Research Interests

My current work uses advanced statistical methods to model associations of biomarkers with kidney disease and cardiovascular disease in HIV infected persons, including cluster analysis to identify phenotypes of risk, Bayesian model averaging for variable selection, and penalized regression to handle problems with large numbers of potentially correlated predictors. In addition, I have developed CKD risk prediction algorithms in two large cohorts: 21,590 HIV-infected men from the Veterans Health Administration cohort and in 24,877 participants from the Reasons for Geographic and Racial Differences in Stroke (REGARDS) study. Also in the VA cohort, I published a paper using inverse probability treatment weighting showing that tenofovir exposure was independently associated with increased risk of kidney disease. My research involves use of marginal structural modeling to handle both drug channeling bias and loss to follow-up, issues which arise in the study of kidney disease in HIV infected patients.