The focus of this project is development and refinement of statistical procedures for the design and analysis of cancer screening and related studies. Problems under investigation include development and comparison of data analysis methods, assessment of case-control studies for screening evaluation, development of models of cancer screening, and approaches to the analysis of categorical data. Properties of case-control studies in the context of screening evaluation are being considered. Screening effect as estimated from a case-control study within a randomized trial was compared with the trial result as a standard, initially using data from the HIP study. Alternative definitions of cases-controls and exposure were assessed. A stage-shift cancer screening model was developed which uses the number of cancers diagnosed by stage in screened and control groups to estimate the number of cases shifted to earlier stages as a result of screening and the mortality impact. Application to the HIP data indicated that the bulk of the mortality reduction resulted from earlier detection of stage I and II cancers. A method called Periodic Screening Evaluation was created which uses data from individuals screened at different ages over a short term with no control group to estimate the impact of long term periodic screening starting at different ages. Research into surrogate endpoints which might be valid proxies for mortality in screening studies focused on stage shift, stage rate and survival variables. The most promising surrogate is the rate of advanced stage disease. Regression models for doubly sampled categorical data with nonignorable sampling mechanisms and partial validation were studied. These focus on the relation between covariates and true response in the categorical data setting where two samples are available, one containing data on covariates and a surrogate response, the other data on true response, surrogate response and perhaps covariates.
