Methodological approaches to estimating prevalence of disability: The case of paralysis

Accurately estimating the prevalence of disabling conditions is a necessary but frequently overlooked step in developing sound, appropriate and effective policies, programs and services targeted at subpopulations. Over the last several decades, various methods have been used to estimate the prevalence of paralysis including registries, hospital discharge datasets, data from suppliers of durable medical equipment and population-based surveys. Estimates from each of these data sources have frequently resulted in widely divergent results. This paper addresses two issues. First, what accounts for the differences in estimates resulting from these various methods? Second, what are the relative strengths and weaknesses of estimation methods that have been used to estimate the prevalence of paralysis? These issues are discussed within the context of a multi-year effort to address shortcomings in historical estimates of paralysis involving researchers from universities across the country and federal agencies.