Department of Biostatistics Seminar/Workshop Series

Analysis Methods for Medical Product Post-Marketing Surveillance

Michael Matheny, MD, MS, MPH

Assistant Professor of Medicine, Assistant Professor of Biomedical Informatics, Vanderbilt University

Wednesday, March 17, 1:30-2:30pm, MRBIII Conference Room 1220

Persons interested in applied statistics, statistical theory, observational cohort analysis methods, statistical process control, epidemiology, health services research, medical product surveillance, and statistical computing.

DELTA (Data Extraction and Longitudinal Time Analysis) is a software program designed to retrospectively and prospectively analyze medical product outcome data for outliers. The system utilizes a database backend and a web-based user interface with an array of tabular and graphical outputs as well as automated email notification of alerts. The software uses a number of frequentist and Bayesian methods, and includes SAS, WinBUGS (through SAS) and R (in development) interface modules.

This presentation will focus on methods and theoretic framework relating to the development and validation of the software, including a number of statistical process control techniques, considerations for selection of event rate expectations and observational cohort risk-adjustment, and if we have time, conclude with the use of risk-adjusted Sequential Probability Ratio Testing for institutional and provider outlier detection.