Eric Schneider (London School of Economics): “Atmospheric Pollution and Child Morbidity from Respiratory Diseases: Evidence from the London Foundling Hospital 1897-1914.”

Abstract:
Atmospheric pollution has been shown to exacerbate respiratory infections and contribute to mortality in both modern and historical settings. However, existing historical studies focus on mortality as an outcome, which makes it difficult to pinpoint the precise mechanism through which pollution affects health outcomes: does it increase the spread of disease or increase the intensity of sickness events already in progress? We also cannot rely solely on estimates of these relationships in modern populations given that improvements in nutrition, hygiene and medical technology may have altered the relationship between pollution and morbidity over time. Finally, focussing on mortality potentially skews our understanding of how pollution affected other health outcomes such as child growth. Repeated respiratory infections may have slowed child growth, but most of these sickness episodes were not fatal and therefore would not appear in the mortality record, making it possible that the disease composition of sickness events and deaths differed substantially. Thus, to understand how historical pollution affected population health, we must study historical pollution and sickness directly.

The paper tests the influence of short-run, exogenous spikes in atmospheric pollution on disease outcomes for children. I use Hanlon (2022)’s data on fog events in London to proxy short-run spikes in pollution since fogs trapped existing pollution close to the ground. Hanlon collected weekly meteorological data as well as the number of light and heavy fogs in a given week. The morbidity data, covering both the incidence (new cases) of disease, sickness duration (time in the infirmary) and the appearance of complications, is reconstructed from rich, individual-level sickness information recorded by an orphanage in London, the Foundling Hospital, from 1897 to 1914. The data is drawn from reports from the institution’s infirmary which cared for sick children between the ages of c. 5 and 15. The infirmary records include over 6,000 sickness events from 235 different diseases (classified into ICD10 codes) for c. 1,000 children.

Preliminary results confirm the importance of atmospheric pollution for morbidity from respiratory diseases. While there is some evidence that pollution events could contribute to the spread of infection (catarrh and tonsillitis), the rarity of some diseases (bronchitis/bronchopneumonia) and the intermittent outbreaks for others (influenza and measles) prevents definitive conclusions for these diseases. However, there is strong and clear evidence that atmospheric pollution influenced sickness duration for a wide range of respiratory diseases (catarrh, tonsillitis, influenza and bronchitis/bronchopneumonia). Thus, pollution affected the severity of sickness events, prolonging them and increasing their costs to health regardless of the mortality consequences.

Applying these results to the health transition in London is complicated by contradictory evidence of long-term trends in pollution. Some argue that pollution began to fall from the 1890s as people adopted cleaner coal for heating and shifted to gas for cooking (Clay and Troesken 2011). However, the introduction of coal-fired power plants throughout central London may have countervailed some of these improvements, and certainly by the 1950s when instrumental records of pollution are available, pollution levels in London were extremely high (Hanlon 2022). Thus, declines in mortality from respiratory diseases occurred in spite of the longer run trends in atmospheric pollution.