Application of simulation techniques for estimating duration of multiple sclerosis derived from prevalence-formed cohorts.

Comparisons of the average duration of multiple sclerosis derived from a prevalence survey of cases alive in 1970 in a low-risk area (Los Angeles County, California) and a high-risk area (King and Pierce Counties, Washington) suggest that patients in the high-risk area had a longer duration of disease than patients in the low-risk area. Because this finding was unexpected and because the underlying population of these two areas has been increasing at different rates, two simulation models were developed to estimate the duration of multiple sclerosis from a prevalence survey, taking into account the effects of changes in the population over time. Comparison of the durations derived from the two simulation studies suggested that underestimation of true backward recurrence time from the simulation studies was similar for the two study areas. Thus, the observed differences in duration between the two areas probably reflect the course of disease rather than differences in rate of growth of the two populations. These studies demonstrate the usefulness of simulation studies in estimating disease duration from cohorts derived from prevalence surveys of non-stable populations.