Please join us online for a talk by , a 鈥渟tatistics virtuoso鈥� on the faculty of the Dartmouth Institute for Health Policy & Clinical Practice and Dartmouth College. He will present "Causal Estimation of Intervention Spillover in a Cluster-Randomized Stepped Wedge Trial: Lessons from a Physician Network" on Wednesday, September 16, at 1:30pm Central Time. For access to this seminar, contact series administrator Cierra Streeter.听
James O鈥橫alley, PhD, FASA, holds Dartmouth鈥檚 Peggy Y. Thomson Professorship in the Evaluative Clinical Sciences. He is also Professor of Biomedical Data Science and Adjunct Professor of Computer Science at Dartmouth College, where he directed the Program in Quantitative Biomedical Sciences (2018鈥�2020) and the PhD Program in Health Policy and Clinical Practice (2019鈥�2024). His methodological interests in statistics span social network analysis, multivariate hierarchical models, causal inference (using instrumental variables), and Bayesian inference. He has published 320 peer-reviewed research papers across statistics, health policy, medical and other journals. He chaired the Health Policy Statistics Section (HPSS) of the American Statistical Association (ASA) in 2008, co-chaired its International Conference in 2011, and serves as Associate Editor for Statistics in Medicine听and听Observational听Studies. In 2011 he received the HPSS Mid-Career Excellence听Award, in 2012听he听became an听elected fellow of the ASA, and听he听was the 2019 recipient of the International Society of Pharmacoeconomics and Outcomes Research (ISPOR) Health Economics and Outcomes听Research Excellence in Methodology Award. In 2025 he was a co-recipient of听Geisel School of Medicine's Research Excellence Award for Senior Faculty in Foundational Science.听
Abstract
Often motivated by concerns about contamination of control subjects, stepped wedge听cluster-randomized trials assign interventions to distinct clusters (e.g., hospitals) and protect against confounding by randomizing the timing of intervention delivery. However, when trial units are embedded in professional networks that span clusters, such trial designs are vulnerable to spillover from intervention subjects to control subjects. We first develop potential outcomes-听based definitions of the causal average direct and the causal average indirect effects of the intervention. We then use a longitudinal cluster-randomized trial and contemporaneous physician professional networks for a national provider organization in the United States to model and estimate the direct, indirect and direct-indirect interaction effects of an Advance Care Planning (ACP) intervention. We show that the direct effect measures the intervention's impact in a counterfactual world absent contamination, the indirect effect quantifies the spillover effect from intervention to control subjects, and the interaction term assesses whether spillover modifies the intervention's effect, a phenomenon known as contamination. In addition, we exploit the staggered nature of the stepped听-wedge听cluster-randomized trial design to test whether intervened-on peers impact the sustainability of the intervention over follow-up and whether spillover reinforces or attenuates the intervention. In our empirical analyses, we find that without adjusting for network structure the intervention showed no significant effect on ACP billing. However, in the network adjusted analyses we detected a large spillover effect with strong evidence of contamination prior to a subject being intervened on and a significant intervention reinforcement effect post intervention. These findings suggest that the diffusion of intervention behaviors through the physician network were primarily driven by a large physician spillover effect that modified the direct effect at the time of intervention and subsequently over follow-up.听