This report presents different postmarket surveillance methods applicable to the distributed data setting with rare outcomes. The workgroup 1) reviewed the statistical and epidemiology literature on methods for survival outcomes that incorporate adjustment for confounders using a causal inference approach; 2) assessed the methods’ applicability to Mini-Sentinel and made recommendations on which strategies were best suited for use within this setting, where rare events, a distributed data environment, and sequential testing introduce new complications; and 3) evaluated, via simulation, the most promising existing approaches and new approaches tailored to the Mini-Sentinel setting.
Andrea J. Cook, PhD; Biostatistics Unit, Group Health Research Institute, Seattle, WA and Department of Biostatistics, University of Washington, Seattle, WA
Azadeh Shoaibi PhD, MHS; Ram C. Tiwari, PhD; Rima Izem, PhD; Rongmei Zhang, PhD; Center for Drug Evaluation and Research, FDA, Silver Spring, MD
Susan R. Heckbert, MD, PhD; Department of Epidemiology, University of Washington, Seattle, WA
Lingling Li, PhD; Department of Population Medicine, Harvard Pilgrim Health Care Center and Harvard Medical School, Boston, MA
Robert D. Wellman, MS; Biostatistics Unit, Group Health Research Institute, Seattle, WA
Jennifer C Nelson, PhD, Biostatistics Unit, Group Health Research Institute and Department of Biostatistics, University of Washington, Seattle, WA