Precision of rate estimation under uniform interval censoring

Medical events are often recorded not at their time of occurrence but rather at a subsequent examination. In survival analysis, this is called interval censoring. This paper presents a relationship between the sampling error of the estimated event rate and the number of examinations, which are assumed to be evenly spaced throughout the study. This relationship should help balance the cost of additional examination rounds against the precision gained. Copyright (c) 2008 John Wiley & Sons, Ltd.