Since joining the Biostatistics Service in 2006, Cami Sima has collaborated with researchers in the health outcomes, epidemiology and clinical services. Currently, her primary clinical collaborations are in the thoracic oncology, thoracic surgery and head and neck areas. She has been working with investigators from these groups on the design and analysis of prospective and retrospective studies, with a focus on evaluating treatment strategies, exploring various biomarkers, and developing prediction and risk models. Dr. Sima's independent research interests are primarily motivated by her ongoing collaboration with the health-outcomes group. Population-based observational data sources (such as SEER-Medicare, State cancer registries/Medicaid linked data, or National Cancer Center Network) are rich resources for building evidence-based knowledge on the effectiveness of medications, devices and other relevant interventions. However, their utility is limited by inherent selection bias, non-measured confounders and non-ignorable missing data. Dr. Sima is interested in applying and extending causal inference methods such as instrumental variables and quasi-experimental designs to facilitate integrated analyses of observational datasets. She has been working with Drs. Schrag and Panageas in developing methods to characterize patterns of chemotherapy diffusion in the USA using the SEER-Medicare linked registries.