Applying synthetic control methods for multiple treatments to study the health impacts of extreme weather events

Abstract

Title: Applying synthetic control methods for multiple treatments to study the health impacts of extreme weather events Extreme weather events such as wildfires, heat waves or flooding are expected to increase in frequency and intensity in our changing climate. In this context, it is particularly important to precisely and accurately quantify the causal effects of such events to inform adaptation strategies and identify vulnerable populations and territories. Quasi-experimental designs (QED) that capitalize on natural experiments can be particularly robust to obtain valid causal estimates under specific identification assumptions. Especially, QED could be useful and flexible to disentangle the effects of extreme weather events from underlying secular changes affecting the health outcomes or other time fixed unmeasured confounders. In the recent years, many methodological developments regarding QED have been proposed in other settings highlighting their flexibility and good performance regarding bias and precision. For example, synthetic control methods (SCM) constitute a recent family of methods that have been used to estimate the effect of various policy changes or intervention on health outcomes. In the recent years, many developments allowed the consideration of multiple treatments and/or heterogeneity in the expected outcome across space and population subgroups. Such recent SCM developments can be very useful to understand how extreme weather events impact population health across space and time especially when the nature of such events and population susceptibility changes across time. In this introduction of the symposium, we will first present recent developments in SCM and how they could be applied to various extreme weather events such as wildfires or heat waves. We will then present an extension of the SCM and focus on an empirical example related to the impacts of the wildfires on the risk of respiratory hospital admissions in California.