Scroll to top

Modeling a Bivariate Residential-Workplace Neighborhood Effect when Estimating the Effect of Proximity to Fast-Food Establishments on Body Mass Index

October 28, 2016

James O’Malley, PhD
Professor of Biostatistics, Department of Biomedical Data Science
Professor, The Dartmouth Institute for Health Policy and Clinical Practice at the Geisel School of Medicine at Dartmouth

About the Presentation: Hierarchical modeling is the preferred approach of modeling neighborhood effects. Given residential and workplace location indicators, a bivariate (residential-workplace) neighborhood random effect whose correlation quantifies the extent that a neighborhood’s residential effect correlates with its workplace effect may be specified. However, statistical model-estimation software typically does not allow correlations between the effects of different clustering variables. We develop a Bayesian model with a bivariate random effect for neighborhood and an accompanying estimation procedure. The model accounts for individuals who reside or work in multiple neighborhoods across their observations, individual heterogeneity (random intercepts, random slopes), and serial correlation between observations on the same individual. We apply the models to the motivating Framingham heart study linked food establishment data to examine whether (i) Proximity to fast-food establishments is associated with Body Mass Index (BMI); (ii) Workplace neighborhood exposure associations are larger than those for residential exposure; (iii) Residential neighborhood exposure associations correlate with workplace neighborhood exposure. For robustness, we perform analyses under multiple specifications of the prior distribution of the neighborhood random effect covariance matrix. In addition, we show that allowing for time varying neighborhood membership, individual heterogeneity, and serial correlation across time yields more precise neighborhood level estimates.

About the Presenter: Dr. O’Malley’s methodological research interests have centered on social network analysis, causal inference (comparative effectiveness research), multivariate-hierarchical modeling, and previously the design and analysis of medical device clinical trials. In these, he has developed novel statistical methods, often involving novel use of Bayesian statistical methods, to solve important methodological and applied problems in health policy and health services research, including the evaluation of treatments and outcomes of health care in multiple areas of medicine. This has led to advances in interventional cardiology, vascular surgery, measuring quality of health care, mental health and long-term care.

Dr. O’Malley’s primary teaching activities currently include teaching a statistics class to medical and health policy students, guest lectures in other courses, and presenting short courses at conferences and other forums. His educational activities include a large number of invited seminars; mentoring colleagues, post-doctoral fellows and students; and service to the statistics profession. In addition, his service to the statistics profession includes eight-years at the forefront of the Health Policy Statistics Section of the ASA, associate editorships at both Statistics in Medicine and Health Services and Outcomes Research Methodology, and as reviewer for over 20 respected academic journals.

In recognition of many of Dr. O’Malley’s above contributions, he was elected fellow of the American Statistical Association (ASA) in 2012 and awarded the 2011 Mid-Career Award by the Health Policy Statistics Section of the ASA (a single award is given biannually).