Abstract
Several investigators have discussed various reasons for the large and consistent difference in the relative risk estimates per unit exposure to air pollution derived from time-series and cohort study designs. This article considers whether the two study designs are estimating fundamentally different parameters even though both are commonly expressed in terms of relative risk coefficients. A frailty model for air pollution mortality is proposed to illustrate some of the theoretical behaviors of heterogeneous populations exposed to time-varying pollution patterns and their implications for the kinds of summary measures widely used in air pollution risk assessment.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have