 2        1  1  1          2    5  South City, 3  1  3  4  7  3  2  1  2  1  1  2  2  29  South City, 4, E.    1      1  1              3  South City, 4, W.    1    1  1  1              4  Dublin City  10  14  6  17  19  18  5  8  1  3  7  5  113  RD  Jan  Feb  Mar  Apr  May  June  July  Aug  Sept  Oct  Nov  Dec  Total  North City, 1, E.  1        1  1              3  North City, 1, W.  5  2    5  2  1  1  2    1    1  20  North City, 2,  2  5  2  3  7  9  3  2    1  2  2  38  North City, 3  1        1  1    1          4  South City, 1    2      2  1    1      1    7  South City, 2        1  1  1          2    5  South City, 3  1  3  4  7  3  2  1  2  1  1  2  2  29  South City, 4, E.    1      1  1              3  South City, 4, W.    1    1  1  1              4  Dublin City  10  14  6  17  19  18  5  8  1  3  7  5  113  Source: GRO dataset. Table 5 Maternal mortality registrations in Dublin, by RD, 1890 RD  Jan  Feb  Mar  Apr  May  June  July  Aug  Sept  Oct  Nov  Dec  Total  North City, 1, E.                  1        1  North City, 1, W.              1            1  North City, 2  2        1    2    1  2    1  9  North City, 3    1        1  1            3  South City, 1  1    1  3  1    1            7  South City, 2      1                    1  South City, 3  1    5  2  1  1          1    11  South City, 4  1  2  1        1  1          6  Dublin City  5  3  8  5  3  2  6  1  2  2  1  1  39  RD  Jan  Feb  Mar  Apr  May  June  July  Aug  Sept  Oct  Nov  Dec  Total  North City, 1, E.                  1        1  North City, 1, W.              1            1  North City, 2  2        1    2    1  2    1  9  North City, 3    1        1  1            3  South City, 1  1    1  3  1    1            7  South City, 2      1                    1  South City, 3  1    5  2  1  1          1    11  South City, 4  1  2  1        1  1          6  Dublin City  5  3  8  5  3  2  6  1  2  2  1  1  39  Source: GRO dataset. Mapping maternal mortality The localisation of deaths to the vicinity of lying-in hospital settings as the maps corresponding to Tables 3, 4 and 5 show is unsurprising. The Rotunda was located in North City No. 2 RD, which clustered 24 maternal deaths in its surrounds in 1870 (Figure 4). What is extraordinary to us is that no deaths were recorded in or near the city’s two workhouses. Again the limited use of the guiding nosology was a factor; in 1870 the Registrar General issued a circular to the Poor Law Commission admonishing Masters of respective workhouses for furnishing ‘defective’ mortality returns.103 Only 33 out of a possible 145 definite causes of death were used in national returns of 10,639 deaths in workhouses for 1870; 15 of these were deaths in childbirth. Similar returns were given in 1880 (12,940 total deaths, 28 in childbirth, See Figure 5) and 1890 (11,256, 9 in childbirth, See Figure 6).104 Like the ARRGs these annual reports are not disaggregated and the PLU records survive piecemeal so it is not possible to do a comprehensive micro-study of maternal deaths. But the Union Hospitals operated midwifery services and the probability of low to no mortality in overcrowded centres seemed dubious to us. There is evidence of the Rotunda accepting some of the more difficult labour cases from the workhouses but that was not routine practice.105 Figure 4 View largeDownload slide Map of maternal deaths, Dublin metropolitan area, 1870 Note: Purple indicates maternal death from puerperal disease; white circle indicates a maternal death from another cause. A small circle indicates a single maternal death. A large circle indicates deaths in a public institution, with the number showing the number of deaths. Source: GRO registers of deaths, Dublin North SRD and Dublin South SRD; city only, superimposed on OSI map of Dublin City. Figure 4 View largeDownload slide Map of maternal deaths, Dublin metropolitan area, 1870 Note: Purple indicates maternal death from puerperal disease; white circle indicates a maternal death from another cause. A small circle indicates a single maternal death. A large circle indicates deaths in a public institution, with the number showing the number of deaths. Source: GRO registers of deaths, Dublin North SRD and Dublin South SRD; city only, superimposed on OSI map of Dublin City. Figure 5 View largeDownload slide Map of maternal deaths, Dublin metropolitan area, 1880 Note: As per Figure 4. Source: GRO registers of deaths, Dublin North SRD and Dublin South SRD; city only, superimposed on OSI map of Dublin City. Figure 5 View largeDownload slide Map of maternal deaths, Dublin metropolitan area, 1880 Note: As per Figure 4. Source: GRO registers of deaths, Dublin North SRD and Dublin South SRD; city only, superimposed on OSI map of Dublin City. Figure 6 View largeDownload slide Map of maternal deaths, Dublin metropolitan area, 1890 Note: As per Figure 4. Source: GRO registers of deaths, Dublin North SRD and Dublin South SRD; city only, superimposed on OSI map of Dublin City. Figure 6 View largeDownload slide Map of maternal deaths, Dublin metropolitan area, 1890 Note: As per Figure 4. Source: GRO registers of deaths, Dublin North SRD and Dublin South SRD; city only, superimposed on OSI map of Dublin City. With the exception of 1880, a year of high general mortality, the most striking feature of these tables is that maternal-mortality numbers appear low in a large city with appalling levels of poverty and deprivation. For 1870 and 1890 only two months (Tables 3 and 5) attract immediate interest; December 1870, when 20 maternal-mortality cases were recorded, thirteen in North City, No. 2 RD, and March 1890, when eight cases were recorded five of which occurred in South City No. 3 RD. The apparent peak in deaths in December 1870 merits particular comment, as it seems a maternal-mortality crisis was occurring in the Rotunda, the location of 12 of the 13 deaths. That year, 22 of the city’s 61 maternal mortality cases occurred in the Rotunda, yet its informants only attended the RD-office on seven occasions during the year. For instance, three informants attended the office on 30 September, and none appeared again until 31 December, when five attended, reporting all deaths occurring in the hospital during the previous three months.106 Of course, this placed the hospital in the unenviable position of breaching the requirement to report deaths within seven days but prosecutions did not ensue. Two of the three informants in September and three of the five in December were illiterate, which may also have compromised the accuracy of the information that was being reported. It is difficult to ascertain the degree to which nurses received training or their literacy levels prior to the Irish registration acts of 1918 for midwives and 1919 for nurses.107 Prior to these acts, as Gerard Fealy argues, poor wages acted as a deterrent for properly trained nurses and attracted the uneducated instead.108 Attempts were made to start ‘instruction in midwifery’ as soon as the Rotunda was established, which Campbell Ross contends was always part of the original plan of the founder, Bartholomew Mosse.109 By the late nineteenth century, it had a strong reputation for its six-month midwifery training programme but that does not mean that all hospital employees were qualified or literate. Kirkpatrick and Jellett note that salaries as low as £10 per year were paid to ‘regular nurses’ in 1878 and thus ‘the class of women who sought employment was not very high; some were quite illiterate, and none of them were required to show evidence of any special training or aptitude for their position’.110 In 1876 the Rotunda initiated a domiciliary service, which was systematic and flourished as the years progressed.111 By 1890 the domiciliary midwifery service was comprehensive, and the use of what McClintock termed ‘intercurrent’ or underlying disease as cause of death may be a factor in hidden maternal mortality (Table 6). Studies have found low maternal mortality in areas that operated domiciliary services, but in some instances the reported rates appear too low to be credible.112 A hospital commission in 1887 reprimanded the Coombe for poor oversight of external maternity cases; while the Rotunda’s external operation was just as big and unwieldy (see Table 6 for patient figures for the Rotunda, 1889–91), it was not subjected to the same criticism.113 Table 6 Rotunda Hospital, dispensary and domiciliary cases, 1889–1891 For year ended 31 March  1876  1889  1890  1891  Lying-in hospital admissions  1,206  1,538  1,599  1,526  Auxiliary hospital admissions  259  417  383  427  Attended at their own homes (extern maternity)  368  1,662  1,687  1,845  Treated at the hospital dispensary  4,261  10,602  9,268  8,624  For year ended 31 March  1876  1889  1890  1891  Lying-in hospital admissions  1,206  1,538  1,599  1,526  Auxiliary hospital admissions  259  417  383  427  Attended at their own homes (extern maternity)  368  1,662  1,687  1,845  Treated at the hospital dispensary  4,261  10,602  9,268  8,624  Source: Report of the Rotunda Hospitals for poor lying-in women, and for the treatment of diseases peculiar to women for year ended 31 March, 1877 (Dublin, 1877). 5. Rotunda Hospitals for Poor Lying-in Women and for the treatment of diseases peculiar to women, for year ended 31 March, 1891, Dublin 1891, 7. Although the risk of puerperal fever was reduced dramatically in the domiciliary context if best practice in practitioner hygiene was followed, other risks associated with poor sanitation, and general poverty must have been complicating factors in Dublin. Nuttall, in her discussion of the Edinburgh Royal Maternity Hospital explains that detailed outdoor ledgers were kept for the institution, which included ‘the names, ages and parity of all patients, with the date of their delivery and the classification of their labour’.114 The Rotunda’s outdoor records are sparse by comparison; they simply list the women attended, the time they were attended and by whom, they make no mention of patients’ medical conditions.115 Puerperal fever aside, it is not improbable that a domiciliary service (see Table 6) exceeding the numbers of in-patient admissions could produce an equal number of direct maternal deaths from obstruction, haemorrhage and placental problems, but they are not in evidence here. Nosological developments If varying degrees of nursing competencies presented problems, these were compounded by on-going difficulties with cursory nosology usage. Between 1869 and 1890 several innovations were introduced to promote specificity in cause of death recording. In 1869 The Nomenclature of Diseases was published and it expanded the nosographical range of ‘affections connected with pregnancy’.116 Contemporary observers continued to worry about the use of secondary causes of death and one commentator stated that ‘a certain percentage of deaths in childbirth are indefinitely certified as peritonitis, pyaemia, or in some other manner, without any reference to the fact of childbirth’.117 This lack of precision perturbed several obstetricians and statisticians as the figures simply did not add up. One physician observed in 1873 that mortality attributed to peritonitis among childbearing women in Scotland was twice that of males but that in cohorts beyond reproductive age such disparities were not evident.118 An analysis of the Irish statistics for 1873 shows a similar pattern: of 121 male deaths 44.6 per cent were aged between 15 and 54, the equivalent for women was 66.9 of 157 deaths.119 The second edition of the Nomenclature of Diseases, published in 1884, recommended that ‘the term “Puerperal fever” should no longer be used’. It advised that more specific terms be employed to distinguish between deaths of women and men from peritonitis, and to specify deaths occurring in though not necessarily of childbirth. In spite of recommendations that ‘puerperal’ should also be used as a prefix ‘to the word denoting the local process’ (like ‘puerperal peritonitis’) Ireland’s officialdom was slow to respond.120 In 1890, six years after the publication of the second edition of the Nomenclature of Diseases, puerperal fever was recorded as the official cause of death for 7 of the 39 maternal deaths in the GRO dataset, the same number recorded against septicaemia. For the first time, eclampsia appeared as a recorded cause of death, accounting for more than one in ten maternal deaths in 1890 indicating that some advances in medical science were working their way from the pages of the BMJ and the Lancet into the Irish civil registration system. Two more revisions of the Nomenclature of Diseases were published in 1896 and 1906, each argued how useless ‘puerperal fever’ was as a means of describing childbirth-related disease.121 It was not until the early years of the twentieth century before the ARRG began to treat ‘puerperal fever’ deaths with a sceptical eye. In 1901, using a more inclusive definition, 624 maternal deaths were accounted for nationally, making the annual death rate for 1901 ‘6.18 per 1,000 deaths’, 220 of which were from puerperal fever.122 That year, 323 deaths were caused by peritonitis none of which were ascribed to puerperal causes.123 Suspicious deaths Dublin’s lying-in hospitals appear to have been relatively compliant in recording deaths in women in childbirth but several other deaths of women aged between 14 and 54 in the GRO registers raised suspicions that they could be cases of direct maternal mortality: 8 for 1870, 81 for 1880 and 15 for 1890. All of these were examined in greater detail. For each, the birth records for a 42-day period previous to the death of the woman were examined, to see if a child was born to her within the timeframe—if a child was identified then we considered it a maternal mortality case. For 1880 and 1890 the BSDH reports were also examined, to provide additional information. By these means it was possible to categorise two of the eight deaths in 1870, nine of the 81 deaths in 1880, and two of the 15 deaths in 1890 as maternal mortality cases. Thus, the MMR for Dublin can be recalibrated (Table 7). Table 7 Recalibrated MMR for Dublin, 1870 and 1890 Year  Certain  Possible  Extra  Total  Births,  MM rate per 100,000 births     MM  MM  MM  MM  Dublin                  Dublin  National  1870  61  8  2  63  7,154  880  687  1880  113  81  9  122  8,400  1,452  696  1890  39  15  2  41  7,416  553  657  Year  Certain  Possible  Extra  Total  Births,  MM rate per 100,000 births     MM  MM  MM  MM  Dublin                  Dublin  National  1870  61  8  2  63  7,154  880  687  1880  113  81  9  122  8,400  1,452  696  1890  39  15  2  41  7,416  553  657  It is important to note that most of these hidden mortality cases were identified because a child was born alive to the deceased woman, and survived, thereby requiring the child to be registered; if the births were unregistered then these additional cases would have remained ‘hidden’. It is plausible that some of the remaining 91 cases were also childbirth-related, as further examination may yet reveal. Conclusion Civil registration law relied heavily on the medical profession to provide faithful accounts of cause of death but the guidance given permitted much leeway in childbirth cases. This paper shows that historic trends in reported MMR for Ireland are inaccurate and points to several reasons why. Here we examined registered, and therefore legally compliant, deaths and have found idiosyncrasies in reporting from institutions and caution that mortality outside institutions was not as assiduously recorded. In the first instance, we posit that the implementation of abstract government policy had several operational difficulties to overcome because Ireland presented additional cultural and Socio-economic challenges to the rest of the United Kingdom. Denominational factors and the coupling of civil registration with poor law administration posed a major impediment to the GRO’s authority. The poor law may have provided a geographically comprehensive public health network but when civil registration was layered over it, no provision was made for Ireland’s unique circumstances. A post-famine economy, social class, gender and denominational factors all compromised the way in which official information was reported, moderated and recorded. This article has identified specific problems about how cause of maternal death may have been hidden in institutions. The way in which Irish registrars could legally manipulate the categories to the advantage of qualified birth attendants remains an issue. On investigation we found that even maternal mortality reporting by lying-in hospitals was problematic. The first major issue is the selective use of the nosology, which distinguished between deaths in and of childbirth—further research may show such distinctions in evidence elsewhere. By admission of the President of the Dublin Obstetric Society in 1869, practitioners attending to births outside institutions where an underlying condition was present were legally permitted to return the incipient disease rather than childbirth as cause of death. For instance, in the case of peritonitis, non-puerperal cases are not identified in the ARRGs until 1902. Tuberculosis was one of the greatest threats to public health in this timeframe and Greta Jones has argued that the mortality estimates underrepresent the scale of the problem. She contends that frequent pregnancies ‘must have depressed the health of a number of women making them liable to infection, or, alternatively, to a recurrence of a tuberculosis infection’.124 Deaths of women from TB, during pregnancy resulting in a stillbirth are not possible to determine, but late maternal deaths, if live births are recorded, offer potential for further research into that element of hidden maternal deaths. The use of underlying conditions like phthisis effectively drove down official maternal mortality statistics, both in Dublin and nationally. Clearly, a more detailed analysis of appropriate records can uncover hidden maternal mortality in Dublin. Tracing the true extent of the Irish MMR would necessitate a much larger sample from a longer range of consecutive years, cross-referenced between parish, General Register Office (GRO), medical and census records to examine family structure, identify sibship intervals and possible maternal deaths but that was beyond the scope of our project. Our initial focus on GRO mortality records led to an examination of lying-in hospital records and the interplay between both data types; this showed how even reported and registered deaths can be deceptive. Our work has raised several other research questions and a call for the integration of historical data using emerging technologies to deepen our understanding of how people engaged with civil registration and how the architecture of that power functioned.125 Brian Gurrin a Research Associate in Irish Social and Economic History at Ulster University, Northern Ireland. Footnotes 1 Report of the committee appointed by the Local Government Board for Ireland to inquire into the public health of the city of Dublin, 1900 [Cd.243] [Cd.244], 3. 2 Ibid., 3–4. Greta Jones, ‘Captain of all these men of death’: the History of Tuberculosis in Nineteenth and Twentieth Century Ireland (Amsterdam: Rodopi Press, 2001), 1–126. 3 Vincent De Brouwere, ‘The Comparative Study of Maternal Mortality over Time: The Role of the Professionalization of Childbirth’, Social History of Medicine, 2007, 20, 541–62, 542. 4 Rebecca Kippen, ‘Counting Nineteenth-century Maternal Deaths: The Case of Tasmania’, Historical Methods: A Journal of Quantitative and Interdisciplinary History, 2005, 38, 14–25; Irvine Loudon, Death in Childbirth: An International Study of Maternal Care and Maternal Mortality, 1800–1950 (Oxford: Clarendon Press, 1992), 23. 5 Cormac Ó Gráda, ‘Dublin’s Demography in the Early Nineteenth Century: Evidence from the Rotunda’, Population Studies, 1991, 45, 43–54; Cormac Ó Gráda, ‘The Rotunda and the People of Dublin: Glimpses from the Hospital’s Archives’, in Alan Browne (ed.), Masters, Midwives and Ladies-in-Waiting the Rotunda Hospital 1745–1995 (Dublin A. & A. Farmar, 1995); Lindsey Earner-Byrne, Mother and Child: Maternity and Child Welfare in Ireland, 1922–60 (Manchester: Manchester University Press, 2007), 34–46; Earner-Byrne, ‘Managing Motherhood: Negotiating a Maternity Service for Catholic Mothers in Dublin, 1930–1954’, Social History of Medicine, 2006, 19, 261–77; Earner-Byrne, ‘Twixt God and Geography: The Development of Maternity Services in Twentieth-century Ireland’, in Janet Greenlees and Linda Bryder, Western Maternity and Medicine, 1880–1990 (London: Pickering & Chatto, 2013), 99–111 at 100–1. 6 Report of the Inter-Departmental Committee on Physical Deterioration, Vol. I, 1904 [Cd. 2175]; Department of Local Government and Public Health, Report (Dublin: Stationery Office, 1927–8), 40. 7 Irvine Loudon, ‘On Maternal and Infant Mortality 1900–1960’, Social History of Medicine, 1991, 4, 29–73 at 29; Loudon, Death in Childbirth, 541–84. 8 Ciara Breathnach and Brian Gurrin, ‘A tale of two citiesinfant mortality and cause of infant death, Dublin, 1864–1910’, Urban History (2016) online first: 1–31, doi:10.1017/S0963926816000717.. 9 Carlo Ginzburg, John Tedeschi and Anne C. Tedeschi, ‘Microhistory: Two or Three Things That I Know about It’, Critical Inquiry, 1993, 20, 10–35 at 28. 10 We are grateful to the GRO for sharing its rich content with our project. Under the terms and conditions of access we anonymised all data. ‘GRO dataset’ will be used throughout to denote data extracted from the GRO death registers for 1870, 1880 and 1890. 11 Michel Foucault, ‘Society Must be Defended’, in Lectures at the College De France 1975–76 (New York: Allen Lane, 2003), 239–63 at 243. Foucault termed these state efforts ‘biopower’ and defined it as ‘an explosion of numerous and diverse techniques for achieving the subjugations of bodies and the control of populations’. 12 Census of Ireland for the year 1871. Abstract of the Enumerators’ Returns, Showing by Provinces, Counties, Cities, and Certain Corporate Towns, 1871 [C.375], 5. We use the term Anglican faith to group together those professing as Protestant and Presbyterian. According to the 1871 census 4,141,933 professed as Roman Catholic, 683,295 were Protestant Episcopalian, 503,461 Presbyterian and 74,070 ‘all others’. 13 Jurgen Habermas, The Philosophical Discourse of Modernity. Tr. by Frederick Lawrence. (Cambridge, MA: MIT Press, 1998), 2. Laurence M. Geary, ‘The Medical Profession, Health Care and the Poor Law in Nineteenth-Century Ireland’, in Virginia Crossman and Peter Gray (eds), Poverty and Welfare in Ireland, 1838–1948 (Dublin: Irish Academic Press, 2011), 189–206. 14 An Act for the registration of births and deaths in Ireland (26 & 27 Vict. 11), sect. 46 (hereafter 1863 Act). 15 For UK statistics, see Loudon, Death in Childbirth, p. 23. In England Loudon notes that the public only ‘slowly’ got used to the idea of compulsory registration after the changes to the act in 1874. 16 G. Dean and C. J. Mulvihill. ‘The Registration of Births and Deaths in Ireland’, Journal of the Irish Medical Association, 1972, 65, 101–5. G. Dean and H. McLoughlin, ‘The Registration and Certification of Deaths in the West of Ireland’, Irish Medical Journal, 1980, 73, 269–70. See J. F. Connolly, A. Cullen, J. Scott and D. Smithwick, ‘Non-registration of Deaths as a Source of Error in Mortality Data in the Irish Republic’, Irish Journal of Psychological Medicine, 1999, 16, 16–17. 17 Registration of Deaths in Ireland: A Statistical Nosology, Comprising the Causes of Death, Classified and Alphabetically Arranged (Dublin, 1864), iii (hereafter Statistical Nosology); First Annual Report of the Registrar-General of Marriages, Births and Deaths in Ireland, 1864, 7 (hereafter First ARRG). 18 Statistical Nosology, iii. 19 Royal College of Physicians of Ireland (RCPI)/2/2/1/1/1 Occasional Committee Minutes, February/March 1863. The Fellows of the RCPI were invited to comment on the civil registration bill in February 1863. A committee was formed to consider it. It expressed dissatisfaction with the legislation and emphasised five main concerns, which all centred around the responsibility placed on physicians for the accuracy of data under the headings of place of residence, age at death, date of last visit (especially for metropolitan practitioners), cause of death and duration of disease. It recommended that these data emanate ‘from sources external to the medical profession’ but their case had no impact on the subsequent legislation. 20 Anne Cameron, ‘Medicine, Meteorology and Vital Statistics: The Influence of the Royal College of Physicians of Edinburgh upon Scottish civil registration, c. 1840–1855’, Journal of the Royal College of Physicians Edinburgh, 2007, 37, 173–80 at 174–6. 21 An Act for registering Births, Deaths, and Marriages in England, 6 & 7 Will. IV. c.86, section 27. 22 Cameron, ‘Medicine, Meteorology and Vital Statistics’, 178–9. 1863 act, section 46. 23 http://www.who.int/healthinfo/statistics/indmaternalmortality/en/, accessed 8 December 2015. 24 Statistical Nosology, iii; First ARRG, 1864, 7. 25 William Farr, Report on the Nomenclature and Statistical Classification of Diseases for Statistical Returns (Oxford, 1856), 11. 26 James Mathew Duncan, On the Mortality of Childbed and Maternity Hospitals (Edinburgh: Adam and Charles Black, 1870), 6. 27 Peter Conrad, ‘Medicalization and Social Control’, Annual Review of Sociology, 1992, 18, 209–32. Conrad defines the process of medicalisation as ‘defining a problem in medical terms, using medical language to describe a problem, adopting a medical framework to understand a problem, or using a medical intervention to “treat” it’. 28 Anon, BMJ, 30 December 1876, 859. ‘The statistics of puerperal mortality are … more imperfect than are those of most other diseases. … In consequence of the present state of the law … it is impossible to arrive at the correct number of child-bearings’. 29 Jones, ‘Captain of all these men of death’, 68. 30 Report from the Select Committee on Midwives’ Registration, 1892 (289), xiv, 1, 25. 31 On criticism from the Brussels Congress on Hygiene, see Inter-Departmental Committee on Physical Deterioration, 45. On the provision created in 1926 and 1938, see Gayle Davis, ‘Stillbirth Registration and Perceptions of Infant Death, 1900–60: The Scottish Case in National Context’, Economic History Review, 2009, 62, 629–54. 32 Stillbirths Registration Act, 1994, http://www.irishstatutebook.ie/1994/en/act/pub/0001/print.html, accessed 19 June 2015. 33 Helen Andrews, ‘McClintock, Alfred Henry’, in James McGuire and James Quinn, ed., Dictionary of Irish Biography (Cambridge: Cambridge University Press, 2009), 839–40. 34 Dr McClintock’s contribution to the, ‘Report of the Dublin Obstetrical Society’, Dublin Journal of Medical Science, 1869, 48, 266. This meeting was conducted over eleven nights by 17 physicians in response to former Master of the Rotunda, Dr Evory Kennedy’s paper ‘Zymotic Diseases, as more Especially Illustrated by Puerperal Diseases’, Dublin Quarterly Journal of Medical Science, 1867, xliv, 514–21, which suggested radical structural reform of the Rotunda to stop the spread of puerperal disease. See also Andrews, ‘Kennedy, Evory’, in Dictionary of Irish Biography, 108–9. 35 Kennedy, ‘Zymotic Diseases’, 514–21. 36 Report of the Commissioners Appointed to Inquire into the Hospitals of Dublin. With appendices [2063], 1856, xix, 7. 37 First Annual Report of the Board of Superintendence of Dublin Hospitals, with appendices [2353, 1857–58] (hereafter BSDH). 38 S. D. Sessoms, ‘Puerperal Fever’, Journal of the National Medical Association, 1915, 7, 104–8. 39 Both cases occurred in Dublin North City No. 3 registrar’s district. 40 Section 46 of the Irish registration act, 1864, describes the procedure Registrars and Medical practitioners were to follow. An Act to provide for the better regulation of Births, Deaths, and Marriages in Scotland, 17 & 18 Vict. c.80 s. 41. 41 Forty-third ARRG, 1906, xvii, xxxii. See Appendix A for a complete list of the Registrar General’s annual reports which we have used in this paper. 42 ARRG, 1970 (Dublin, 1973), xliv. 43 First and Second Reports from the Select Committee on Death Certification, 46. 44 Stephan Curtis, ‘Midwives and their Role in the Reduction of Direct Obstetric Deaths during the late Nineteenth Century: The Sundsvall Region of Sweden (1860–1890)’, Medical History, 2005, 49, 321–50 at 326–7. 45 ‘An Address Delivered in the Section of Public Medicine’, BMJ, 1879, 2, 242. 46 Edward Higgs, ‘A Cuckoo in the Nest?: The Origins of Civil Registration and State Medical Statistics in England and Wales’, Continuity and Change, 1996, 11, 115–34, at 115–21. 47 J.K.L., Letters on the State of Ireland (Dublin: Coyne, 1825), 96; K. H. Connell, The Population of Ireland, 1750–1845 (Oxford: Clarendon Press, 1950), 2. 48 E. Margaret Crawford, Counting the People; A Survey of the Irish Censuses, 1813–1911 (Dublin: Four Courts Press, 2003), 15. 49 An Act for Marriages in England (6 & 7 William IV, c. 85); An Act for Registering Births, Deaths and Marriages in England (6 & 7 William IV, c. 85). 50 An Act for Marriages in Ireland, and for registering such marriages (7 & 8 Vict., c. 81) (hereafter Irish Marriage act, 1844). 51 Anne Cameron, ‘The Establishment of Civil Registration in Scotland’, The Historical Journal, 2007, 50, 377–97 at 378–9. Civil registration in Scotland commenced on 1 January 1855, with the enactment of an Act to provide for the better registration of births, deaths and marriages in Scotland (17 & 18 Vict., c. 80). Other Bills, in 1837, 1847, 1848 and 1849, had failed to pass into law. 52 Irish Marriage act, 1844, sect. 52. 53 Ibid., sects 57, 58. 54 First Report of the Registrar-General of Marriages in Ireland, under the Provisions of the Act 7 & 8 Vic. Chap. 81 [1130], 1850, xxv, 16–17. 55 The census of Ireland for the year 1851, part vi: General report, with appendix, county table, miscellaneous tables, and index to names of places [2134], 1856, xxxi, I, lix. Crawford, Counting the People, 21–2. Simon Szreter, Fertility, Class and Gender in Britain, 1860–1940 (Cambridge: Cambridge University Press, 1996), 85. 56 House of Commons Journal, ci, part. 2, 1211, 1222 (leave to introduce the Bill granted, 14 August 1846; Bill introduced, 17 August 1846); A bill for registering births, deaths and marriages in Ireland, H.C. 1846 (637), iii, 515, p. 1 (sects 1, 6). 57 Registration of births, &c., (Ireland) Bill H.C. 1859, session 1 (68), ii, 547. In 1860 Opposition bill introduced by Lord Naas, 8 May 1860. Government bill (number 2) introduced by Chief Secretary Cardwell, 10 May 1860. Cardwell’s bill was considered by a committee and amended, 21 May 1860. Both bills were withdrawn, 5 July 1860 (Commons jn, cxv, 230, 235, 257, 356); Hansard Parliamentary Debates, clvii (24 April 1860–6 June 1860), cols 886–7; clix (7 June 1860–20 July 1860), cols 209–10. In 1861 Government bill introduced 11 February 1861. Naas’s Bill introduced 22 February 1861. Select Committee’s modifications, 11 July 1861, which removed birth and death aspects, and amended Cardwell’s Bill with portions of Naas’s. Bill withdrawn, 22 July 1861 (Commons jn, cxvi, pp 37, 74, 354, 381; Hansard Parliamentary Debates, clxiv (28 June 1861–6 August 1861), col. 1285). In 1862 Hugh Cairns’ Marriage Bill introduced 19 February 1862, and Peel’s Births and Deaths Bill introduced on 20 February 1862. Both Bills withdrawn on 2 July 1862 (Commons jn., 58, 60, 305; Hansard Parliamentary Debates, clxvii (27 May 1862–7 July 1862), cols 622, 1313, 1321–2). 58 1863 Act, sect. 4; Irish Times, 9 September 1876. 59 1863 Act, sect. 17 (if a PLU boundary changed then the boundary of the SRD could be modified, to reflect the new union boundary). 60 Ibid., sects 18, 20, 22, 23. 61 Poor Law and Medical Charities (Ireland) Act 14 & 15 Vict c. 68. Geary, ‘The Medical Profession’, 189–206 at 190. Geary argues that the transition from medical charity to poor law was strongly opposed by ‘a substantial segment of the medical profession’ and adds that it ‘regarded any connection with the poor law system as socially and professionally degrading’. Any efforts to mount an ‘opposition campaign’, he also notes, were ‘undone by the Great Famine’. 62 Ronald D. Cassells, Medical Charities, Medical Politics; The Irish Dispensary System and the Poor Law, 1836–1872 (Rochester, NY, 1997), 78. 63 Laurence M. Geary, Medicine and Charity in Ireland 1718–1851 (Dublin: University College Dublin Press, 2004), 209–10. Anon, ‘The Local Government Board of Ireland and Dispensary Doctors’, BMJ, 1896, 1:1827, 54; Anon, ‘Irish Dispensary Doctors’ Grievances Deputation to the Chief Secretary’, BMJ, 1892, 1:1634, 880. Catherine Cox, ‘Access and Management: The Medical Dispensary Service in Post-Famine Ireland’, in Catherine Cox and Maria Luddy, ed., Cultures of Care in Irish Medical History (Basingstoke: Palgrave Macmillan, 2010), 57–78 at 60. 64 1863 Act. 65 Medical Officers Superannuation Act (Ireland), 1869 (32 & 33 Vict., c. 50). 66 1863 Act, sects 31–33. 67 Ibid., sects 36–38. 68 1863 Act; Irish Marriage Act, 1844. 69 Geary, ‘The Medical Profession’, 198–9. 70 Report of the Royal commissioners appointed to inquire into the sewerage and drainage of the city of Dublin, 1880 [C.2605]. 71 Thom’s Irish Almanac and Official Directory of the United Kingdom of Great Britain and Ireland for the year 1877 (Dublin: Thom’s Publications, 1877), 1070. 72 13 & 14 Car II c12, (1662). 73 Dublin city, covering 3,808 acres in 1871, had expanded to 7,911 acres by 1901 (Vaughan and Fitzpatrick, Irish Historical Statistics, Population, 29). 74 Ibid., 243, 247–8. 75 Sixteenth ARRG, 1879, 5–6. Public Health (Ireland) Act, 1878 (41 & 42 Vict., c. 52), sect. 191; Public Health (Ireland) Amendment Act, 1879 (42 & 43 Vict., c. 57), sect. 7. 76 First ARRG, 1864, 64; Second ARRG, 1865, 24; Third ARRG, 1866, 24; Fourth ARRG, 1867, 24; Fifth ARRG, 1868, 24; Sixth ARRG, 1869, 58; Seventh ARRG, 1870, 58; Eight ARRG, 1871, 56; Ninth ARRG, 1872, 56; Tenth ARRG, 1873, 54; Eleventh ARRG, 1874, 58; Twelfth ARRG, 1875, 58; Thirteenth ARRG, 1876, 58; Fourteenth ARRG, 1877, 58; Fifteenth ARRG, 1878, xix, 58; Sixteenth ARRG, 1879, 58; Seventeenth ARRG, 1880, 58; Eighteenth ARRG, 1881, 55; Nineteenth ARRG, 1882, 46; Twentieth ARRG, 1883, 46; Twenty-first ARRG, 1884, 48; Twenty-second ARRG, 1885, 52; Twenty-third ARRG, 1886, 62; Twenty-fourth ARRG, 1887, 62; Twenty-fifth ARRG, 1888, 70; Twenty-sixth ARRG, 1889, 70; Twenty-seventh ARRG, 1890, 70; Twenty-eight ARRG, 1891, 70; Twenty-ninth ARRG, 1892, 72; Thirtieth ARRG, 1893, 72; Thirty-first ARRG, 1894, 74; Thirty-second ARRG, 1895, 74; Thirty-third ARRG, 1896, 74; Thirty-fourth ARRG, 1897, 74; Thirty-fifth ARRG, 1898, 74; Thirty-sixth ARRG, 1899, 50; Thirty-seventh ARRG, 1900, 50; Thirty-eighth ARRG, 1901, 51. 77 Census of Ireland, 1871, pt ii, Vital Statistics, vol. ii, Report and Tables Relating to Deaths [C. 1000], 1874, lxxiv, ci. 78 On the cholera outbreak, see Ibid., c. 79 Report of Royal Commissioners to inquire into the sewerage and drainage, city of Dublin, 32. 80 Ciara Breathnach, ‘Handywomen and Birthing in Rural Ireland 1851–1955’, Gender & History, 2016, 28, 36–58. 81 On the division into regional divisions, see First ARRG, 1864, 18–20. 82 First ARRG, 1864, 18. 83 GRO dataset. 84 Rotunda Hospital Archive is held at the National Archives of Ireland (NAI), usage is by prior approval and strict appointment only. The collection is cited as follows NAI/PRIV/1263. 85 First Annual Report of the BSDH, 3. 86 ‘Report of the Dublin Obstetrical Society’, 290–4. 87 T. Percy Kirkpatrick and Henry Jellett, The Book of the Rotunda Hospital (London, 1913), 158. 88 Ibid., 158. Report of the Rotunda Hospitals for poor lying-in women, and for the treatment of diseases peculiar to women (Dublin, 1877), 5–6. 89 Robley Dunglison, Medical Lexicon: A Dictionary of Medical Science (Philadelphia: Henry C. Lea, 1868), 16. 90 NAI PRIV1263/2/22. 91 Loudon, Death in Childbirth, 49–84. 92 Dr W. J. Smyly, ‘The Maternal Mortality In Childbed’, The Lancet, 1900, 11, 385–8. 93 M. Best and D. Neuhauser, ‘Heroes and Martyrs of Quality and Safety: Ignaz Semmelweis and the birth of infection control’, Quality & Safety in Health Care, 2004, 13, 233–4. 94 Phil Gorey, ‘Puerperal Fever in Dublin: The Case of the Rotunda Lying-in’, in Lisa Marie Griffith and Ciarán Wallace, eds, Grave Matters: Death and Dying in Dublin 1500 to the Present (Dublin, 2016), 46–60 at 56. 95 Reports of Societies’, BMJ, 18 July 1885, 102–3 at 103. 96 Thirty-second report of the BSDH [6073] 1890, 12. 97 Seventh ARRG, 1870, 58. 98 Ibid., 11, 131. 99 Seventeenth ARRG, 1880, 48, 58. 100 Irish Times, 13 March 1894; Irish Times, 20 March 1894. 101 Twenty-seventh ARRG, 1890, 70. 102 Ibid., 1, 14, 19. 103 Annual report of the commissioners for administering the laws for relief of the poor in Ireland, including the twenty-fourth report under the 10 & 11 Vic., c. 90, and the nineteenth report under the 14 & 15 Vic., c. 68; with appendices, 1871 [C.361], 46. 104 Ibid., 15. ARLGB, being the ninth report under “the Local Government Board (Ireland) Act”, 35 & 36 Vic., c. 69. 1881 [C.2926] [C.2926-I], 15 (hereafter ARLGB); Nineteenth ARLGB for Ireland, 1890–91 [C.6439], 16. 105 Arthur Macan, Master of the Rotunda Hospital, to Dublin Hospital’s Commission, Report of the Committee of Inquiry, 1887 [C. 5042], 1887, xxxv, 89. 106 Attending on 30 September were M.G. (illiterate), who informed on six deaths, A.O’C. (literate), informing on three deaths, and M.S. (illiterate) who informed on one death. Attending on 31 December were C.M. (illiterate), informing on four deaths, A.O’C. (literate; five deaths), M.S. (illiterate; six deaths), K.B. (literate; three deaths), and L.W. (illiterate; one death) (GRO records, qr 3, Dublin North City, No. 2 records, p. 340; ibid., qr 4, Dublin North City, No. 2 records, 393–5). 107 Breathnach, ‘Handywomen and birthing’, 36–58. Ciara Breathnach, ‘“ … it would be preposterous to bring a Protestant here”: religion, provincial politics and district nurses in Ireland, 1890–1904’, in D. S. Lucey and Virginia Crossman, eds, Healthcare in Ireland and Britain 1850–1970: Voluntary, Regional and Comparative Perspectives (London, 2015), 161–80. 108 Gerard Fealy, A History of Apprenticeship Nurse Training in Ireland (Abingdon: Routledge, 2006), 24–8. 109 Ian Campbell Ross, ‘Midwifery’ in Ian Campbell Ross, ed., Public Virtue, Public Love: The Early Years of the Dublin Lying-in Hospital, the Rotunda (Dublin: O’Brien Press, 1986), 125–64 at 159. 110 Henry Jellett, The Book of the Rotunda Hospital: An Illustrated History of the Dublin Lying-in Hospital from its Foundation in 1745 to the Present Time (Dublin), 181. 111 Report of the Rotunda Hospitals (Dublin, 1880), 8. The 1886 report notes that Midwifery outcases started in 1874 with 95 cases, it almost doubled to 177 in 1875 and grew steadily in the subsequent ten years to 1892 cases as against 1807 admissions in 1886. Report of the Rotunda Hospital (Dublin, 1886), 7–9 (From 1885 Rotunda Hospital was dropped). 112 Stanley A. Seligman, ‘The Royal Maternity Charity: The First Hundred Years’, Medical History, 1980, 24, 403–18 at 417. Seligman raises suspicions about the two maternal mortality cases reported for London’s Royal Maternity Charity in 1856 out of 3,297 live births (of which 53 were of twins and three of triplets): ‘when the hospitals of London, Vienna, Paris, and Dublin were ravaged by childbed fever’. 113 Dublin Hospitals Commission. Report of the Committee of Inquiry, 1887, together with minutes of evidence and appendices, 1887 [C.5042] xx–xxii. 114 Alison Nuttall,‘Passive Trust or Active Application: Changes in the Management of Difficult Childbirth and the Edinburgh Royal Maternity Hospital, 1850–1890’, Medical History, 2006, 50, 351–72 at 353–4. 115 NAI /PRIV/1263/4. 116 The Nomenclature of Diseases (1st edn, London, 1869), vi, x. It was prepared by a joint committee (which included Farr) appointed by the Royal College of Physicians of London, and was deemed ‘suitable to England, and to all countries where the English language is in common use’. 117 Anon, BMJ, 30 December 1876, 858. Loudon, Death in Childbirth, 28. 118 BMJ, 30 December 1876, 858. 119 Tenth ARRG, 1875, 102–3. 120 The Nomenclature of Diseases (2nd edn, London, 1884), 9, 117. 121 Ibid. (3rd edn, London, 1896), 11; Ibid. (4th edn, London, 1906), 7. Ibid. (5th edn, London, 1918), 4). 122 Thirty-eighth ARRG, 1901, xviii, 18. 123 Ibid. 124 Jones, ‘Captain of all these men of death’, 69. 125 Foucault, ‘Society Must be Defended’. Appendix A. Annual reports of Registrar General First Annual Report of the Registrar-General of Marriages, Births and Deaths in Ireland, 1864 [4137], 1868–9, xvi (hereafter ARRG). Second ARRG, 1865 [C. 4], 1870, xvi. Third ARRG, 1866 [C. 130], 1870, xvi. Fourth ARRG, 1867 [C. 238], 1871, xv. Fifth ARRG, 1868 [C. 238], 1871, xv. Sixth ARRG, 1869 [C. 554], 1872, xvii. Seventh ARRG, 1870 [C. 785], 1873, xx. Eight ARRG, 1871 [C. 968], 1874, xiv. Ninth ARRG, 1872 [C. 1099], 1874, xiv. Tenth ARRG, 1873 [C. 1376], 1876, xviii. Eleventh ARRG, 1874 [C. 1495], 1876, xviii. Twelfth ARRG, 1875 [C. 1617], 1876, xix. Thirteenth ARRG, 1876 [C. 1937], 1878, xxii. Fourteenth ARRG, 1877 [C. 2301], 1878–79, xix. Fifteenth ARRG, 1878 [C. 2388], 1878–79, xix. Sixteenth ARRG, 1879 [C. 2688], 1880, xvi. Seventeenth ARRG, 1880 [C. 3046], 1881, xxvii. Eighteenth ARRG, 1881 [C. 3368], 1882, xix. Nineteenth ARRG, 1882 [C. 3795], 1883, xx. Twentieth ARRG, 1883 [C. 4149], 1884, xx. Twenty-first ARRG, 1884 [C. 4538], 1884–85, xvii. Twenty-second ARRG, 1885 [C. 4801], 1886, xvii. Twenty-third ARRG, 1886 [C. 5153], 1887, xxiii. Twenty-fourth ARRG, 1887 [C. 5537], 1888, xxx. Twenty-fifth ARRG, 1888 [C. 5844], 1889, xxv. Twenty-sixth ARRG, 1889 [C. 6147], 1890, xxiv. Twenty-seventh ARRG, 1890 [C. 6520], 1890–91, xxiii. Twenty-eight ARRG, 1891 [C. 6787], 1892, xxiv. Twenty-ninth ARRG, 1892 [C. 7255], 1893–94, xxi. Thirtieth ARRG, 1893 [C. 7535], 1894, xxv. Thirty-first ARRG, 1894 [C. 7800], 1895, xxiii. Thirty-second ARRG, 1895 [C. 8236], 1896, xxiii. Thirty-third ARRG, 1896 [C. 8560], 1897, xxii. Thirty-fourth ARRG, 1897 [C. 8949], 1898, xviii. Thirty-fifth ARRG, 1898 [C. 9491], 1899, xvi. Thirty-sixth ARRG, 1899 [Cd. 295], 1900, xv. Thirty-seventh ARRG, 1900 [Cd. 697], 1901, xv. Thirty-eighth ARRG, 1901 [Cd. 1225], 1902, xviii. Thirty-ninth ARRG, 1902 [Cd. 1711], 1903, xvi. Fortieth ARRG, 1903 [Cd. 2222], 1904, xiv. Forty-first ARRG, 1904 [Cd. 2673], 1905, xvii. Forty-second ARRG, 1905 [Cd. 3123], 1906, xx. Forty-third ARRG, 1906 [Cd. 3663], 1907, xvi. Forty-fourth ARRG, 1907 [Cd. 4233], 1908, xvii. Forty-fifth ARRG, 1908 [Cd. 4769], 1908, xi. Acknowledgements This work was supported by the Irish Research Council, Interdisciplinary Research Project Grant [2013–3]. We gratefully acknowledge the cooperation of the General Register Office, which has kindly shared its data with the project. © The Author 2017. Published by Oxford University Press on behalf of the Society for the Social History of Medicine. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Social History of Medicine Oxford University Press

Maternal Mortality, Dublin, 1864–1902

Loading next page...
 
/lp/ou_press/maternal-mortality-dublin-1864-1902-Suln4b3gbz
Publisher
Oxford University Press
Copyright
© The Author 2017. Published by Oxford University Press on behalf of the Society for the Social History of Medicine.
ISSN
0951-631X
eISSN
1477-4666
D.O.I.
10.1093/shm/hkx010
Publisher site
See Article on Publisher Site

Abstract

Summary Using civil registration as a prism this article examines the complicated interface between the authorities, medical professionals, women in childbirth and the aftermath of delivery. It argues that Irish maternal mortality was underestimated and provides a number of arguments in relation to the complex socio-medical environment to explain why. Our research shows how a combination of cavalier attitudes towards cause of death classification and liberal interpretations of William Farr’s Statistical Nosology, served to obscure the true extent of maternal mortality in Dublin City from 1864 to 1902. By offering a microhistory of maternal mortality reporting and registration this article problematises the merits of using civil registration data as a resource for the social history of medicine. maternal mortality, gender, Ireland, civil registration Introduction Late nineteenth-century Dublin City held the unenviable record of having one of the United Kingdom’s worst urban mortality rates. The average death rate of 29.5 per thousand in the ten years from 1890 to 1899 was far in excess of London, and ‘the thirty-three large towns of England and Wales’; it was so appalling that it prompted a commission of inquiry in 1900.1 Annual returns to parliament included analyses of age cohorts and cause of death, but, in an era that was characterized by worryingly high mortality in the 5–35 age groups (from tuberculosis and pulmonary diseases), deaths during pregnancy, in childbirth or in the post-natal context did not receive much focused political attention.2 Vincent De Brouwere found in his comparative analysis of the historiography of European maternal mortality that maternal death was too narrowly defined. In addition he argued that because historic maternal mortality rates (MMR) omit deaths caused indirectly by childbirth and late maternal deaths (occurring after 42 days), the official returns are inaccurate.3 Using late-nineteenth-century metropolitan Dublin as a case study, this research aims to examine if underreporting of vital events occurred and to demonstrate whether, as a phenomenon, ‘hidden maternal mortality’ was as inherent to Irish civil registers as, for instance, Rebecca Kippen found in Tasmania.4 Apart from Cormac Ó Gráda and Lindsey Earner-Byrne’s work, scholars of Ireland have largely ignored the issues of maternal and infant mortality.5 This is partly due to the fact that, despite various government commissions on the health of nations, maternal health only began to emerge as a discrete medical issue in Ireland in the late 1920s.6 Historically, maternal mortality has been defined by its relationship to live births, which is why, with few exceptions it is usually studied in conjunction with infant mortality. Irvine Loudon has argued that analysing infant and maternal mortality together can provide insightful contrasts.7 But, given the nascent state of Irish scholarship in the respective fields, we posit that more useful insights can be garnered by separating them.8 This article is primarily concerned with mortality reporting under the civil registration system, which was introduced fully in 1864. As distinct from Loudon’s methods, which relied on national statistics, a combination of macro- and microhistory approaches are used to problematise the data presented in official Irish mortality returns.9 For example, an outline of overall mortality in Dublin from 1864 to 1901 is given to contextualise maternal mortality over that period, and is followed by a micro-study of the death registers in three pre-census years to assess the accuracy of the returns.10 Using these techniques we identified peculiarities in the registration process and patterns in the localisation of maternal death. In the data we use here 74 per cent of the maternal deaths were certified; all occurred in hospital settings shortly after delivery (mostly within one week, and rarely above one month), which raises questions, as De Brouwere cautions, about those occurring outside institutions, those attributed to indirect (underlying causes) and late maternal mortality. We trace these peculiarities to Ireland’s unique set of political and socioeconomic circumstances, which made the implementation of census taking and civil registration (both crucial to effective governance) very challenging.11 There are several reasons for this that range in origin from class and creed to culture. Historically the Irish people had a fractious relationship with British administration in all its guises and therefore studies of official census and civil registration returns should acknowledge that these data are deficient. Religion, power, social class and gender are critical issues in the timeframe we examine: the majority of the population (roughly 75 per cent nationally, but higher again for Dublin) was Roman Catholic and poor but men of Protestant faiths (primarily Protestant Episcopalians and Presbyterians) were the main powerbrokers.12 That civil registration was placed under the auspices of a public medical system, which was grappling with professionalisation, specialisation and later modernisation, did not aid its progress.13 The act stipulated that deaths should be medically certified, but the legislation paid little heed to the importance of accurate cause of death reporting.14 Furthermore, the matters of uncertified deaths, unspecified cause of death and imprecise categorisation of cause combined to reduce the accuracy of the records. The remainder of this article is divided into four sections. In Section I we outline our source material and explain our methodology. We discuss the merits of using civil registration data for medical history research and contend that some of the problems arising from the data emanate from the architecture of the system itself. While there are instances of poor compliancy with civil registration across the United Kingdom, casual observance of the law was more pronounced in Ireland where mortality under-registration persisted into the 1990s.15 A series of studies conducted from the 1970s to the 1990s advocated that undertakers form part of the solution to the problem of underreporting in the West of Ireland.16 To provide an understanding of why the Irish were reluctant to engage with certain aspects of civil registration, Section II outlines the foundation of the system and contextualises its operation within broader Irish administration. The General Register Office (GRO) was crudely grafted onto the Irish Poor Law in 1864, which created scope for general mistrust and engendered a fear of surveillance. We contend that the pre-history of the poor law, and its strong associations with the Great Famine, naturally inhibited the success of the GRO: it entrenched the power of the former system and was a retrograde step. In Section III we use two GRO data types—aggregate annual returns to parliament and individual register entries—to isolate overall mortality figures from 1864 to 1913 in order to contextualise the MMR. Section IV endeavours to give a better understanding of the nature of maternal mortality for metropolitan Dublin by examining a variety of available data. In the first instance, we extracted entries of maternal deaths from the 1870, 1880 and 1890 registers and mapped the data to see if local factors such as professional maternity services had an impact on the rate. Unsurprisingly maternal deaths were localised to maternity hospitals but the geographical dispersion shows a curious absence of deaths in workhouse hospitals, which operated maternity services. The city also had several ad hoc domiciliary nursing services, so the clustering of the majority of deaths near maternal hospitals raised our suspicions about underreporting elsewhere. The history of hidden maternal deaths is methodologically difficult to unravel and this article aims to show potential avenues for future research. To that end our concluding section returns to the matter of cause of death catergorisation and traces how international discourses on disease classification only gradually caused an improvement in cause of maternal death reporting in Ireland. Overall we show that deaths reported by health care professionals are problematic up to 1902. I. Methodology: Problematising the Sources In this section we focus on registered maternal deaths and how they were defined and categorised to show how the published aggregate data are fundamentally inaccurate. For the purposes of the 1863 act the Registrar General supplied a reprint of William Farr’s Statistical Nosology of Causes of Death to all registrars and registered medical practitioners in Ireland to standardise civil-registration entries, and achieve ‘uniformity in the registry of the causes of death’.17 The GRO in England first published Farr’s Nosology in 1845, a few years after civil registration had commenced there. In the Irish edition, the preface emphasised that the process was not intended to be doctrinaire and the nosology was pitched as ‘suggestions … to medical practitioners’.18 Unlike their Scottish counterparts, who mounted a sustained albeit unsuccessful campaign against the adoption of the ‘flawed English schedule’, Irish medical practitioners did not respond so vociferously.19 Instead, with respect to maternal death, they later used Farr’s guidelines to make a distinction between death in but not of childbirth.20 Under the English system non-medical personnel, such as undertakers, could issue death certificates.21 Opposition from the Fellows of the Royal College of Physicians of Edinburgh gave rise to the inclusion of a system of fines for uncertified deaths under its civil registration law. Irish physicians benefited from the Scottish advances and similar clauses were included in the 1863 Irish act, but only in cases where the deceased had sought medical counsel prior to death.22 World Health Organisation (WHO) defines a death occurring ‘within 42 days of the termination of pregnancy’, irrespective of duration, site or outcome, as a direct maternal death.23 Historically it was largely understood to mean a death that occurred within six weeks of the commencement of labour, following a successful pregnancy. Farr’s Statistical Nosology listed 145 causes of death.24 It made some allowances for arcane medical terminology through the provision of ‘synonymes’, which could be used ‘at the discretion of the medical informants’.25 Obstetrics was a nascent field in the 1860s and even recognised experts such as the Scottish physician, Dr James Matthews Duncan, used loose terminology of ‘say four weeks’ post-partum to define a maternal death.26 Childbirth was increasingly ‘medicalised’ in the timeframe we examine, but deaths in women who did not go to full term could technically be classified under categories such as fever or other underlying health conditions.27 Despite the fact that doctors were encouraged to note ‘pregnancy’ where applicable in cause of death recording, it was not a legal requirement, so maternal morbidity was poorly captured in the Irish records. The same provisions formed part of the English registration act and troubled contributors to the British Medical Journal (BMJ).28 Another factor that obscures the full extent of maternal mortality and morbidity in the United Kingdom is that stillbirths were not registered. Ireland was in the grip of a tuberculosis epidemic in the late nineteenth century and Greta Jones draws attention to the physiological effects of childbearing in aggravating latent cases. She also speculates (although it is difficult to quantify precise mortality) that tuberculosis was ‘implicated in stillbirths and in perinatal deaths in children’.29 As early as 1892 a House of Commons select committee heard that ‘every country in Europe, except Russia, registers still-born children’.30 Despite further criticism at the Brussels Congress on Hygiene in 1903 regarding the absence of stillbirth registration in the United Kingdom, provision was not made until 1926 for England and Wales and 1938 for Scotland.31 Ireland’s lag is quite remarkable: a national system for registering stillbirths was not introduced until 1994.32 The absence of stillbirths from the historical record means that an accurate study of Irish fertility and antenatal health from GRO records over a longue durée is impossible. From the outset we recognised that most deaths were recorded because they occurred in maternity hospitals but that there was a strong possibility even these were underestimated. Dr McClintock, Master of the Rotunda Lying-In Hospital, north of the river Liffey (founded in 1745) between 1854 and 1861, summarised how doctors understood their duties at a meeting of the Dublin Obstetrical Society in 1869.33 … if the cause of death is known to be puerperal fever—or anything pertaining thereto—quite a panic is created in the neighbourhood, and both doctor and nurse come in for more than their fair share of blame. Hence for their own sake, as well as the charitable motive of not alarming all the pregnant women in the community, the death is imputed to any other cause rather than the dreaded puerperal. … The defect lies in our system of registration—not in those who supply the returns. Practitioners make very proper distinction between dying in childbirth, and dying of childbirth. When a woman happens to die in childbed of some intercurrent disease—as phthisis, pneumonia, dysentery, apoplexy, albuminuria, bronchitis, morbis cordis, &c—this alone is returned and rightly so—to the registrar as cause of death. Consequently all these deaths have no place in the registration reports of deaths in childbed. (Emphasis in original)34 This was penned in response to an open letter written by the former Master of the Rotunda, Dr Evory Kennedy, on the structural failures of the hospital to cope with puerperal fever outbreaks.35 McClintock infers that even lying-in hospitals, could manipulate their cause of death statistics if an underlying disease was present, which points to a major weakness in these data. Given the way in which maternal death was defined and classified, a central issue for our research was whether or not lying-in hospitals might be guilty of manipulating their statistics, by attributing maternal deaths to other causes. Therefore, where extant, we examined their records to determine the accuracy of the published data. The two principal maternity hospitals in Dublin were the Rotunda and the Coombe Lying-In Hospital (founded in 1826), in the south west of the city.36 Both hospitals were charitable institutions and catered primarily for Dublin’s poor. Maternal mortality statistics from the Rotunda and the Coombe were published from 1858 in the Board of Superintendence of Dublin Hospitals (BSDH) annual reports.37 We used these in tandem with the Rotunda hospital in-patient registers (registers for the Coombe hospital pre-1914 are not extant) and the GRO death entries to enable us to ascertain how these records related to one another and to see if there were any discrepancies. As it was clear that we were dealing with deficient data we employed a second process in our examination of the GRO registers, where we identified other possible direct maternal deaths of women aged between 14 and 54 and used birth information and other sources to test our suspicions. Apart from being labour intensive, we encountered many difficulties. First, maternal mortality cases are not clear-cut. Obviously childbirth-related diseases such as ‘puerperal fever’ present no ambiguity, but other conditions such as peritonitis (a potential outcome in the final stages of puerperal fever) or phthisis do.38 The second difficulty is that the verification of a suspected pregnancy case depends on whether a live birth ensued and if it was subsequently registered. In one instance we identified an additional maternal death in 1880 caused by ‘unavoidable haemorrhage and pleurisy’ because of the accidental entry of a stillbirth in the GRO birth record that was subsequently struck out by the registrar when the error was realised.39 Uncertified deaths It is important to stress that death notification underwent an iterative process of translation, which increased the potential for inaccuracy and allowed scope for an interpretation of the cause of each death. The duty of recording maternal deaths occurring in hospital settings fell to the medical attendant, who was responsible for categorising the precise cause of death.40 While Irish law was clear on the registration procedure, the method followed by registrars is difficult to determine. A GRO death entry required information on whether it was certified or not, but there was no requirement to record who certified it, thereby reducing accountability. The Annual Reports of the Registrar General (ARRG) began to acknowledge the problem of uncertified deaths in 1906, when it was estimated that 24.4 per cent of all registered deaths were uncertified, and in rural areas that percentage was much higher.41 It was not until 1961 that the proportion of uncertified deaths nationally fell below 5 per cent.42 Uncertified deaths occurring in domestic settings are even more challenging because descriptions of causes of death were filtered through ordinary individuals, often family members with no medical expertise, or through midwives (qualified and unqualified) with varying degrees of medico-legal knowledge. Efforts to standardise disease classification were frustrated by uncertified deaths on which Thomas Wrigley Grimshaw (Registrar General, 1879–1900) commented that the cause of death ‘is a mere statement … it may all be incorrect’.43 In uncertified cases, the circumstances of a death were coloured in the first instance by vernacular accounts of cause of death, which in turn was complicated by the way in which registrars moderated the information they received. It was then returned to central administration for transcription and collation, and each stage created multiple opportunities for error or inaccuracy. An example of how the process of moderating cause of death data can distort the eventual returns is shown in Stephan Curtis’ analysis of Sundsvall in Sweden. His study of 213 direct maternal deaths among women aged 15–45 notes a curious absence of deaths from haemorrhage and suggests that they might have been classified under ‘difficult birth’ in the parish registers.44 Deficiencies in the data troubled nineteenth-century observers who invariably voiced their concerns in medical journals. For instance, Dr Francis Vacher, medical officer of health for Birkenhead, in an 1887 address to the Sanitary Institute of Great Britain, bemoaned the poor quality of cause of death data in the English mortality returns.45 If the death registration system was found to be flawed in England then post-famine Ireland was an even greater challenge, which Section II explains in greater detail. II. The Origins of Irish Civil Registration Edward Higgs has argued that the English GRO is erroneously perceived in binary terms, as a secular vital-registration agency on one hand, or as an organ of the public health movement on the other. He reminds us that Thomas Cromwell’s original 1538 parish registration act was a secular endeavour that aimed to regulate property rights.46 No such property regulations were in place in Ireland and civil registration most certainly became an important instrument of the public health movement. The process of introducing full civil registration was delayed primarily for socio-political reasons, but it was also hampered by religious differences, which had posed problems for the English authorities in Ireland for centuries. For example, there was strong resistance to the early Irish censuses in 1813–15 (which was incomplete) and again in 1821. Commenting on prevailing attitudes during the 1821 census, James Doyle, Roman Catholic bishop of Kildare and Leighlin, observed that ‘The Catholics have ever been unwilling to make known their numbers to any agent of the Government. Having too often experienced from it what they deemed treachery or injustice …’47 In spite of the difficulties, E. Margaret Crawford describes the 1821 census as ‘a great improvement on its predecessor’, not least because it was the first all-Ireland estimate of Irish population to see completion.48 Thereafter, national censuses were held decennially until 1911. Civil registration commenced in the United Kingdom in 1837, with the passage of two statutes in 1836, which applied to England and Wales only.49 In Ireland a first step towards civil registration was taken with the passage of the 1844 marriage registration act which stipulated that from 1 April 1845 all non-Catholic Irish marriages were to be registered by the civil authorities.50 Following a number of failed attempts, Scotland introduced full civil registration in 1855 leaving Ireland as the only part of the United Kingdom without a system of birth and death registration.51 Under the 1844 marriage registration act the Lord Lieutenant of Ireland was authorised to establish the GRO in Dublin and appoint a Registrar General of Marriages.52 The act provided for the division of the country into ‘districts’, with a register office and a district registrar, established in each.53 Initially 130 civil registration districts were created. These were coterminous with the Poor Law Unions (PLU) but employed a discrete staff. William Donnelly was appointed Registrar General and he proposed that the administration of civil registration and the poor law be kept separate; the Lord Lieutenant concurred.54 Gradually the GRO came to absorb other functions; it became responsible for carrying out the decennial census in 1851 until 1911, staffing was remarkably consistent and of excellent calibre.55 With respect to Irish civil registration five bills were brought before the House of Commons in 1846, 1859, 1860, 1861 and 1862, but all failed to pass.56 The 1846 proposal, introduced by Morgan John O’Connell and Benjamin Chapman was structurally unsound, whereas the Bills introduced between 1859 and 1862 foundered because of political differences between the parties as to whether the registration process should be tethered to the policing or the Poor Law systems; the Conservatives favoured the former, and the Liberals the latter. Financial constraints led to a Liberal victory, as the poor law option was cheaper.57 In 1863 the matter was finally resolved with the passage of two acts, one to register births and deaths, and the second to register Catholic marriages. The Act for the Registration of Births and Deaths in Ireland came into force on 1 January 1864. William Donnelly was assigned the position of Registrar General of Births and Deaths, which he held until his resignation in 1876.58 Like the 1844 act, the registration process for births and deaths was based on the PLUs, with each Union being deemed a ‘Superintendent Registrar’s District’ (SRD).59 Each SRD had a register office funded out of PLU funds, and a Superintendent Registrar, who was usually the Clerk of the Union. Each SRD was subdivided into ‘Registrar’s Districts’ (RD), with a registrar responsible for registering births and deaths.60 The Poor Law dispensary system had been established under the 1851 Medical Charities Act, and provided the poor with free medical care.61 Cassells argues that the 1851 act conferred to the Poor Law Commission a level of power over public health and medical affairs that was ‘unprecedented in Ireland and unparalleled in the United Kingdom’.62 It created 723 Dispensary Districts from the then 163 PLUs, each was under the control of a dispensary doctor who was answerable to a management committee.63 When registration commenced, dispensary doctors became responsible for registering births and deaths in their areas.64 Payment for duties such as civil registration and successful vaccinations were undoubtedly a welcome addition to the basic salary of a dispensary doctor, which was determined annually by the Board of Guardians and paid out of the poor rate.65 Civil registration placed a great deal of responsibility on individuals to develop a medico-legal awareness. In the case of a birth, the law tasked parents with informing the registrar. If they were unable to do so, responsibility then devolved to the birth attendant or occupiers of the same dwelling. Births were to be notified within twenty-one days.66 Similar to the procedure for a birth, a death was to be reported to the registrar by somebody present at the event or occupiers of the same dwelling. In normal circumstances deaths were to be notified within seven days, and an informant could be summoned to sign the register within the next fourteen days.67 To incentivise prompt, accurate reporting, a complicated system of fines was introduced for registrars and informants alike.68 Geary argues that despite establishment claims of Ireland being better served medically than any other country, the poor did not share the ‘enthusiasm for the workhouse hospitals’.69 It seems they were also reluctant to engage with other services associated with the PLU system, Thomas Wrigley Grimshaw in 1879 conceded that 9 per cent of deaths in Dublin in the previous decade ‘escaped registration’.70 It is to the matter of Dublin city’s MMR that our attention now turns. III. Dublin mortality rates in context Given the emphasis that the 1863 act placed on administrative boundaries, this section begins by tracing how it was applied to the Dublin metropolitan area. It proceeds by outlining national mortality rates between 1864 and 1913, in order to contextualise our micro-study of maternal mortality in the chosen sample pre-census years of 1870, 1880 and 1890. County Dublin was covered by seven SRDs, and two of these, Dublin North and Dublin South, encompassed all of Dublin city, and its rural hinterland (Figure 1). Dublin North SRD was divided into nine RDs, three (and later four), of which were in the city. Dublin South SRD comprised ten RDs, including four city divisions.71 Figure 1 View largeDownload slide Superintendent registrars’ districts in County Dublin Source: Author’s own. Figure 1 View largeDownload slide Superintendent registrars’ districts in County Dublin Source: Author’s own. Civil registration data for large urban centres present specific research challenges: the birth and death rates of RDs containing public institutions, such as maternity hospitals and workhouses, were artificially inflated by non-residents and therefore do not reflect natural population change within the RD. Ireland did not have laws of settlement, which would have caused demographic readjustment with the official return of strangers to their parishes of origin.72 The ARRGs report on the number of births and deaths per year at various administrative levels, including to the level of RD. The pattern for deaths in Dublin varied from the national picture (Figures 2 and 3). While between 1871 and 1911 the national population declined at each successive census, Dublin’s population increased. The increase can be partially explained by an expansion in the city’s borough boundaries by the 1898 Local Government (Ireland) Act but that change had no impact on the city’s RDs.73 In the thirty-four years between 1864 and 1901 a mean of 7,470 deaths per year were recorded in the two urban SRDs (Figure 3). Figure 2 View largeDownload slide Number of deaths recorded in Ireland, 1864–1901 Note: Horizontal line shows the average number of deaths per year (89,582) in Ireland, 1864–1901. Source: Vaughan and Fitzpatrick, Irish Historical Statistics, Population, 247–8. Figure 2 View largeDownload slide Number of deaths recorded in Ireland, 1864–1901 Note: Horizontal line shows the average number of deaths per year (89,582) in Ireland, 1864–1901. Source: Vaughan and Fitzpatrick, Irish Historical Statistics, Population, 247–8. Figure 3 View largeDownload slide Number of deaths recorded in Dublin city, 1864–1901 Note: Horizontal line shows the average number of deaths per year (7,470) in Dublin city, 1864–1901. Source: ARRGs, 1864–1901 (see footnote 76 for complete references). Figure 3 View largeDownload slide Number of deaths recorded in Dublin city, 1864–1901 Note: Horizontal line shows the average number of deaths per year (7,470) in Dublin city, 1864–1901. Source: ARRGs, 1864–1901 (see footnote 76 for complete references). A notable peak in deaths occurred in the three-year period 1878–1880, and in the latter two years the national death total exceeded 100,000, the only occasions when recorded deaths in a calendar year breached that benchmark. The 105,089 deaths registered in 1879 was the highest number of deaths recorded in the period we examine.74 Although these years coincided with the most serious agricultural crisis since the 1840s, the ARRG enthusiastically observed that the increase was not solely due to increased mortality; instead it cited increased engagement with civil registration and better accounting for burials under the Public Health (Ireland) Acts, 1878–1879.75 Similarly the outstanding feature of Dublin city’s death records (Figure 3) is the mortality peak in 1879 and 1880. We discuss the ramifications for maternal mortality further in the next section but in both years the mortality level exceeded the city’s mean level in the period 1864–1901 (7,470 deaths) by between 25 and 30 per cent. In two other years (1887 and 1899) the number of deaths recorded exceeded the mean level (1868–1901) by more than 10 per cent. The 9,653 deaths recorded in the city’s RDs in 1880 formed a peak in mortality in Dublin, which was not exceeded within our timeframe.76 The ARRGs only reported on deaths that were recorded by the various registrars, but, since not all deaths were registered, the published figures were deficient. The GRO was not blind to this and indeed noted that the number of burials in Dublin’s cemeteries and graveyards exceeded the number of registered deaths of city residents by almost 10 per cent during the first seven years of registration (see Table 1).77 Table 1 Dublin city burials and death registrations, 1864–1870   1864  1865  1866  1867  1868  1869  1870  1864–1870  Burials  7,116  7,536  8,338  8,493  7,248  7,237  7,352  53,320  Reg. deaths  6,260  6,959  7,571  7,374  6,804  6,557  6,625  48,150  Excess burials  856  577  767  1,119  444  680  727  5,170  % discrepancy  12.0%  7.7%  9.2%  13.2%  6.1%  9.4%  9.9%  9.7%    1864  1865  1866  1867  1868  1869  1870  1864–1870  Burials  7,116  7,536  8,338  8,493  7,248  7,237  7,352  53,320  Reg. deaths  6,260  6,959  7,571  7,374  6,804  6,557  6,625  48,150  Excess burials  856  577  767  1,119  444  680  727  5,170  % discrepancy  12.0%  7.7%  9.2%  13.2%  6.1%  9.4%  9.9%  9.7%  Source: Census of Ireland, 1871, pt ii, vital statistics, vol. ii, report and tables relating to deaths, ci. Conformity with death registration law appears to have been particularly bad during two of these seven years, 1864 and 1867, and the suggested reasons are illustrative. Understandably, deficiencies were evident for 1864, the first year of full civil registration, but the significant divergence between burials and registered deaths in 1867 was attributed to an outbreak of cholera in the city, which raises the prospect that the death registration figure was even more inaccurate during epidemics.78 Grimshaw estimated that the under-reporting of deaths in the municipal area of Dublin was 9 per cent in the previous decade so from 1879 he used burial returns to ‘correct’ the registered statistics.79 Maternal deaths pose a range of other research issues in relation to the definition, classification and certification, which Section IV examines in finer detail. IV. Maternal Mortality in Dublin, 1870, 1880 and 1890 Unlike their rural contemporaries, women in Dublin had a range of birthing options available to them; these included qualified, assisted delivery in domestic locations, lying-in hospitals and charitable institutions. The Rotunda, one of Europe’s oldest and largest lying-in centres, offered free medical assistance, as did the Coombe, so they were primarily patronised by the poor. There was also a proliferation of domiciliary midwifery schemes operating in the late nineteenth century. As Table 6 shows, domiciliary services such as the Rotunda’s were used extensively. Further to this a network of unqualified midwives or handywomen operated in the city.80 Unfortunately, the ARRG did not publish maternal mortality statistics to RD level. From the outset of civil registration the GRO had divided Ireland into eight regional ‘Divisions’, for reporting purposes, and in the ARRG the lowest level for maternal mortality figures were to ‘Division’ level.81 Since Dublin lay within the Eastern Division, which spanned all of three and parts of seven other counties, that reporting level is of little use in examining maternal mortality in the city.82 (These higher-level divisions were germane to English, Welsh and Scottish registers too.) For such reasons we examined each individual GRO death register entry in the North and South Dublin city’s RDs in order to determine the extent of recorded maternal mortality levels for three pre-census years—1870, by which time the returns were increasingly accurate, 1880, a year of general high mortality, and 1890, when the city’s mortality rates were reportedly in decline.83 The Rotunda registers survive so we used them to provide a benchmark of reporting accuracy from lying-in hospitals.84 As several inconsistencies emerged in how, or indeed if, deaths were registered we examined the BSDH annual reports to cross reference registered deaths with those located in the lying-in hospitals. Established by statute in 1856, the BSDH reported on a variety of medical matters in the eleven Dublin hospitals under its remit, including the Rotunda and the Coombe.85 The information reported to the BSDH by the two maternity hospitals varied over time, and occasionally one or other failed to report. For most years the information available in the annual reports included details on the childbirth-related deaths occurring during the previous year, and often date and cause of death were related. Detailed information on the diseases prevalent in the hospitals is provided too, and for six years for the Rotunda and one year for the Coombe the names of the deceased women were recorded in the reports. A complicating factor, however, is that the BSDH reports span a period from 1 April to 31 March the following year, making it difficult to match figures in the reports with maternal deaths recorded by registrars in a calendar year. Take the following case as an example. In 1869 as part of the rebuttal of Evory Kennedy’s damning report on puerperal fever in hospitals, the Coombe returned a detailed table of maternal mortality (from January 1861 until December 1868) to the Dublin Obstetric Society.86 In that timeframe 54 women died, eight from puerperal fever, but more problematically for our purposes 14 deaths were from peritonitis, which is categorised in non-gender specific terms in the ARRGs until 1902 and were probably not therefore returned as maternal. When the list was compared to the BDSH annual reports only 43 of the 54 deaths could be accounted for because the reporting periods are out of synchrony. As the admission registers for the Rotunda hospital are extant, we examined these in great detail. What is evident from official statistics emerging from the Rotunda is that mortality rates declined year on year. An unprecedented 80 childbirth-related deaths occurred in the hospital in 1861–1862, but above 30 deaths in a year were reported on seven other occasions; all of these occurred by the mid 1870s. In the quarter century after 1880–1881 the number of reported maternity deaths in what it termed the ‘labour ward’ did not exceed 17 during any year, but we suspect some cases were hidden because terminally ill maternity patients could be transferred to the on-site auxiliary hospital (which only ‘treated diseases peculiar to women’), or even be discharged from the hospital.87 The auxiliary hospital was administratively separate to the maternity section and, from 1876, it dealt exclusively with chronic illness and ‘gynaecology cases’.88 Of the 26 deaths occurring in the auxiliary hospital in 1880, for example, two were associated with abortion (defined as the expulsion of the fetus before seven months89), three with peritonitis and another five with septicaemia, diseases which are commonly associated with pregnancy or childbirth.90 But these deaths were not necessarily recorded as maternal deaths, thus falsely improving the official statistics for the maternity hospital. Puerperal or childbirth fever was ‘dimly’ recognised as infectious from the seventeenth century. But in the 1840s advances were made in America with the publication in 1843 of Oliver Wendell Holmes’ essay entitled ‘The Contagiousness of Puerperal Fever’ in the New England Quarterly Journal of Medicine and in Europe by Ignaz Semmelweis. His work at the Allgemeine Krankenhaus in Vienna showed that puerperal fever among women delivered by doctors was much higher than those delivered by midwives. He concluded that poor hand hygiene among doctors was the cause and introduced a rigorous hand-washing policy.91 His results were stark but emphatically rejected, as miasma theory still held sway. It was not until the 1870s that Joseph Lister’s anti-septic surgical practices were applied in maternity hospitals.92 Further to advances in medical science, Louis Pasteur’s discoveries in microbiology provided irrefutable evidence to debunk the arguments of the anti-contagionists.93 Under the Mastership of Lombe Atthill (1875–1882) the Rotunda became a strong proponent of ‘germ theory’ by encouraging the use of carbolic soap and hand washing using carbolic solution.94 Arthur V. Macan, who replaced Atthill, had travelled and received further training in Europe 1870s and he intensified the antiseptic strategies. In his first year of office 1,090 women were admitted and only six died, two of these from causes other than septicaemia. It was later noted that ‘This excellent result had been obtained by careful antiseptic precautions’.95 Improvements were less evident in the Coombe, where the worst year for maternal deaths was 1888–1889, when 16 deaths were returned, out of 430 cases.96 Our comparative analysis of the records supports the fact that the larger lying-in hospitals were usually careful in reporting mortality statistics to the BSDH. But a comparison between the 1880 BSDH figures for the Rotunda with the GRO entries for that year shows some anomalies. The hospital reported 22 maternal deaths occurring in its labour wards in 1880 (Table 2). No names were published in the BSDH, but the dates of admittance, delivery and death are available. By comparing the dates of delivery and death with the maternal-mortality cases identified in the GRO entries, it is possible to identify most of these patients in the Rotunda’s registers. A closer examination of the GRO registers shows that 36 maternal deaths occurred in the Rotunda; 14 more than the number reported by the hospital for its labour wards and our examination of registered births shows that each of these additional deaths was linked to a pregnancy. Table 2 Maternal mortality cases reported by the Rotunda Hospital, 1880 Month  GRO entries  BSDH maternal mortality totals     (additional possible cases)  Labour wards  Auxiliary hospital  January  2 (1)  2 (2 identified)  Data not available in usable form.  February  0 (0)  0  March  2 (0)  1 (1 identified)  April  7 (0)  5 (5 identified)  2 (2 identified)  May  7 (0)  4 (4 identified)  5 (3 identified)  June  8 (0)  2 (1 identified)  3 (2 identified)  July  2 (0)  2 (2 identified)  0  August  2 (0)  2 (2 identified)  0  September  2 (0)  0  2 (2 identified)  October  1 (1)  2 (2 identified)  0  November  0 (0)  0  0  December  3 (1)  2 (1 identified)  2 (2 identified)  1880 total  36 (3)  22 (20 identified)  14 (11 identified)  Month  GRO entries  BSDH maternal mortality totals     (additional possible cases)  Labour wards  Auxiliary hospital  January  2 (1)  2 (2 identified)  Data not available in usable form.  February  0 (0)  0  March  2 (0)  1 (1 identified)  April  7 (0)  5 (5 identified)  2 (2 identified)  May  7 (0)  4 (4 identified)  5 (3 identified)  June  8 (0)  2 (1 identified)  3 (2 identified)  July  2 (0)  2 (2 identified)  0  August  2 (0)  2 (2 identified)  0  September  2 (0)  0  2 (2 identified)  October  1 (1)  2 (2 identified)  0  November  0 (0)  0  0  December  3 (1)  2 (1 identified)  2 (2 identified)  1880 total  36 (3)  22 (20 identified)  14 (11 identified)  Source: NAI/Rotunda Hospital Register NAI/PRIV1263/2/20. In 1870 61 cases of maternal deaths were recorded in the city, but using our age-range criteria (14–54) we identified another eight cases of possible maternal mortality. With 7,154 births being recorded in the city that year, this equates to an MMR of between 853 and 964 per 100,000 births.97 The city’s MMR exceeded the reported national rate, which was 687 per 100,000 births that year.98 For 1880, a vastly different set of circumstances prevailed; 113 definite and a further 80 possible cases of maternal mortality were identified in Dublin city. With the GRO reporting 8,400 births in the city, the MMR ranged from 1,345 to 2,298 per 100,000 births, well ahead of the national rate of 695 per 100,000 births.99 Using the BSDH reports and GRO birth records, we were able to confirm that nine of the 80 possible maternal mortality cases corresponded to a registered birth, thus increasing the minimum MMR in Dublin to 1,452 per 100,000 births that year and providing further evidence of ‘hidden maternal mortality’. By 1890, with the crisis of the early 1880s over, and with a third maternity hospital functioning in Holles Street from 1884 (although it welcomed patients of all creeds ‘its management was exclusively Catholic’), 39 maternal mortality cases were recorded.100 We found another 15 suspect cases, equating to a MMR of between 526 and 728 per 100,000 births, out of the 7,416 births recorded in the city that year.101 This compared favourably with the national rate, of 657 per 100,000 births.102 The RD specific figures are presented in Tables 3, 4 and 5. Table 3 Maternal mortality registrations in Dublin, by RD, 1870 RD  Jan  Feb  Mar  Apr  May  June  July  Aug  Sept  Oct  Nov  Dec  Total  North City, 1      1  1    1  2        1  1  7  North City, 2  3    2    2  3  1    1      13  25  North City, 3        1          1    3  2  7  South City, 1  1  1    2  1  1  1            7  South City, 2  1            1      2      4  South City, 3              1    1      4  6  South City, 4    2  2            1        5  Dublin City  5  3  5  4  3  5  6    4  2  4  20  61  RD  Jan  Feb  Mar  Apr  May  June  July  Aug  Sept  Oct  Nov  Dec  Total  North City, 1      1  1    1  2        1  1  7  North City, 2  3    2    2  3  1    1      13  25  North City, 3        1          1    3  2  7  South City, 1  1  1    2  1  1  1            7  South City, 2  1            1      2      4  South City, 3              1    1      4  6  South City, 4    2  2            1        5  Dublin City  5  3  5  4  3  5  6    4  2  4  20  61  Table 4 Maternal mortality registrations in Dublin, by RD, 1880 RD  Jan  Feb  Mar  Apr  May  June  July  Aug  Sept  Oct  Nov  Dec  Total  North City, 1, E.  1        1  1              3  North City, 1, W.  5  2    5  2  1  1  2    1    1  20  North City, 2,  2  5  2  3  7  9  3  2    1  2  2  38  North City, 3  1        1  1    1          4  South City, 1    2      2  1    1      1    7  South City, 2        1  1  1          2    5  South City, 3  1  3  4  7  3  2  1  2  1  1  2  2  29  South City, 4, E.    1      1  1              3  South City, 4, W.    1    1  1  1              4  Dublin City  10  14  6  17  19  18  5  8  1  3  7  5  113  RD  Jan  Feb  Mar  Apr  May  June  July  Aug  Sept  Oct  Nov  Dec  Total  North City, 1, E.  1        1  1              3  North City, 1, W.  5  2    5  2  1  1  2    1    1  20  North City, 2,  2  5  2  3  7  9  3  2    1  2  2  38  North City, 3  1        1  1    1          4  South City, 1    2      2  1    1      1    7  South City, 2        1  1  1          2    5  South City, 3  1  3  4  7  3  2  1  2  1  1  2  2  29  South City, 4, E.    1      1  1              3  South City, 4, W.    1    1  1  1              4  Dublin City  10  14  6  17  19  18  5  8  1  3  7  5  113  Source: GRO dataset. Table 5 Maternal mortality registrations in Dublin, by RD, 1890 RD  Jan  Feb  Mar  Apr  May  June  July  Aug  Sept  Oct  Nov  Dec  Total  North City, 1, E.                  1        1  North City, 1, W.              1            1  North City, 2  2        1    2    1  2    1  9  North City, 3    1        1  1            3  South City, 1  1    1  3  1    1            7  South City, 2      1                    1  South City, 3  1    5  2  1  1          1    11  South City, 4  1  2  1        1  1          6  Dublin City  5  3  8  5  3  2  6  1  2  2  1  1  39  RD  Jan  Feb  Mar  Apr  May  June  July  Aug  Sept  Oct  Nov  Dec  Total  North City, 1, E.                  1        1  North City, 1, W.              1            1  North City, 2  2        1    2    1  2    1  9  North City, 3    1        1  1            3  South City, 1  1    1  3  1    1            7  South City, 2      1                    1  South City, 3  1    5  2  1  1          1    11  South City, 4  1  2  1        1  1          6  Dublin City  5  3  8  5  3  2  6  1  2  2  1  1  39  Source: GRO dataset. Mapping maternal mortality The localisation of deaths to the vicinity of lying-in hospital settings as the maps corresponding to Tables 3, 4 and 5 show is unsurprising. The Rotunda was located in North City No. 2 RD, which clustered 24 maternal deaths in its surrounds in 1870 (Figure 4). What is extraordinary to us is that no deaths were recorded in or near the city’s two workhouses. Again the limited use of the guiding nosology was a factor; in 1870 the Registrar General issued a circular to the Poor Law Commission admonishing Masters of respective workhouses for furnishing ‘defective’ mortality returns.103 Only 33 out of a possible 145 definite causes of death were used in national returns of 10,639 deaths in workhouses for 1870; 15 of these were deaths in childbirth. Similar returns were given in 1880 (12,940 total deaths, 28 in childbirth, See Figure 5) and 1890 (11,256, 9 in childbirth, See Figure 6).104 Like the ARRGs these annual reports are not disaggregated and the PLU records survive piecemeal so it is not possible to do a comprehensive micro-study of maternal deaths. But the Union Hospitals operated midwifery services and the probability of low to no mortality in overcrowded centres seemed dubious to us. There is evidence of the Rotunda accepting some of the more difficult labour cases from the workhouses but that was not routine practice.105 Figure 4 View largeDownload slide Map of maternal deaths, Dublin metropolitan area, 1870 Note: Purple indicates maternal death from puerperal disease; white circle indicates a maternal death from another cause. A small circle indicates a single maternal death. A large circle indicates deaths in a public institution, with the number showing the number of deaths. Source: GRO registers of deaths, Dublin North SRD and Dublin South SRD; city only, superimposed on OSI map of Dublin City. Figure 4 View largeDownload slide Map of maternal deaths, Dublin metropolitan area, 1870 Note: Purple indicates maternal death from puerperal disease; white circle indicates a maternal death from another cause. A small circle indicates a single maternal death. A large circle indicates deaths in a public institution, with the number showing the number of deaths. Source: GRO registers of deaths, Dublin North SRD and Dublin South SRD; city only, superimposed on OSI map of Dublin City. Figure 5 View largeDownload slide Map of maternal deaths, Dublin metropolitan area, 1880 Note: As per Figure 4. Source: GRO registers of deaths, Dublin North SRD and Dublin South SRD; city only, superimposed on OSI map of Dublin City. Figure 5 View largeDownload slide Map of maternal deaths, Dublin metropolitan area, 1880 Note: As per Figure 4. Source: GRO registers of deaths, Dublin North SRD and Dublin South SRD; city only, superimposed on OSI map of Dublin City. Figure 6 View largeDownload slide Map of maternal deaths, Dublin metropolitan area, 1890 Note: As per Figure 4. Source: GRO registers of deaths, Dublin North SRD and Dublin South SRD; city only, superimposed on OSI map of Dublin City. Figure 6 View largeDownload slide Map of maternal deaths, Dublin metropolitan area, 1890 Note: As per Figure 4. Source: GRO registers of deaths, Dublin North SRD and Dublin South SRD; city only, superimposed on OSI map of Dublin City. With the exception of 1880, a year of high general mortality, the most striking feature of these tables is that maternal-mortality numbers appear low in a large city with appalling levels of poverty and deprivation. For 1870 and 1890 only two months (Tables 3 and 5) attract immediate interest; December 1870, when 20 maternal-mortality cases were recorded, thirteen in North City, No. 2 RD, and March 1890, when eight cases were recorded five of which occurred in South City No. 3 RD. The apparent peak in deaths in December 1870 merits particular comment, as it seems a maternal-mortality crisis was occurring in the Rotunda, the location of 12 of the 13 deaths. That year, 22 of the city’s 61 maternal mortality cases occurred in the Rotunda, yet its informants only attended the RD-office on seven occasions during the year. For instance, three informants attended the office on 30 September, and none appeared again until 31 December, when five attended, reporting all deaths occurring in the hospital during the previous three months.106 Of course, this placed the hospital in the unenviable position of breaching the requirement to report deaths within seven days but prosecutions did not ensue. Two of the three informants in September and three of the five in December were illiterate, which may also have compromised the accuracy of the information that was being reported. It is difficult to ascertain the degree to which nurses received training or their literacy levels prior to the Irish registration acts of 1918 for midwives and 1919 for nurses.107 Prior to these acts, as Gerard Fealy argues, poor wages acted as a deterrent for properly trained nurses and attracted the uneducated instead.108 Attempts were made to start ‘instruction in midwifery’ as soon as the Rotunda was established, which Campbell Ross contends was always part of the original plan of the founder, Bartholomew Mosse.109 By the late nineteenth century, it had a strong reputation for its six-month midwifery training programme but that does not mean that all hospital employees were qualified or literate. Kirkpatrick and Jellett note that salaries as low as £10 per year were paid to ‘regular nurses’ in 1878 and thus ‘the class of women who sought employment was not very high; some were quite illiterate, and none of them were required to show evidence of any special training or aptitude for their position’.110 In 1876 the Rotunda initiated a domiciliary service, which was systematic and flourished as the years progressed.111 By 1890 the domiciliary midwifery service was comprehensive, and the use of what McClintock termed ‘intercurrent’ or underlying disease as cause of death may be a factor in hidden maternal mortality (Table 6). Studies have found low maternal mortality in areas that operated domiciliary services, but in some instances the reported rates appear too low to be credible.112 A hospital commission in 1887 reprimanded the Coombe for poor oversight of external maternity cases; while the Rotunda’s external operation was just as big and unwieldy (see Table 6 for patient figures for the Rotunda, 1889–91), it was not subjected to the same criticism.113 Table 6 Rotunda Hospital, dispensary and domiciliary cases, 1889–1891 For year ended 31 March  1876  1889  1890  1891  Lying-in hospital admissions  1,206  1,538  1,599  1,526  Auxiliary hospital admissions  259  417  383  427  Attended at their own homes (extern maternity)  368  1,662  1,687  1,845  Treated at the hospital dispensary  4,261  10,602  9,268  8,624  For year ended 31 March  1876  1889  1890  1891  Lying-in hospital admissions  1,206  1,538  1,599  1,526  Auxiliary hospital admissions  259  417  383  427  Attended at their own homes (extern maternity)  368  1,662  1,687  1,845  Treated at the hospital dispensary  4,261  10,602  9,268  8,624  Source: Report of the Rotunda Hospitals for poor lying-in women, and for the treatment of diseases peculiar to women for year ended 31 March, 1877 (Dublin, 1877). 5. Rotunda Hospitals for Poor Lying-in Women and for the treatment of diseases peculiar to women, for year ended 31 March, 1891, Dublin 1891, 7. Although the risk of puerperal fever was reduced dramatically in the domiciliary context if best practice in practitioner hygiene was followed, other risks associated with poor sanitation, and general poverty must have been complicating factors in Dublin. Nuttall, in her discussion of the Edinburgh Royal Maternity Hospital explains that detailed outdoor ledgers were kept for the institution, which included ‘the names, ages and parity of all patients, with the date of their delivery and the classification of their labour’.114 The Rotunda’s outdoor records are sparse by comparison; they simply list the women attended, the time they were attended and by whom, they make no mention of patients’ medical conditions.115 Puerperal fever aside, it is not improbable that a domiciliary service (see Table 6) exceeding the numbers of in-patient admissions could produce an equal number of direct maternal deaths from obstruction, haemorrhage and placental problems, but they are not in evidence here. Nosological developments If varying degrees of nursing competencies presented problems, these were compounded by on-going difficulties with cursory nosology usage. Between 1869 and 1890 several innovations were introduced to promote specificity in cause of death recording. In 1869 The Nomenclature of Diseases was published and it expanded the nosographical range of ‘affections connected with pregnancy’.116 Contemporary observers continued to worry about the use of secondary causes of death and one commentator stated that ‘a certain percentage of deaths in childbirth are indefinitely certified as peritonitis, pyaemia, or in some other manner, without any reference to the fact of childbirth’.117 This lack of precision perturbed several obstetricians and statisticians as the figures simply did not add up. One physician observed in 1873 that mortality attributed to peritonitis among childbearing women in Scotland was twice that of males but that in cohorts beyond reproductive age such disparities were not evident.118 An analysis of the Irish statistics for 1873 shows a similar pattern: of 121 male deaths 44.6 per cent were aged between 15 and 54, the equivalent for women was 66.9 of 157 deaths.119 The second edition of the Nomenclature of Diseases, published in 1884, recommended that ‘the term “Puerperal fever” should no longer be used’. It advised that more specific terms be employed to distinguish between deaths of women and men from peritonitis, and to specify deaths occurring in though not necessarily of childbirth. In spite of recommendations that ‘puerperal’ should also be used as a prefix ‘to the word denoting the local process’ (like ‘puerperal peritonitis’) Ireland’s officialdom was slow to respond.120 In 1890, six years after the publication of the second edition of the Nomenclature of Diseases, puerperal fever was recorded as the official cause of death for 7 of the 39 maternal deaths in the GRO dataset, the same number recorded against septicaemia. For the first time, eclampsia appeared as a recorded cause of death, accounting for more than one in ten maternal deaths in 1890 indicating that some advances in medical science were working their way from the pages of the BMJ and the Lancet into the Irish civil registration system. Two more revisions of the Nomenclature of Diseases were published in 1896 and 1906, each argued how useless ‘puerperal fever’ was as a means of describing childbirth-related disease.121 It was not until the early years of the twentieth century before the ARRG began to treat ‘puerperal fever’ deaths with a sceptical eye. In 1901, using a more inclusive definition, 624 maternal deaths were accounted for nationally, making the annual death rate for 1901 ‘6.18 per 1,000 deaths’, 220 of which were from puerperal fever.122 That year, 323 deaths were caused by peritonitis none of which were ascribed to puerperal causes.123 Suspicious deaths Dublin’s lying-in hospitals appear to have been relatively compliant in recording deaths in women in childbirth but several other deaths of women aged between 14 and 54 in the GRO registers raised suspicions that they could be cases of direct maternal mortality: 8 for 1870, 81 for 1880 and 15 for 1890. All of these were examined in greater detail. For each, the birth records for a 42-day period previous to the death of the woman were examined, to see if a child was born to her within the timeframe—if a child was identified then we considered it a maternal mortality case. For 1880 and 1890 the BSDH reports were also examined, to provide additional information. By these means it was possible to categorise two of the eight deaths in 1870, nine of the 81 deaths in 1880, and two of the 15 deaths in 1890 as maternal mortality cases. Thus, the MMR for Dublin can be recalibrated (Table 7). Table 7 Recalibrated MMR for Dublin, 1870 and 1890 Year  Certain  Possible  Extra  Total  Births,  MM rate per 100,000 births     MM  MM  MM  MM  Dublin                  Dublin  National  1870  61  8  2  63  7,154  880  687  1880  113  81  9  122  8,400  1,452  696  1890  39  15  2  41  7,416  553  657  Year  Certain  Possible  Extra  Total  Births,  MM rate per 100,000 births     MM  MM  MM  MM  Dublin                  Dublin  National  1870  61  8  2  63  7,154  880  687  1880  113  81  9  122  8,400  1,452  696  1890  39  15  2  41  7,416  553  657  It is important to note that most of these hidden mortality cases were identified because a child was born alive to the deceased woman, and survived, thereby requiring the child to be registered; if the births were unregistered then these additional cases would have remained ‘hidden’. It is plausible that some of the remaining 91 cases were also childbirth-related, as further examination may yet reveal. Conclusion Civil registration law relied heavily on the medical profession to provide faithful accounts of cause of death but the guidance given permitted much leeway in childbirth cases. This paper shows that historic trends in reported MMR for Ireland are inaccurate and points to several reasons why. Here we examined registered, and therefore legally compliant, deaths and have found idiosyncrasies in reporting from institutions and caution that mortality outside institutions was not as assiduously recorded. In the first instance, we posit that the implementation of abstract government policy had several operational difficulties to overcome because Ireland presented additional cultural and Socio-economic challenges to the rest of the United Kingdom. Denominational factors and the coupling of civil registration with poor law administration posed a major impediment to the GRO’s authority. The poor law may have provided a geographically comprehensive public health network but when civil registration was layered over it, no provision was made for Ireland’s unique circumstances. A post-famine economy, social class, gender and denominational factors all compromised the way in which official information was reported, moderated and recorded. This article has identified specific problems about how cause of maternal death may have been hidden in institutions. The way in which Irish registrars could legally manipulate the categories to the advantage of qualified birth attendants remains an issue. On investigation we found that even maternal mortality reporting by lying-in hospitals was problematic. The first major issue is the selective use of the nosology, which distinguished between deaths in and of childbirth—further research may show such distinctions in evidence elsewhere. By admission of the President of the Dublin Obstetric Society in 1869, practitioners attending to births outside institutions where an underlying condition was present were legally permitted to return the incipient disease rather than childbirth as cause of death. For instance, in the case of peritonitis, non-puerperal cases are not identified in the ARRGs until 1902. Tuberculosis was one of the greatest threats to public health in this timeframe and Greta Jones has argued that the mortality estimates underrepresent the scale of the problem. She contends that frequent pregnancies ‘must have depressed the health of a number of women making them liable to infection, or, alternatively, to a recurrence of a tuberculosis infection’.124 Deaths of women from TB, during pregnancy resulting in a stillbirth are not possible to determine, but late maternal deaths, if live births are recorded, offer potential for further research into that element of hidden maternal deaths. The use of underlying conditions like phthisis effectively drove down official maternal mortality statistics, both in Dublin and nationally. Clearly, a more detailed analysis of appropriate records can uncover hidden maternal mortality in Dublin. Tracing the true extent of the Irish MMR would necessitate a much larger sample from a longer range of consecutive years, cross-referenced between parish, General Register Office (GRO), medical and census records to examine family structure, identify sibship intervals and possible maternal deaths but that was beyond the scope of our project. Our initial focus on GRO mortality records led to an examination of lying-in hospital records and the interplay between both data types; this showed how even reported and registered deaths can be deceptive. Our work has raised several other research questions and a call for the integration of historical data using emerging technologies to deepen our understanding of how people engaged with civil registration and how the architecture of that power functioned.125 Brian Gurrin a Research Associate in Irish Social and Economic History at Ulster University, Northern Ireland. Footnotes 1 Report of the committee appointed by the Local Government Board for Ireland to inquire into the public health of the city of Dublin, 1900 [Cd.243] [Cd.244], 3. 2 Ibid., 3–4. Greta Jones, ‘Captain of all these men of death’: the History of Tuberculosis in Nineteenth and Twentieth Century Ireland (Amsterdam: Rodopi Press, 2001), 1–126. 3 Vincent De Brouwere, ‘The Comparative Study of Maternal Mortality over Time: The Role of the Professionalization of Childbirth’, Social History of Medicine, 2007, 20, 541–62, 542. 4 Rebecca Kippen, ‘Counting Nineteenth-century Maternal Deaths: The Case of Tasmania’, Historical Methods: A Journal of Quantitative and Interdisciplinary History, 2005, 38, 14–25; Irvine Loudon, Death in Childbirth: An International Study of Maternal Care and Maternal Mortality, 1800–1950 (Oxford: Clarendon Press, 1992), 23. 5 Cormac Ó Gráda, ‘Dublin’s Demography in the Early Nineteenth Century: Evidence from the Rotunda’, Population Studies, 1991, 45, 43–54; Cormac Ó Gráda, ‘The Rotunda and the People of Dublin: Glimpses from the Hospital’s Archives’, in Alan Browne (ed.), Masters, Midwives and Ladies-in-Waiting the Rotunda Hospital 1745–1995 (Dublin A. & A. Farmar, 1995); Lindsey Earner-Byrne, Mother and Child: Maternity and Child Welfare in Ireland, 1922–60 (Manchester: Manchester University Press, 2007), 34–46; Earner-Byrne, ‘Managing Motherhood: Negotiating a Maternity Service for Catholic Mothers in Dublin, 1930–1954’, Social History of Medicine, 2006, 19, 261–77; Earner-Byrne, ‘Twixt God and Geography: The Development of Maternity Services in Twentieth-century Ireland’, in Janet Greenlees and Linda Bryder, Western Maternity and Medicine, 1880–1990 (London: Pickering & Chatto, 2013), 99–111 at 100–1. 6 Report of the Inter-Departmental Committee on Physical Deterioration, Vol. I, 1904 [Cd. 2175]; Department of Local Government and Public Health, Report (Dublin: Stationery Office, 1927–8), 40. 7 Irvine Loudon, ‘On Maternal and Infant Mortality 1900–1960’, Social History of Medicine, 1991, 4, 29–73 at 29; Loudon, Death in Childbirth, 541–84. 8 Ciara Breathnach and Brian Gurrin, ‘A tale of two citiesinfant mortality and cause of infant death, Dublin, 1864–1910’, Urban History (2016) online first: 1–31, doi:10.1017/S0963926816000717.. 9 Carlo Ginzburg, John Tedeschi and Anne C. Tedeschi, ‘Microhistory: Two or Three Things That I Know about It’, Critical Inquiry, 1993, 20, 10–35 at 28. 10 We are grateful to the GRO for sharing its rich content with our project. Under the terms and conditions of access we anonymised all data. ‘GRO dataset’ will be used throughout to denote data extracted from the GRO death registers for 1870, 1880 and 1890. 11 Michel Foucault, ‘Society Must be Defended’, in Lectures at the College De France 1975–76 (New York: Allen Lane, 2003), 239–63 at 243. Foucault termed these state efforts ‘biopower’ and defined it as ‘an explosion of numerous and diverse techniques for achieving the subjugations of bodies and the control of populations’. 12 Census of Ireland for the year 1871. Abstract of the Enumerators’ Returns, Showing by Provinces, Counties, Cities, and Certain Corporate Towns, 1871 [C.375], 5. We use the term Anglican faith to group together those professing as Protestant and Presbyterian. According to the 1871 census 4,141,933 professed as Roman Catholic, 683,295 were Protestant Episcopalian, 503,461 Presbyterian and 74,070 ‘all others’. 13 Jurgen Habermas, The Philosophical Discourse of Modernity. Tr. by Frederick Lawrence. (Cambridge, MA: MIT Press, 1998), 2. Laurence M. Geary, ‘The Medical Profession, Health Care and the Poor Law in Nineteenth-Century Ireland’, in Virginia Crossman and Peter Gray (eds), Poverty and Welfare in Ireland, 1838–1948 (Dublin: Irish Academic Press, 2011), 189–206. 14 An Act for the registration of births and deaths in Ireland (26 & 27 Vict. 11), sect. 46 (hereafter 1863 Act). 15 For UK statistics, see Loudon, Death in Childbirth, p. 23. In England Loudon notes that the public only ‘slowly’ got used to the idea of compulsory registration after the changes to the act in 1874. 16 G. Dean and C. J. Mulvihill. ‘The Registration of Births and Deaths in Ireland’, Journal of the Irish Medical Association, 1972, 65, 101–5. G. Dean and H. McLoughlin, ‘The Registration and Certification of Deaths in the West of Ireland’, Irish Medical Journal, 1980, 73, 269–70. See J. F. Connolly, A. Cullen, J. Scott and D. Smithwick, ‘Non-registration of Deaths as a Source of Error in Mortality Data in the Irish Republic’, Irish Journal of Psychological Medicine, 1999, 16, 16–17. 17 Registration of Deaths in Ireland: A Statistical Nosology, Comprising the Causes of Death, Classified and Alphabetically Arranged (Dublin, 1864), iii (hereafter Statistical Nosology); First Annual Report of the Registrar-General of Marriages, Births and Deaths in Ireland, 1864, 7 (hereafter First ARRG). 18 Statistical Nosology, iii. 19 Royal College of Physicians of Ireland (RCPI)/2/2/1/1/1 Occasional Committee Minutes, February/March 1863. The Fellows of the RCPI were invited to comment on the civil registration bill in February 1863. A committee was formed to consider it. It expressed dissatisfaction with the legislation and emphasised five main concerns, which all centred around the responsibility placed on physicians for the accuracy of data under the headings of place of residence, age at death, date of last visit (especially for metropolitan practitioners), cause of death and duration of disease. It recommended that these data emanate ‘from sources external to the medical profession’ but their case had no impact on the subsequent legislation. 20 Anne Cameron, ‘Medicine, Meteorology and Vital Statistics: The Influence of the Royal College of Physicians of Edinburgh upon Scottish civil registration, c. 1840–1855’, Journal of the Royal College of Physicians Edinburgh, 2007, 37, 173–80 at 174–6. 21 An Act for registering Births, Deaths, and Marriages in England, 6 & 7 Will. IV. c.86, section 27. 22 Cameron, ‘Medicine, Meteorology and Vital Statistics’, 178–9. 1863 act, section 46. 23 http://www.who.int/healthinfo/statistics/indmaternalmortality/en/, accessed 8 December 2015. 24 Statistical Nosology, iii; First ARRG, 1864, 7. 25 William Farr, Report on the Nomenclature and Statistical Classification of Diseases for Statistical Returns (Oxford, 1856), 11. 26 James Mathew Duncan, On the Mortality of Childbed and Maternity Hospitals (Edinburgh: Adam and Charles Black, 1870), 6. 27 Peter Conrad, ‘Medicalization and Social Control’, Annual Review of Sociology, 1992, 18, 209–32. Conrad defines the process of medicalisation as ‘defining a problem in medical terms, using medical language to describe a problem, adopting a medical framework to understand a problem, or using a medical intervention to “treat” it’. 28 Anon, BMJ, 30 December 1876, 859. ‘The statistics of puerperal mortality are … more imperfect than are those of most other diseases. … In consequence of the present state of the law … it is impossible to arrive at the correct number of child-bearings’. 29 Jones, ‘Captain of all these men of death’, 68. 30 Report from the Select Committee on Midwives’ Registration, 1892 (289), xiv, 1, 25. 31 On criticism from the Brussels Congress on Hygiene, see Inter-Departmental Committee on Physical Deterioration, 45. On the provision created in 1926 and 1938, see Gayle Davis, ‘Stillbirth Registration and Perceptions of Infant Death, 1900–60: The Scottish Case in National Context’, Economic History Review, 2009, 62, 629–54. 32 Stillbirths Registration Act, 1994, http://www.irishstatutebook.ie/1994/en/act/pub/0001/print.html, accessed 19 June 2015. 33 Helen Andrews, ‘McClintock, Alfred Henry’, in James McGuire and James Quinn, ed., Dictionary of Irish Biography (Cambridge: Cambridge University Press, 2009), 839–40. 34 Dr McClintock’s contribution to the, ‘Report of the Dublin Obstetrical Society’, Dublin Journal of Medical Science, 1869, 48, 266. This meeting was conducted over eleven nights by 17 physicians in response to former Master of the Rotunda, Dr Evory Kennedy’s paper ‘Zymotic Diseases, as more Especially Illustrated by Puerperal Diseases’, Dublin Quarterly Journal of Medical Science, 1867, xliv, 514–21, which suggested radical structural reform of the Rotunda to stop the spread of puerperal disease. See also Andrews, ‘Kennedy, Evory’, in Dictionary of Irish Biography, 108–9. 35 Kennedy, ‘Zymotic Diseases’, 514–21. 36 Report of the Commissioners Appointed to Inquire into the Hospitals of Dublin. With appendices [2063], 1856, xix, 7. 37 First Annual Report of the Board of Superintendence of Dublin Hospitals, with appendices [2353, 1857–58] (hereafter BSDH). 38 S. D. Sessoms, ‘Puerperal Fever’, Journal of the National Medical Association, 1915, 7, 104–8. 39 Both cases occurred in Dublin North City No. 3 registrar’s district. 40 Section 46 of the Irish registration act, 1864, describes the procedure Registrars and Medical practitioners were to follow. An Act to provide for the better regulation of Births, Deaths, and Marriages in Scotland, 17 & 18 Vict. c.80 s. 41. 41 Forty-third ARRG, 1906, xvii, xxxii. See Appendix A for a complete list of the Registrar General’s annual reports which we have used in this paper. 42 ARRG, 1970 (Dublin, 1973), xliv. 43 First and Second Reports from the Select Committee on Death Certification, 46. 44 Stephan Curtis, ‘Midwives and their Role in the Reduction of Direct Obstetric Deaths during the late Nineteenth Century: The Sundsvall Region of Sweden (1860–1890)’, Medical History, 2005, 49, 321–50 at 326–7. 45 ‘An Address Delivered in the Section of Public Medicine’, BMJ, 1879, 2, 242. 46 Edward Higgs, ‘A Cuckoo in the Nest?: The Origins of Civil Registration and State Medical Statistics in England and Wales’, Continuity and Change, 1996, 11, 115–34, at 115–21. 47 J.K.L., Letters on the State of Ireland (Dublin: Coyne, 1825), 96; K. H. Connell, The Population of Ireland, 1750–1845 (Oxford: Clarendon Press, 1950), 2. 48 E. Margaret Crawford, Counting the People; A Survey of the Irish Censuses, 1813–1911 (Dublin: Four Courts Press, 2003), 15. 49 An Act for Marriages in England (6 & 7 William IV, c. 85); An Act for Registering Births, Deaths and Marriages in England (6 & 7 William IV, c. 85). 50 An Act for Marriages in Ireland, and for registering such marriages (7 & 8 Vict., c. 81) (hereafter Irish Marriage act, 1844). 51 Anne Cameron, ‘The Establishment of Civil Registration in Scotland’, The Historical Journal, 2007, 50, 377–97 at 378–9. Civil registration in Scotland commenced on 1 January 1855, with the enactment of an Act to provide for the better registration of births, deaths and marriages in Scotland (17 & 18 Vict., c. 80). Other Bills, in 1837, 1847, 1848 and 1849, had failed to pass into law. 52 Irish Marriage act, 1844, sect. 52. 53 Ibid., sects 57, 58. 54 First Report of the Registrar-General of Marriages in Ireland, under the Provisions of the Act 7 & 8 Vic. Chap. 81 [1130], 1850, xxv, 16–17. 55 The census of Ireland for the year 1851, part vi: General report, with appendix, county table, miscellaneous tables, and index to names of places [2134], 1856, xxxi, I, lix. Crawford, Counting the People, 21–2. Simon Szreter, Fertility, Class and Gender in Britain, 1860–1940 (Cambridge: Cambridge University Press, 1996), 85. 56 House of Commons Journal, ci, part. 2, 1211, 1222 (leave to introduce the Bill granted, 14 August 1846; Bill introduced, 17 August 1846); A bill for registering births, deaths and marriages in Ireland, H.C. 1846 (637), iii, 515, p. 1 (sects 1, 6). 57 Registration of births, &c., (Ireland) Bill H.C. 1859, session 1 (68), ii, 547. In 1860 Opposition bill introduced by Lord Naas, 8 May 1860. Government bill (number 2) introduced by Chief Secretary Cardwell, 10 May 1860. Cardwell’s bill was considered by a committee and amended, 21 May 1860. Both bills were withdrawn, 5 July 1860 (Commons jn, cxv, 230, 235, 257, 356); Hansard Parliamentary Debates, clvii (24 April 1860–6 June 1860), cols 886–7; clix (7 June 1860–20 July 1860), cols 209–10. In 1861 Government bill introduced 11 February 1861. Naas’s Bill introduced 22 February 1861. Select Committee’s modifications, 11 July 1861, which removed birth and death aspects, and amended Cardwell’s Bill with portions of Naas’s. Bill withdrawn, 22 July 1861 (Commons jn, cxvi, pp 37, 74, 354, 381; Hansard Parliamentary Debates, clxiv (28 June 1861–6 August 1861), col. 1285). In 1862 Hugh Cairns’ Marriage Bill introduced 19 February 1862, and Peel’s Births and Deaths Bill introduced on 20 February 1862. Both Bills withdrawn on 2 July 1862 (Commons jn., 58, 60, 305; Hansard Parliamentary Debates, clxvii (27 May 1862–7 July 1862), cols 622, 1313, 1321–2). 58 1863 Act, sect. 4; Irish Times, 9 September 1876. 59 1863 Act, sect. 17 (if a PLU boundary changed then the boundary of the SRD could be modified, to reflect the new union boundary). 60 Ibid., sects 18, 20, 22, 23. 61 Poor Law and Medical Charities (Ireland) Act 14 & 15 Vict c. 68. Geary, ‘The Medical Profession’, 189–206 at 190. Geary argues that the transition from medical charity to poor law was strongly opposed by ‘a substantial segment of the medical profession’ and adds that it ‘regarded any connection with the poor law system as socially and professionally degrading’. Any efforts to mount an ‘opposition campaign’, he also notes, were ‘undone by the Great Famine’. 62 Ronald D. Cassells, Medical Charities, Medical Politics; The Irish Dispensary System and the Poor Law, 1836–1872 (Rochester, NY, 1997), 78. 63 Laurence M. Geary, Medicine and Charity in Ireland 1718–1851 (Dublin: University College Dublin Press, 2004), 209–10. Anon, ‘The Local Government Board of Ireland and Dispensary Doctors’, BMJ, 1896, 1:1827, 54; Anon, ‘Irish Dispensary Doctors’ Grievances Deputation to the Chief Secretary’, BMJ, 1892, 1:1634, 880. Catherine Cox, ‘Access and Management: The Medical Dispensary Service in Post-Famine Ireland’, in Catherine Cox and Maria Luddy, ed., Cultures of Care in Irish Medical History (Basingstoke: Palgrave Macmillan, 2010), 57–78 at 60. 64 1863 Act. 65 Medical Officers Superannuation Act (Ireland), 1869 (32 & 33 Vict., c. 50). 66 1863 Act, sects 31–33. 67 Ibid., sects 36–38. 68 1863 Act; Irish Marriage Act, 1844. 69 Geary, ‘The Medical Profession’, 198–9. 70 Report of the Royal commissioners appointed to inquire into the sewerage and drainage of the city of Dublin, 1880 [C.2605]. 71 Thom’s Irish Almanac and Official Directory of the United Kingdom of Great Britain and Ireland for the year 1877 (Dublin: Thom’s Publications, 1877), 1070. 72 13 & 14 Car II c12, (1662). 73 Dublin city, covering 3,808 acres in 1871, had expanded to 7,911 acres by 1901 (Vaughan and Fitzpatrick, Irish Historical Statistics, Population, 29). 74 Ibid., 243, 247–8. 75 Sixteenth ARRG, 1879, 5–6. Public Health (Ireland) Act, 1878 (41 & 42 Vict., c. 52), sect. 191; Public Health (Ireland) Amendment Act, 1879 (42 & 43 Vict., c. 57), sect. 7. 76 First ARRG, 1864, 64; Second ARRG, 1865, 24; Third ARRG, 1866, 24; Fourth ARRG, 1867, 24; Fifth ARRG, 1868, 24; Sixth ARRG, 1869, 58; Seventh ARRG, 1870, 58; Eight ARRG, 1871, 56; Ninth ARRG, 1872, 56; Tenth ARRG, 1873, 54; Eleventh ARRG, 1874, 58; Twelfth ARRG, 1875, 58; Thirteenth ARRG, 1876, 58; Fourteenth ARRG, 1877, 58; Fifteenth ARRG, 1878, xix, 58; Sixteenth ARRG, 1879, 58; Seventeenth ARRG, 1880, 58; Eighteenth ARRG, 1881, 55; Nineteenth ARRG, 1882, 46; Twentieth ARRG, 1883, 46; Twenty-first ARRG, 1884, 48; Twenty-second ARRG, 1885, 52; Twenty-third ARRG, 1886, 62; Twenty-fourth ARRG, 1887, 62; Twenty-fifth ARRG, 1888, 70; Twenty-sixth ARRG, 1889, 70; Twenty-seventh ARRG, 1890, 70; Twenty-eight ARRG, 1891, 70; Twenty-ninth ARRG, 1892, 72; Thirtieth ARRG, 1893, 72; Thirty-first ARRG, 1894, 74; Thirty-second ARRG, 1895, 74; Thirty-third ARRG, 1896, 74; Thirty-fourth ARRG, 1897, 74; Thirty-fifth ARRG, 1898, 74; Thirty-sixth ARRG, 1899, 50; Thirty-seventh ARRG, 1900, 50; Thirty-eighth ARRG, 1901, 51. 77 Census of Ireland, 1871, pt ii, Vital Statistics, vol. ii, Report and Tables Relating to Deaths [C. 1000], 1874, lxxiv, ci. 78 On the cholera outbreak, see Ibid., c. 79 Report of Royal Commissioners to inquire into the sewerage and drainage, city of Dublin, 32. 80 Ciara Breathnach, ‘Handywomen and Birthing in Rural Ireland 1851–1955’, Gender & History, 2016, 28, 36–58. 81 On the division into regional divisions, see First ARRG, 1864, 18–20. 82 First ARRG, 1864, 18. 83 GRO dataset. 84 Rotunda Hospital Archive is held at the National Archives of Ireland (NAI), usage is by prior approval and strict appointment only. The collection is cited as follows NAI/PRIV/1263. 85 First Annual Report of the BSDH, 3. 86 ‘Report of the Dublin Obstetrical Society’, 290–4. 87 T. Percy Kirkpatrick and Henry Jellett, The Book of the Rotunda Hospital (London, 1913), 158. 88 Ibid., 158. Report of the Rotunda Hospitals for poor lying-in women, and for the treatment of diseases peculiar to women (Dublin, 1877), 5–6. 89 Robley Dunglison, Medical Lexicon: A Dictionary of Medical Science (Philadelphia: Henry C. Lea, 1868), 16. 90 NAI PRIV1263/2/22. 91 Loudon, Death in Childbirth, 49–84. 92 Dr W. J. Smyly, ‘The Maternal Mortality In Childbed’, The Lancet, 1900, 11, 385–8. 93 M. Best and D. Neuhauser, ‘Heroes and Martyrs of Quality and Safety: Ignaz Semmelweis and the birth of infection control’, Quality & Safety in Health Care, 2004, 13, 233–4. 94 Phil Gorey, ‘Puerperal Fever in Dublin: The Case of the Rotunda Lying-in’, in Lisa Marie Griffith and Ciarán Wallace, eds, Grave Matters: Death and Dying in Dublin 1500 to the Present (Dublin, 2016), 46–60 at 56. 95 Reports of Societies’, BMJ, 18 July 1885, 102–3 at 103. 96 Thirty-second report of the BSDH [6073] 1890, 12. 97 Seventh ARRG, 1870, 58. 98 Ibid., 11, 131. 99 Seventeenth ARRG, 1880, 48, 58. 100 Irish Times, 13 March 1894; Irish Times, 20 March 1894. 101 Twenty-seventh ARRG, 1890, 70. 102 Ibid., 1, 14, 19. 103 Annual report of the commissioners for administering the laws for relief of the poor in Ireland, including the twenty-fourth report under the 10 & 11 Vic., c. 90, and the nineteenth report under the 14 & 15 Vic., c. 68; with appendices, 1871 [C.361], 46. 104 Ibid., 15. ARLGB, being the ninth report under “the Local Government Board (Ireland) Act”, 35 & 36 Vic., c. 69. 1881 [C.2926] [C.2926-I], 15 (hereafter ARLGB); Nineteenth ARLGB for Ireland, 1890–91 [C.6439], 16. 105 Arthur Macan, Master of the Rotunda Hospital, to Dublin Hospital’s Commission, Report of the Committee of Inquiry, 1887 [C. 5042], 1887, xxxv, 89. 106 Attending on 30 September were M.G. (illiterate), who informed on six deaths, A.O’C. (literate), informing on three deaths, and M.S. (illiterate) who informed on one death. Attending on 31 December were C.M. (illiterate), informing on four deaths, A.O’C. (literate; five deaths), M.S. (illiterate; six deaths), K.B. (literate; three deaths), and L.W. (illiterate; one death) (GRO records, qr 3, Dublin North City, No. 2 records, p. 340; ibid., qr 4, Dublin North City, No. 2 records, 393–5). 107 Breathnach, ‘Handywomen and birthing’, 36–58. Ciara Breathnach, ‘“ … it would be preposterous to bring a Protestant here”: religion, provincial politics and district nurses in Ireland, 1890–1904’, in D. S. Lucey and Virginia Crossman, eds, Healthcare in Ireland and Britain 1850–1970: Voluntary, Regional and Comparative Perspectives (London, 2015), 161–80. 108 Gerard Fealy, A History of Apprenticeship Nurse Training in Ireland (Abingdon: Routledge, 2006), 24–8. 109 Ian Campbell Ross, ‘Midwifery’ in Ian Campbell Ross, ed., Public Virtue, Public Love: The Early Years of the Dublin Lying-in Hospital, the Rotunda (Dublin: O’Brien Press, 1986), 125–64 at 159. 110 Henry Jellett, The Book of the Rotunda Hospital: An Illustrated History of the Dublin Lying-in Hospital from its Foundation in 1745 to the Present Time (Dublin), 181. 111 Report of the Rotunda Hospitals (Dublin, 1880), 8. The 1886 report notes that Midwifery outcases started in 1874 with 95 cases, it almost doubled to 177 in 1875 and grew steadily in the subsequent ten years to 1892 cases as against 1807 admissions in 1886. Report of the Rotunda Hospital (Dublin, 1886), 7–9 (From 1885 Rotunda Hospital was dropped). 112 Stanley A. Seligman, ‘The Royal Maternity Charity: The First Hundred Years’, Medical History, 1980, 24, 403–18 at 417. Seligman raises suspicions about the two maternal mortality cases reported for London’s Royal Maternity Charity in 1856 out of 3,297 live births (of which 53 were of twins and three of triplets): ‘when the hospitals of London, Vienna, Paris, and Dublin were ravaged by childbed fever’. 113 Dublin Hospitals Commission. Report of the Committee of Inquiry, 1887, together with minutes of evidence and appendices, 1887 [C.5042] xx–xxii. 114 Alison Nuttall,‘Passive Trust or Active Application: Changes in the Management of Difficult Childbirth and the Edinburgh Royal Maternity Hospital, 1850–1890’, Medical History, 2006, 50, 351–72 at 353–4. 115 NAI /PRIV/1263/4. 116 The Nomenclature of Diseases (1st edn, London, 1869), vi, x. It was prepared by a joint committee (which included Farr) appointed by the Royal College of Physicians of London, and was deemed ‘suitable to England, and to all countries where the English language is in common use’. 117 Anon, BMJ, 30 December 1876, 858. Loudon, Death in Childbirth, 28. 118 BMJ, 30 December 1876, 858. 119 Tenth ARRG, 1875, 102–3. 120 The Nomenclature of Diseases (2nd edn, London, 1884), 9, 117. 121 Ibid. (3rd edn, London, 1896), 11; Ibid. (4th edn, London, 1906), 7. Ibid. (5th edn, London, 1918), 4). 122 Thirty-eighth ARRG, 1901, xviii, 18. 123 Ibid. 124 Jones, ‘Captain of all these men of death’, 69. 125 Foucault, ‘Society Must be Defended’. Appendix A. Annual reports of Registrar General First Annual Report of the Registrar-General of Marriages, Births and Deaths in Ireland, 1864 [4137], 1868–9, xvi (hereafter ARRG). Second ARRG, 1865 [C. 4], 1870, xvi. Third ARRG, 1866 [C. 130], 1870, xvi. Fourth ARRG, 1867 [C. 238], 1871, xv. Fifth ARRG, 1868 [C. 238], 1871, xv. Sixth ARRG, 1869 [C. 554], 1872, xvii. Seventh ARRG, 1870 [C. 785], 1873, xx. Eight ARRG, 1871 [C. 968], 1874, xiv. Ninth ARRG, 1872 [C. 1099], 1874, xiv. Tenth ARRG, 1873 [C. 1376], 1876, xviii. Eleventh ARRG, 1874 [C. 1495], 1876, xviii. Twelfth ARRG, 1875 [C. 1617], 1876, xix. Thirteenth ARRG, 1876 [C. 1937], 1878, xxii. Fourteenth ARRG, 1877 [C. 2301], 1878–79, xix. Fifteenth ARRG, 1878 [C. 2388], 1878–79, xix. Sixteenth ARRG, 1879 [C. 2688], 1880, xvi. Seventeenth ARRG, 1880 [C. 3046], 1881, xxvii. Eighteenth ARRG, 1881 [C. 3368], 1882, xix. Nineteenth ARRG, 1882 [C. 3795], 1883, xx. Twentieth ARRG, 1883 [C. 4149], 1884, xx. Twenty-first ARRG, 1884 [C. 4538], 1884–85, xvii. Twenty-second ARRG, 1885 [C. 4801], 1886, xvii. Twenty-third ARRG, 1886 [C. 5153], 1887, xxiii. Twenty-fourth ARRG, 1887 [C. 5537], 1888, xxx. Twenty-fifth ARRG, 1888 [C. 5844], 1889, xxv. Twenty-sixth ARRG, 1889 [C. 6147], 1890, xxiv. Twenty-seventh ARRG, 1890 [C. 6520], 1890–91, xxiii. Twenty-eight ARRG, 1891 [C. 6787], 1892, xxiv. Twenty-ninth ARRG, 1892 [C. 7255], 1893–94, xxi. Thirtieth ARRG, 1893 [C. 7535], 1894, xxv. Thirty-first ARRG, 1894 [C. 7800], 1895, xxiii. Thirty-second ARRG, 1895 [C. 8236], 1896, xxiii. Thirty-third ARRG, 1896 [C. 8560], 1897, xxii. Thirty-fourth ARRG, 1897 [C. 8949], 1898, xviii. Thirty-fifth ARRG, 1898 [C. 9491], 1899, xvi. Thirty-sixth ARRG, 1899 [Cd. 295], 1900, xv. Thirty-seventh ARRG, 1900 [Cd. 697], 1901, xv. Thirty-eighth ARRG, 1901 [Cd. 1225], 1902, xviii. Thirty-ninth ARRG, 1902 [Cd. 1711], 1903, xvi. Fortieth ARRG, 1903 [Cd. 2222], 1904, xiv. Forty-first ARRG, 1904 [Cd. 2673], 1905, xvii. Forty-second ARRG, 1905 [Cd. 3123], 1906, xx. Forty-third ARRG, 1906 [Cd. 3663], 1907, xvi. Forty-fourth ARRG, 1907 [Cd. 4233], 1908, xvii. Forty-fifth ARRG, 1908 [Cd. 4769], 1908, xi. Acknowledgements This work was supported by the Irish Research Council, Interdisciplinary Research Project Grant [2013–3]. We gratefully acknowledge the cooperation of the General Register Office, which has kindly shared its data with the project. © The Author 2017. Published by Oxford University Press on behalf of the Society for the Social History of Medicine.

Journal

Social History of MedicineOxford University Press

Published: Feb 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 12 million articles from more than
10,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Unlimited reading

Read as many articles as you need. Full articles with original layout, charts and figures. Read online, from anywhere.

Stay up to date

Keep up with your field with Personalized Recommendations and Follow Journals to get automatic updates.

Organize your research

It’s easy to organize your research with our built-in tools.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve Freelancer

DeepDyve Pro

Price
FREE
$49/month

$360/year
Save searches from Google Scholar, PubMed
Create lists to organize your research
Export lists, citations
Read DeepDyve articles
Abstract access only
Unlimited access to over
18 million full-text articles
Print
20 pages/month
PDF Discount
20% off