Professor of Medicine (Biostatistics), HMS

Dr. Rosner’s research activities are focused on several areas currently including (a) longitudinal data analysis, (b) analysis of clustered continuous, binary and ordinal data, and (c) methods for the adjustment of regression models for measurement error.

Dr. Rosner’s work on longitudinal data analysis has involved comparative analyses of several longitudinal data models on the same datasets involving serial measurements of blood pressure and pulmonary function over a 20-30 year period. These analyses are important since they enable one to predict future blood pressure levels based on current levels, which is important for screening purposes.

Dr. Rosner’s work on the analysis of clustered binary data has involved analyses of opthalmologic and otolaryngologic data where clustered data are the rule rather than the exception. Clustered data also appear frequently in coronary regression studies where multiple diseased arteries are sampled from the same subject at different points in time.

His work on measurement error methods has been applied mainly in nutritional epidemiology where dietary exposures are measured with error and one wants to assess the relationship between cancer and cardiovascular disease outcomes and nutrient intake, while adjusting for measurement error.

Ph.D., 1971, Harvard University