The purpose of this project is to provide consultation on statistical and epidemiological methodology in the design, interpretation, and evaluation of clinical trials of diagnosis, treatment, and prevention of cancer, and other studies requiring this kind of expertise. For some studies the Section provides full statistical support, including development of detailed study plans; assistance in the design of appropriate study forms; supervision of randomization (for trials); collection, processing, and editing of data; performance of interim analyses during the progress of the study; preparation of progress reports; final analysis of study data; and collaboration in the preparation of scientific papers. During the past year the Section has continued to collaborate extensively on the Community Intervention Trial for Smoking Cessation (COMMIT). Key activities during this year were design and implementation of the final cohort and cross-sectional surveys, and planning the statistical analysis for major endpoints, involving the exploration of statistical techniques for adjusting observed differences using baseline prognostic factors and for handling missing data using baseline covariates and patterns of intermediate endpoints. The Section has continued to provide full statistical support for the randomized clinical trials of multimodality treatment conducted by the Brain Tumor Cooperative Group. Other important activities this past year include consulting on design issues in large-scale randomized prevention trials relating to breast cancer and prostate cancer; analyses of data from feeding studies measuring intakes of carotenoids; and collaborating on a physician practice study, involving a three-arm randomized design, of the adoption and use of the NCI Primary Care Nutrition Guide.
