摘要：The aim of event history analysis is to study the occurrence and timing of events, such as premature psychotherapy termination or drinking onset, and to relate the probability of event occurrence to relevant covariates. In the social sciences, the timing of the event is often measured discretely by using time intervals, which implies the exact timing of event occurrence is unknown. The optimal number of subjects and time intervals need to be decided upon in the design stage of a trial with discrete-time survival endpoints. This paper shows how the optimal design depends on the underlying survival function and the costs to include a subject relative to the costs to take a measurement. Furthermore, the effects of attrition on the optimal design are studied. An example on drinking onset illustrates the proposed methodology. Keywords: longitudinal study , optimal design , study duration , sample size , time intervals
关键词：longitudinal study; optimal design; study duration; sample size; time intervals
Loading...
