This project provides consultation concerning statistical issues for COMMIT, a large-scale community-based study intended to promote smoking cessation among heavy smokers. Biometry Branch staff initially devised the basic design for the study~eleven matched pairs of communities with one member of each pair chosen at random for intervention and the other serving as a control. The study was designed to detect a 10% difference in the smoking quit rate between the intervention and control communities. Staff of the CDT Section have been actively involved in all meetings of the Steering Committee and have analyzed and presented data to the independent Policy Advisory Committee. The COMMIT intervention ended in 1992, and the final surveys were performed in 1993, finishing in January 1994. The four principal final surveys were: 1) Endpoint Cohort Survey~The full cohorts of heavy and light-to-moderate smokers were contacted in the spring of 1993 to determine the self- reported, 6-month cigarette smoking cessation rates in each of the 22 communities. This measure is the primary endpoint of the COMMIT trial . 2) Cotinine Validation Study~After the completion of the Endpoint Cohort Survey, the """"""""quitters"""""""" in the heavy smoker cohort were re-contacted for participation in this study. Saliva samples were collected from eligible and consenting participants for cotinine measurement. The study is designed to measure cessation misrepresentation rates between the COMMIT intervention and comparison communities. 3) Evaluation Cohort Survey~Cohorts of 400 adults from each of the 22 communities were contacted a third and final time in the spring of 1993 to measure the population-wide impact of COMMIT on intervention awareness, participation, and the decline of the social acceptability of smoking. 4) Final Prevalence Survey~A cross-sectional sample of 3000 adults in each COMMIT community were surveyed to determine adult smoking prevalence. The survey was conducted between August 1993 and January 1994 and is providing a measure of the trial~s secondary endpoint, smoking prevalence. The Branch has planned the statistical analyses for the major COMMIT endpoints. Other work has involved approaches for the weighting and adjustment of results from the prevalence surveys, with consideration of probability of inclusion of each telephone number in the sampling frame, number of telephones in each household, nonresponse, and age-sex distribution for each community based on 1990 census data. The major effort this year has been the analysis of the major endpoints and preparation of the two major publications reporting the outcome.
