From the mid-nineteenth century onwards we have achieved enormous gains in life expectancy leading to what may be called the great transformation of our societies. This still ongoing process has had, beyond any doubt, a decisive impact on the way we live our lives, both collectively as societies, and individually, as human beings. The consequences have not only been increases in longevity, but also taller height of younger birth cohorts, and strong increases in economic productivity by an ever healthier population. The mortality decline is therefore not only one of the most important processes in human history but also a precondition for the formation of our modern post-industrial society.
The magnitude of the change in life expectancy is huge. We have every reason to assume that since the Middle Ages until the mid-nineteenth century, life expectancy has been fluctuating around the age of 30-40 years, after which it rose to its current position of approximately 80 years or even over in only 150 years. What is most relevant and striking for historians and historical demographers is the fact that a large part of that rise in life expectancy took place long before the arrival of modern day curative medicine, such as antibiotics and mass vaccination programmes. These latter instruments, which in most western countries appeared in the 1940s and 1950s, finally enabled mankind to successfully win the fight against infectious diseases.
The crucial point is that already before the 1940s large advances in life expectancy had been made. The reduction of mortality due to infectious diseases, such as diphtheria, scarlet fever, tuberculosis and whooping cough, in the Netherlands began from approximately the 1870s onwards. Between 1870 and 1939 life expectancy in the Netherlands increased from age 37 to age 67. This massive proportional change presents an important explanatory challenge for historians and historical demographers: what drove this important change in mortality and life expectancy and how did it come about? What explanatory factors can help us understand those great leaps forward? And which diseases were driving the decline in mortality, and how can that be explained?
A new approach to understand health progress
These are not easy questions to answer. Mortality and health are the outcomes of complex and multi-causal processes. In this historic extension of life expectancy beyond age 30-40 many factors have played a role, ranging from increased personal hygiene, public health policies, higher incomes, improved nutrition, behavioural change, infant feeding practices, and improved education for the majority of the population. In a large research project currently being carried out at the Radboud Group for Historical Demography and Family History we aim to increase our understanding of this complex process. The strength of this project, titled Lifting the burden of disease, lies in the data that we use: the individual level cause-of-death data for the city of Amsterdam between 1854 and 1940. These are unprecedentedly rich data listing the cause of death for every individual that died in the city in this period, including the address of the deceased, his/her age, sex, marital status, and occupation. We can use the city of Amsterdam as a unique historical laboratory, in which we investigate the role of economic factors, public hygiene, and cultural factors in the nineteenth- and early twentieth-century reduction of mortality.
One of the first results from this project relate to the issue of infant mortality. In the nineteenth century the first year of life was enormously dangerous, and that was no different in Amsterdam. In the 1850s one-fifth of all new-born infants died before reaching age 1. Obviously, this huge death toll weighs heavily on life expectancy from birth. Above age 1 mortality hazards were much lower, so that bringing down the mortality hazards for infants, and also young children between age 1 and 5, would therefore have large effects on life expectancy. The figure displayed below charts the development of infant mortality in Amsterdam between 1856 and 1926 by cause of death. It uses rates rather than counts so that we can take into account the number of infants born in those years. This result already enables us to draw a number of important and interesting conclusions. The high level of infant mortality of 200 or more per 1,000 new-borns continued until the mid-1880s, after which a steady decline occurred which gained speed in the 1890s, ultimately dropping below 50 deaths per 1,000 new-borns at the end of the period. Until the 1890s infant mortality was also influenced by epidemics striking the city, such as cholera in the 1850s or small pox in the early 1870s.
Regarding the causes of death, we can see that until approximately 1900 three disease categories are important which contain vague historical disease terms. These are ‘congenital and birth disorders’, ‘weakness’ and ‘convulsions’. The first two categories both contain disease terms referring to weakness and congenital weakness. The third one, ‘convulsions’, is often thought of as related to water- and food-borne diseases. However, our seasonality tests do not support such an assumption, as we do not find any evidence for the expected summer peak. Towards the end of the period investigated here we find that vague disease terms have disappeared, due to the improvement of diagnostic practices. Another important conclusion is that from approximately 1910 onwards water- and food-borne diseases are largely disappearing. No doubt factors such as improved hygiene, piped water for instance, have played an important role in this development.
In the research project Lifting the burden of disease we will combine these individual level cause-of-death data with a host of other data sets, such as data on the consumption of piped water, the spread of sewerage in the city, the level of urban crowding, the rental value of the houses to determine social economic status of individuals and households, data on wages and prices to determine economic development, and data on spatial location within the city to determine how the changing urban landscape impacted upon disease and mortality patterns.
Moving forward: avoiding the pitfalls of aggregated data
This approach in which we combine different avenues of research will contribute to an unprecedented precision in the attempt to understand not only the changing epidemiological profile of the city but also to improve our understanding of its determinants. Finally, the fact that we can determine the cause of death at the level of individual diseases makes it possible to circumvent the use of aggregated cause-of-death data at the city level, or even higher levels such as provinces or countries. These aggregated data confront researchers with serious limitations because of their dependency upon nineteenth-century disease classifications which distinguish only few and often vague disease categories and contain no information on how diseases were actually classified. Adopting an individual level approach such as we present here, will provide an important stimulus to the field of the history of mortality and epidemiological change.
Full professor (Radboud University Nijmegen / Maastricht University)
Visit personal website