Confounding in Short-term and Long-term Studies
Utrecht University, Netherlands
Abstract Number: 446
Preference: Invited Plenary Speaker
Last modified: January 5, 2010
Working Group: sq7
In population studies, the health effects of air pollution are typically studied on the temporal scale or the spatial scale.
Time series studies focus on the relationships between short-term variations in air pollution, and equally short-term variations in health parameters such as daily mortality, hospital admissions or symptom status in panels of patients with respiratory or cardiovascular disease. In such studies, only factors which vary over similarly short time scales can act as confounders. These include weather variables such as temperature and humidity, but also influenza epidemics and the like.
Long-term studies usually compare the incidence of disease or death in populations living in areas with contrasting air pollution exposure. In such studies, any other determinant of disease or death incidence can act as a confounder of the association between air pollution and health, and careful consideration to these factors is needed at the design and/or analysis stage of the study. Typical factors include but are not limited to smoking and dietary habits, occupational exposure, individual and neighborhood level socioeconomic status etc.
The presentation will focus on some practical examples from the published literature of important potential confounders of temporal and spatial associations between air pollution and health.