What was Killing Babies in Trondheim? An Investigation of Infant Mortality Using Individual Level Causes of Death, 1830–1907

This paper examines infant mortality amongst newborns in Trondheim city, 1830–1907, working specifically with individual level cause of death data. Findings show that infant mortality in the city started to drop from 1895, primarily as a result of a decline in post-neonatal mortality. At the start of the decline air-borne diseases accounted for nearly half of the deaths, and water-food borne for around one third. The drop was predominantly driven by a decline in these two causal groups, and seasonal fluctuations became less pronounced. Because of the fall in post-neonatal mortality, the relative risk of dying amongst neonates rose towards the end of the period. Although 'convulsions' accounted for 50–70% of all infant deaths between 1830 and 1860, this cause had faded away to near insignificance by the beginning of the 1900s. Here we aim to assess the extent to which this particular aspect of decline can be explained by alterations to official instructions regarding registration and in registration practice itself. This article proposes that the decline in deaths from 'convulsions' can be explained by a relabelling of such deaths into 'congenital and birth disorders' amongst neonates, and a mix of 'water-food borne' and 'air-borne diseases' amongst post-neonates. This argument is supported by the fact that the timing of the decline corresponds with the introduction of cause of death certificates issued by medical practitioners, and which most likely resulted in fewer causes of death being reported by lay informants who could only offer vague symptoms rather than informed diagnoses.