Special Topics in Env. Health: From Climate Events to Health Evidence: Research Methods in Practice
EH 550
Subject & Catalog Number
Course Information
Description
How do we know that a heat wave, a hurricane, or a wildfire smoke episode harmed health, by how much, and for whom? This intensive January course teaches the research methods used to turn climate events into rigorous health evidence. The course is organized by method rather than by topic: each session pairs an exposure-assessment or causal-inference approach with real climate events, and is taught by Harvard faculty and guest researchers who use these methods in their own work.
We kick-off the course with an introductory lecture (recorded) ahead of the first session on climate epidemiology, the causal questions it asks, and how exposure timing (acute versus chronic, short versus long windows) shape study design. Course topics include satellite- and machine-learning-based exposure assessment and linkage to health data; methods for multiple, co-occurring climate hazards across the lifespan; time-series regression and distributed lag non-linear models (DLNMs) with multi-site pooling; case-crossover, difference-in-differences, and other quasi-experimental designs; Bayesian hierarchical and small-area models; excess mortality estimation; extreme event attribution science; and the translation of evidence into policy, clinical practice, and intervention evaluation.
Each session consists of a lecture, followed by a hands-on lab using the R programming language. The course culminates in a group capstone in which students apply a method from the course to a climate event and present and defend their analytic choices and code.
Class Notes
Taught by: John S. Evans, Lisa A. Robinson
The course is designed for MS and PhD students in Environmental Health and other departments who intend to develop, apply or interpret risk assessments in support of environmental decision making and policy.
It asks each student to consider how they would decide which environmental exposures pose risks so large that they must be controlled, and which are so small that they may safely be ignored. For those which must be controlled, it asks them to develop an approach for deciding how much control is necessary and how to distinguish circumstances where they must act now from those where uncertainty is so large that control decisions should be deferred (in order to allow researchers time to conduct studies intended to improve understanding of the risks involved). The distinct roles of science and values are discussed. We consider the complexity introduced when individual values (health vs. wealth, risk neutrality-aversion-proneness) differ substantially across the affected population.
The course introduces the concepts and framework of economic evaluation; reviews approaches for valuing risks to health and longevity; discusses methods for incorporating consideration of the distribution and equity of control costs and health benefits. It examines the interpretation of evidence from epidemiology and toxicology in risk assessment. It considers the sources of uncertainty inherent in estimates of reference doses and cancer potencies derived from animal studies, in vitro analyses and structure-activity relationships, and effect estimates from epidemiological studies and introduces approaches for characterization and analysis of the propagation of uncertainty. With this knowledge in hand, we explore the impact of uncertainty on decisions, both about emissions control and about research needs.
We seek a set of principles, solidly grounded in theory, and a toolbox of approaches for applying these principles, which will support more informed decision making about environmental controls.
Prerequisites: EH 206: Foundations of Environmental Health OR in previous years -- RDS 500: Risk Assessment and EH 510: Exposure Assessment OR permission of instructor.
Available for Harvard Cross Registration
