Abstract
Defining the effect of an exposure of interest and selecting an appropriate estimation method are prerequisites for causal inference. Current understanding of the ways in which an association between heat waves (ie, consecutive days of extremely high temperature) and an outcome depends on whether adjustment was made for temperature and how such adjustment was conducted is limited. In this paper we aim to investigate this dependency, demonstrate that temperature is a confounder in heat-wave-outcome associations, and introduce a new modeling approach with which to estimate a new heat-wave-outcome relationship: E[R(Y)|HW = 1, Z]/E[R(Y)|T = OT, Z], where HW is a daily binary variable used to indicate the presence of a heat wave; R(Y) is the risk of an outcome, Y; T is a temperature variable; OT is optimal temperature; and Z is a set of confounders including typical confounders but also some types of T as a confounder. We recommend characterization of heat-wave-outcome relationships and careful selection of modeling approaches to understand the impacts of heat waves under climate change. We demonstrate our approach using real-world data for Seoul, South Korea. Our demonstration suggests that the total effect of heat waves may be larger than what may be inferred from the extant literature. An R package, HEAT, has been developed and made publicly available. This article is part of a Special Collection on Environmental Epidemiology.