The Limits and Possibilities of Cause of Death Categorisation for Understanding Late Nineteenth Century Mortality
Year: 2022
Type: article
Abstract: Over the last decade, the number of large-scale databases in research on the great mortality decline and the epidemiological transition that has taken place since the eighteenth century has increased enormously. Moreover, the advances in Information and Communication Technology (ICT) and quantitative methods have greatly enriched the uses that can be made of such datasets. Yet, comparative research is severely hampered because diagnostics and nosology have varied over time and across countries, regions, localities and even among doctors. As a result, historical cause of death registration and classifications can be problematic in many respects and require careful interpretation. The introduction of this special issue centres on the challenges posed, but also the opportunities offered, by the use of historical cause of death statistics in mortality research. In this introduction, we first discuss the academic debate in which most historians using cause of death data situate themselves, and why these types of data can help illuminate the substantive issues of that particular field. Following that, we present a brief overview of the different cause of death classification systems that researchers have used to analyse the nineteenth-century cause of death data focussing on the pitfalls and possibilities of each system. This is followed by a discussion of the four articles in this special issue. Some authors make use of ICD-10 (the 10th revision of the International Classification of Diseases), others combine different systems or construct a new one specific to their historical source or research aims. We compare and critically assess their use to examine historical changes in mortality and cause of death patterns. Finally, we also make a plea to develop interdisciplinary efforts in this field. The long-run changes in mortality have so far been the almost exclusive domain of historical demographers. The field would benefit from strong interdisciplinary cooperation with medical historians and epidemiologists. Long-run changes in mortality and life expectancy count among the great revolutionary changes of the past 200 years.1 Substantial reduction in mortality caused life expectancy to increase from an average of as low as 30–40 years in the early to mid-nineteenth century to as much as 84 years in some countries of the western world today. In tandem with these mortality changes, we find radical changes in epidemiological profiles. The type of diseases people are suffering from today, and the causes of death people die of, have changed in important ways. In the western world, infectious diseases have given way to cardiovascular diseases and various types of cancers, the so-called epidemiological transition.2 Causal factors threatening our health today have changed accordingly. Whereas in the past, mortality risks were greatest for infants and young children, today, the burden of disease and mortality is coming down primarily in the highest age categories. In other words, the patients, doctors are seeing today, are vastly different from the ones who were calling on their help in the nineteenth century. These entangled and interrelated changes in mortality and disease are still not fully understood. A large part of these transformations took place long before the arrival of modern-day curative medicine, such as antibiotics and mass vaccination programmes, which in most western countries only appeared in the 1940s and 1950s. Debates are continuing as to what caused these improvements in life expectancy before the 1940s, whether it is increases in living standards and hence nutritional intakes, improvements in public and/or personal hygiene, behavioural change, infant feeding practices or higher levels of education for certain groups within the population. This is a complex puzzle to resolve, which requires multi-disciplinary approaches spanning the fields of history and historical demography, medical history and historical epidemiology. It also requires innovative data and data technologies, ideally incorporating cause of death data at the individual level. Cause of death data at the individual level will greatly enhance our understanding of mortality patterns and changes over time. Knowing what disease someone died of, will crucially elucidate our understanding of why that person died in that particular place, at that particular moment in time and at that particular age. It will enhance our capacity to better capture what American medical historian Günther Risse once called the ‘shifting ecology of disease’, the complex and dynamic interactions between biological and non-biological factors impacted upon changing disease patterns.3 Aggregated cause of death statistics, by comparison, pose the risk of revealing only spurious trends in shifting patterns of death and disease. This problem is exacerbated by the complex interrelated changes occurring in medical science, in social and economic contexts and changing bureaucratic procedures, by state or other authorities, surrounding the drafting of causes of death. However, we wish to underline that these aggregated statistics should not be ignored by historians. In fact, in this special issue, one of the papers, by Nynke Van den Boomen, demonstrates ways how to test and probe these types of aggregated data to make them fruitfully speak to us on the nineteenth-century mortality decline. It should be clear, however, that historical cause of death data is not unproblematic, whether they come in the form of individual-level data, without being obscured by historical nosologies, or in the aggregated format. They can never be used in historical research without extensive internal data checking through complex routines assessing patterns, deviations and anomalies in the use of disease terms, for example, over time and/or location, by age, gender, disease categories, etc. The ICT and data revolutions also make this an easier job to execute, especially when datasets are available for millions of individuals over long periods.4 The challenges involved in the use of historical cause of death data and classification systems have already been discussed before by historical demographers and medical historians.5 Since then, the data revolutions taking place in many countries, with digital techniques making large numbers of a great variety of source material easily available to researchers, have expanded the possibilities for careful and critical use of cause of death data. Cause of death data can now be compared with contemporary medical reports and medical knowledge, as well as information on how registration practices were organised. Attempts to identify and classify individual diseases and causes of death go back a long way in history.6 Although some groupings of diseases were already apparent in the humoral theory of the ancient Greeks, the first classification of diseases is generally credited to the French physician François Bossier de Lacroix, better known as Sauvages, who published in 1763 in his Nosologia Medica a system listing 10 major classes of diseases and 2,400 individual diseases. Still, the systematic involvement of physicians in these attempts has emerged only very recently, that is, from the late nineteenth century onwards. Before that time, mainly state and city authorities were involved, and in some countries, most notably the Scandinavian countries, from the eighteenth century onwards religious authorities were the main actors and bodies driving and determining registration practices of diseases and causes of death. Religious authorities had already begun to record causes of death systematically in the seventeenth century. This practice became even more widespread in the eighteenth century. Local parishes in many European countries kept registers of vital events (births, marriages and deaths) to monitor their populations and the degree of religiosity of their flock. In particular, in the Nordic countries, these registers were well kept and provided a wealth of information. For instance, in Norway, from the early decades of the nineteenth century onwards, the priests were also obliged to register the cause a person had died of. To fulfil their duties adequately, the priests also received some medical training, and by the mid-nineteenth century, it appears that the parishes had managed to set up a complete recording of causes of death in the burial registers. However, the variation between parishes and priests was likely quite large, so comparability between regions and communities within the country was nearly impossible. Hence, Norwegian national authorities began to impose regulations on doctors and priests as well as an official nomenclature of diseases in 1861.7 That list provided 122 standardised causes of death and included instructions on how to use the Norwegian and Latin terms. This development was initiated and inspired by the first Statistical Congresses taking place in 1853 and 1855, in Brussels and Paris, respectively. Similar developments were taking place in other western countries. The development of a cause of death registration occurred in most cases against the background of a growing governmental interest in collecting statistical data on all kinds of demographic, economic and social processes. Medical professionals and hygienists, in particular, stressed the need for objective measures of health, including empirical research, standardisation of data collection and increased dissemination of data. In Belgium, for instance, a national initiative for the compulsory cause of death registration was taken in 1851 in response to a series of epidemics in the 1840s. Besides the cause of death, specific data had to be collected using pre-printed forms with the name of the deceased, surname, sex, occupation, age, marital status, and date and place of death. Although the cause of death had to be determined by a doctor, it was common practice for civil servants to fill out the certificates, usually relying on family members or neighbours to provide the cause. In large cities, however, a doctor was appointed for the task.8 Also, in Sweden and the Netherlands, the central state began to issue regulations requiring that the cause of death be recorded by a physician. In Sweden, that was the case in 1860, whereas the Netherlands followed a few years later in 1869 as a result of various local initiatives by physicians.9 This not a from all some of this as a of In the of the a national classification system of categories, with of them of disease and the the numbers of that had received medical In Belgium, members of the for up a list of causes of death which as the first nomenclature in These national classification systems were a of such as or and other more for that be more such as or initiatives were taking place at the national the growing in the of medical statistics also had an In at an in Paris, efforts to at an classification to make the cause of death recording across all countries were and in the Classification of The first of diseases and an classification of It was a of the classification of causes of death used by the city of which was a of and The classification was on of between diseases and to a particular was from French other and use the Since then, the has been almost taken medical and new it is by the and is the most used statistical classification system for diseases in the The the on it is almost as as the The of in the of has in historical research. To across time and place, have been to and one is to a of diseases to as as that each cause of death is the from one to possibilities focussing on specific causes of death that were over time or the to a with a some have their classification French demographers and for instance, in the late a system for aggregated data by and for the the of causes of death from one revision to their has made it to causes of death data for countries such as and the Netherlands across the by other of the since the the to the medical and of the time. In the historian a for individual-level cause of death data, which in with other the is on a the one by using and by on disease As in different case the causes of death to their of and and other and better the of death. the classification is very the much of use for research of the of the epidemiological transition when diseases the mortality cause of death data are also to the European called which historical medical historians and historical in the most to health changes across time and began in as a focussing on hence the the history of In the has to to communities in This has increased the number of and the number of regions and countries within that can be of the main of the is to construct an historical system for disease terms, called which can be used across countries and and for approaches the eighteenth century to the present The with disease patterns is as the system is an of the the has been and is now being on a set of in different European The will also a set of classifications to be used for different age groups and research The and classification systems will be made available as to the research These will comparative between European countries over long and will researchers to make important to the of long-run changes in mortality and epidemiology. The four articles in this special issue make use of cause of death data which are taken from different types of death certificates, cause of death as well as from different countries in make use of large-scale one of them (the even an the others or of these is focussing on changes in the history of mortality and from infant mortality to and from and to and The on the cause of death data how these data can important to more in the field of medical the history of and how these over time. The first in this issue by Nynke den a of how to test a nineteenth-century cause of death classification before using the data in historical research. The to the and of the first national classification system used to register all causes of death in the Netherlands occurring between and This register of a series of aggregated statistics which all mortality data by disease and a for without medical The whether and how this source material be in a in historical and epidemiological research for the of infant This is through a to anomalies and deviations between and In the case of the Netherlands, this is a as the health was also at the and these likely have the registration of causes of death as well as the of diseases within their of the nineteenth-century nomenclature to make it for the of infant This to age is an important to other who wish to historical or contemporary The of the extensive by Van den that pitfalls in the material are in particular to medical such as and diseases of the and and that is that medical and the the aggregated as well as and the causes of death, and that in this and were dynamic processes. This not to the use of these types of statistics for historical research. However, issues a and to the and history of these types of material before using it in the of Whereas the was on a historical classification system in the of the nineteenth century, the by and is at different the ICD-10 and the by This system the main as by between and diseases. that historical causes of death are not as causes of death to the standards of medical science, the authors very set out how these data can be the set out in the by Van den only use historical cause of death data a of the data and the taken in the research Moreover, careful should be given to the specific time and place to which the research are as well as the medical of that The main of this is to a test of the classification systems on a with the causes of death for individuals that had died in the city of in the years careful standardisation of the cause of death were historical terms. the not quite is that when an of the ICD-10 this a number of cases in the that is for causes of death. For this number is even also quite to The a better job with only of cases and for The authors that when with large historical databases it is to use a on to the of different As part of that they also the classification system by in This demonstrates the that a degree of is in research on the historical cause of death data, and a for new to the field. this also demonstrates is the great of individual-level cause of death data. These types of data generally have a higher degree of compared to the aggregated statistics that Van den has to with and which the to individual diseases various classification systems of The by and in this special issue is also on individual-level cause of death data, this time taken from nineteenth-century death for the large city of The to to the history of European by and why died the of the nineteenth century. most occurring at or in one of the and what people die some health more in at with occurring at the authors to the of the system in from a place people primarily to die to a place they were to be main is to that history as an is an for such a given that it a long of for the and the as far back as the century. the mid-nineteenth century, the was a major medical of with a of 10 and In to this large the city also a great number of and The death that have been used for this are of a wealth of the death had to be and by a doctor before a be The doctor was also to go to the to examine the which the of the cause of death of information on the To the causes of death, the authors for a nineteenth-century classification system that was used in at that time. This is by research is need to to epidemiological patterns or of diseases. they wish to to disease doctors were at that time when whether or not to in the The authors examine all occurring in the city in and 1869 to the death in a or at to the the authors are to analyse the of age, gender, and the of family members in the on the of in or at They find that who died in were or coming from the and with few to This the that were still a place for the the and the However, they not this as a with today, the ones who die in at are still the ones with family and with However, on an of the causes of death, the authors also find that the doctors of the were at that time already some degree of medical They appeared to have a special interest in diseases to the and their not make a between the and the This in the by the authors that to some the of the and the doctors was the of the medical historians that causes of death are social not only in the past but also today. This that causes of death are within the of and social in is the of the by and The on how of disease in the diseases between and because of biological which the authors to the biological of For instance, the biological are in early are likely to die of diseases Moreover, mortality generally also as they are to infectious diseases to and However, in epidemiological patterns also as a result of the of and disease are through of and and through in and other of and In the European past, all in these patterns even to primarily for and young using individual-level cause of death data for in and in the Netherlands, the authors wish to how of diseases the way mortality and cause of death were for and young The way causes of death were in the past varied greatly between time and However, we have on whether that was also the case between and we that, for instance, diseases such as and were as in the of death, the and registration of the causes of death were also by of is The individual-level cause of death data these researchers have at their them to construct their and out specific diseases of They to make use of an of the by researchers and which we have already at in this special issue, in the on The data them to on in diseases that are most likely to have been by biological and behavioural infectious and more and an mortality for over in the Norwegian and the For and in the age most notably were driving the mortality but the these patterns were different for and in have been the for have been risks to their duties which kept them and for and other family The communities were also for because of the higher risks of and The authors that these were from the set of individuals who have to or other infectious diseases they had not died In that these different groups of causes were as an interrelated set of For risks were in and which were easily for doctors and priests The great of individual-level cause of death data is also in this The authors a and of local registration practices in the To what were registration and the of the causes of death in the communities is the they For instance, given the of in the mortality of and the authors great to the different but also ways in which was Finally, they that the priests who were the recording of the causes of death were very in the disease they were and were also at to use a variety of disease their physician in the issues central to research on the nineteenth-century mortality the articles in this special issue have provided better and the registration and classification of causes of death, be it by priests or civil The authors of the four articles are to the critical use of cause of death data, and of how researchers can make such data The on and the comparative on and the Netherlands that the opportunities in mortality research can be greatly when the individual-level causes of death are The is first of because individual disease can be without being historical The individual-level data researchers to construct their Although this the comparability of the it the authors to test these classifications in various ways. should be in of the research objective of the Hence, in the on the authors for a contemporary nosology as they to to the medical of doctors in the nineteenth century. For the other such a would be The comparative on and the Netherlands, for instance, has a particular interest in infectious the classification they use is that In this health changes can be A why these individual-level causes of death data greatly increase opportunities is their The causes of death registers a of such as date of death, age, sex, marital status, and of the This a of for research changing epidemiological over time, as well as the economic and these Still, aggregated statistics should not be ignored by In fact, in this special issue, Van den ways to test and probe these types of aggregated data to make them fruitfully speak to us on the nineteenth-century mortality decline. The the and need of with medical historians. changes in medical as well as changes in the of diagnostics and the practices of are for using historical on causes of death. It is that medical of diseases in the past was not as extensive or as as it is today, and that changed over the past 200 years or Moreover, techniques and better recording practices can between as causes of death are by their and local it is important to how the medical professionals and the diseases at the time and how this As the registration of causes of death is with the social history of
Source: Social History of Medicine
DOI: 10.1093/shm/hkac040
Authors: Angélique Janssens, Isabelle Devos
Language: English
Subfield: Demography
Field: Social Sciences
Domain: Social Sciences
Open Access status: hybrid
Estimated APC paid: $3,287
OpenAlex ID: W4293858993