In medical diagnostic testing problems, the covariate adjusted receiver operating characteristic (ROC) curves have been discussed recently for achieving the best separation between case and control. Due to various constraints, the sample sizes for some covariate values are not large enough to support reliable direct estimations of ROCs for all the underlying covariates of interest. The authors develop statistical methods to effectively utilize the information provided by the data using the semiparametric exponential tilting models. In these models, the density functions from different covariate levels share a common baseline density and the parameters in the exponential tilting component reflect the difference among covariates. The new covariate adjusted ROC is much smoother and more efficient than the nonparametric counterpart. A simulation study and a real data application are reported.

Zhong Guan is an Associate Professor of Statistics in the Department of Mathematical Sciences at Indiana University South Bend. He received his PhD in statistics from The University of Toledo in 2001. Before joining the faculty at IU South Bend in 2004, he was a postdoctoral research associate at Yale University School of Medicine. His research focuses on empirical likelihood method semiparametric and nonparametric models and bioinformatics.

Jing Qin is a Mathematical Statistician at the Biostatistics Research Branch in the National Institute of Allergy and Infectious Diseases. After graduating from the University of Waterloo (1992), he spent one year as a postdoctoral fellow at Stanford University before joining the faculty at the University of Maryland. Before moving to the National Institute of Health (2004), he worked at the Memorial Sloan-Kettering Cancer Center for 5 years. Dr. Qin's research interests include the empirical likelihood method, case-control study, length bias sampling, econometrics, survival analysis, missing data, causal inference, genetic mixture models, generalized linear models, survey sampling and microarray data analysis. He was elected as a Fellow of the American Statistical Association in 2006. He was a winner of the Pierre Robillard award in 1993 (The best statistical thesis defended in Canada in 1992).

Biao Zhang is a Professor in the Department of Mathematics and Statistics at the University of Toledo. He completed his PhD in statistics at the University of Chicago in 1992. Before coming to the University of Toledo, he was a visiting assistant professor at the State University of New York at Buffalo. From 2000-2001 Biao was a visiting associate professor in the Department of Biostatistics at the University of Michigan. His areas of research interest include empirical likelihood, missing data analysis, ROC curve analysis, and semiparametric statistical inference under density ratio models.

The award-winning paper will be presented by Dr. Qin at the 41th Annual Meeting of the Statistical Society of Canada to be held in Edmonton, Alberta, May 26-29, 2013.