STATISTICAL METHODS FOR DISEASE PREVENTION TRIALS

点击翻译按钮获取中文摘要

This continuing project aims to develop and evaluate improved methods for the design, conduct and analysis of Disease Prevention and Risk Factor Intervention Trials. A continuing major emphasis will be placed on the development and evaluation of methods for the analysis of correlated failure time data as may arise in prevention trials having multiple disease outcomes, or in risk factor intervention trials with behavioral or disease outcomes by virtue of group randomization. Estimating equations for marginal hazard ratio parameters and for cumulative hazard correlation parameters will be adapted to apply to common baseline hazard models, and estimation procedures for parameters in semiparametric marginal hazard function models and semiparametric covariance rate models will be developed and evaluated. Summary indices that may be used as an aid to assessing benefits versus risks in prevention trial monitoring will be -proposed and studied, including composite univariate indices and multivariate indices. Various additional trial monitoring topics in multiarm and factorial prevention trials will also be considered, along with related estimation issues. Various topics in the analysis of community intervention trials will be addressed, including estimating equation, bootstrap and rank-based methods, with particular emphasis on group randomized trials in which the number of randomization units is fairly small. Methods for extracting additional information from prevention Intervention trials will also be identified and evaluated, primarily in the context of prevention trials in which the investigators have ongoing responsibilities. These include methods for explanatory analyses that aim to identify the elements of multi-faceted interventions responsible for any observed disease benefit or risk; methods for extrapolation of results beyond the specific randomized comparisons, based on the possibility of introducing clinical trial-based bias correction terms into the analysis of related observational data; and methods for the identification and use of auxiliary endpoint data to strengthen tests and estimates, and for the identification of surrogate endpoints to streamline subsequent trials of related interventions. Collectively, the proposed research has the potential to increase the efficiency and reduce the cost of prevention trials, and enhance the scientific and public health knowledge gained.