266 – Statistics in Disease Mapping and Spatial Epidemiology: New Insights and New Frontiers

Multivariate Gaussian Markov Random Fields, MCARs, and Bayesian Disease Mapping

At the invitation from JSM 2014 Program Chair for the Section on Statistics in Epidemiology (SIE) to organize an invited session on Statistics in Disease Mapping and Spatial Epidemiology (to be sponsored by the Section on SIE ), I have worked to put together the following presentation program for an invited session of four speakers. We are very fortunate to have four outstanding statisticians committed to present their most recent research work at the session, they are Drs Martin Kulldorff, Andrew Lawson, Lola Ugrate, and Brain Neelon. Drs. Kulldorff, Lawson, and Ugrate are senior statisticians with Full Professor ranking. They are leading statisticians in the field (of statistics in disease mapping and spatial epidemiology) and each has made significant and scholarly contributions to the development of important statistical models and related methods of inference for analysis of spatial data with applications in disease/health mapping and epidemiology. Briefly, Professor Kulldorff is a world renowned biostatistician and the developer of the well known SaTScan freeware. His presentation will also be a much needed introduction to the highly anticipated development of the SaTScan II, the second generation of the SaTScan. Professor Lawson is also an internationally well known Biostatistician and a prolific researcher and writer on Bayesian disease mapping and statistical methods for spatial epidemiology. He is a WHO advisor in Disease Mapping and Risk Assessment, and has a wide range of publications in this area, including 5 books. He will present new development of conditional autoregression ( CAR) models that could have considerable potential to offer models for data of increased complexity and to extend the scope of the CAR disease mapping methodology and application. With more than two decades research work on the development of statistical methods for disease mapping and spatial epidemiology, Prof Ugarte is a Spanish statistician and a leading expert on the subject and related applications, particularly in the development and evaluation of semi-parametric spatiotemporal disease mapping models for small area disease/health mapping. It is anticipated that her presentation will offer new insights into the roles and applications of spline-based spatiotemporal models ( with spline smoothing over spatial, temporal, and spatiotemporal dimensions) for mapping and analysis of increasingly complex and rich spatiotemporal disease mapping data. To provide a platform for younger and relatively junior statistician to present innovative work for statistics in epidemiology, we will have Dr. Neelon, an Assistant Professor from the Duke University, to present his research work on the subject. Dr. Neelon's research on multivariate spatial mixture models for areal data has the potential to offer new tools for mapping and analysis of spatial data on large lattice. In addition, Neelon's presentation will illustrate statistics in spatial epidemiology in the context of social science research and application. I anticipate the invited presentations to attract JSM 2014 attendees with related statistics interests and applied background and/or experience in health science, social science (such as economics and education), political science, and behavioural science in which a broad spectrum of outcomes has a spatial dimension and the notion of spatial epidemiology is becoming increasingly relevant and is gaining increasing recognition, particularly in this data rich era and for its global scale. Speaker 1: Martin Kulldorff, PhD, Full Professor, Biostatistician Department of Population Medicine, Harvard Medical School and Harvard Pilgrim Health Care Institute Email: martin_kulldorff@hms.harvard.edu Tentative title: Space-time scan statistics for detecting anti-microbial resistance outbreaks. Speaker 2: Andrew B. Lawson, Full Professor of Biostatistics Department of Public Health Sciences, Medical University of South Carolina, Charleston, SC USA Email: lawsonab@musc.edu Tentative title: A Shared neighbor Conditional Autoregressive Model for small area spatial data Speaker 3: Lola Ugarte, Full Professor of Statistics Department of Statistics and Operational Research, Public University of Navarre, Pamplona, Spain Email: lola@unavarra.es Tentative title: Semi-parametric spatiotemporal models for disease mapping Speaker 4: Brain Neelon, Assistant Profesor Department of Biostatistics and Bioinformatics, Duke University Email: neel003@duke.edu Tentative title: A Multivariate Spatial Mixture Model for Areal Data: Examining Regional Differences in Standardized Test Scores