A SIMPLE PROGRAM TO CREATE EXACT PERSON-TIME DATA IN COHORT ANALYSES

Background. Before disease rates can be calculated a tabulation of the length of follow-up for each person in the cohort has to be made. In complicated analyses such tabulations are often stratified by many cha racteristics, some which show no change with time, such as gender or y ear of birth, and some which do change with time, such as age or cumul ative exposure. Available computer programs often restrict the way the se tables can be made, particularly when handling time-dependent varia bles. Methods. This paper presents a simple computer program which cal culates the length of follow-up for each person in a study. Results. P erson-time data can be tabulated by a large number of variables using this method. This program is extremely flexible in the way that time-d ependent variables can be created, can categorize observations by any unit of person-time, and will run on a range of platforms including a personal computer. Conclusions. This method should simplify the task o f creating person-time data for analyses of disease rates in epidemiol ogical studies.