STATISTICAL ISSUES IN THE EARLY DETECTION OF DISEASE

点击翻译按钮获取中文摘要

DESCRIPTION: The major goals of this project are to investigate and develop statistical models and procedures related to the early detection chronic diseases. The research is motivated from current problems in the early detection of cancer --- especially breast cancer. The research is divided into three major areas: (i) planning clinical trials to evaluate the benefits of early detection of disease; (ii) estimation of parameters (sensitivities, pre-clinical sojourn time distributions) during the course of a trial as well as after the last scheduled examination; and (iii) develop and extend quantitative methods to aid public health planners in the planning of early detection programs to benefit larger populations. The major issues in planning large public health programs for the early detection of disease are: at which ages should individuals begin to participate, the time between examinations and the number of examinations. Methods will be developed for evaluating any proposed public health program. The economic and health implications of applying these models are potentially large. The major issues in planning early detection clinical trials are to take advantage of the special features of these trials; e.g., number of exams, spacing between exams, the changing incidence of disease with age, and the history of each subject with regard to earlier exams. Experimental design problems arise as in some disease area control groups (no exams) may not be feasible. The major issues in parameter estimation relate to having more than one detection modality, possibly resulting in different pre-clinical sojourn time distributions, ascertaining if sensitivity and sojourn time distributions are age dependent, compliance, and investigating if different detection modalities are independent.