The diagnostic test accuracy of a screening questionnaire and algorithm in the identification of adults with epilepsy

Summary Objective Accurate estimates of the incidence and prevalence of epilepsy allow us to better assess its societal impact. Prevalence and incidence studies often use unvalidated screening tools resulting in estimates of uncertain accuracy. We present the Canadian Longitudinal Study on Aging—Epilepsy Algorithm ( CLSA ‐ EA ) as well as the results of our validation study designed to estimate the diagnostic accuracy of this epilepsy ascertainment algorithm. Methods We administered English or French‐language versions of the CLSA ‐ EA questionnaire to a consecutive sample of participants from a population‐based cohort of 50,000 individuals aged between 45 and 85 years at baseline, as well as a consecutive sample of individuals from an epilepsy‐enriched general neurology clinic. Every participant was also assessed by a study neurologist who, blinded to the results of the CLSA ‐ EA , determined whether the person had epilepsy or not. Results We recruited 242 consecutive participants, 34 of whom were diagnosed with epilepsy by a study neurologist. The sensitivity and specificity of the CLSA ‐ EA for a lifetime history of epilepsy were 97.1% and 98.1%, and for active epilepsy were 100% and 98.6%, when we defined a positive screening test result as a positive response to the antiepileptic drug question and either the single self‐report diagnosis or any of the symptom‐based questions. Significance The CLSA ‐ EA was found to have a high sensitivity and specificity for the identification of adults with a lifetime history of epilepsy and active epilepsy. Although validation in other settings and age groups is required, the future application of this algorithm to population‐based studies such as the CLSA should help to ensure more accurate estimates of the prevalence and incidence of epilepsy in the general population when a physician assessment is impossible.