Identification of prevalent patients with epilepsy using administrative data: the Tuscany experience.

Overall prevalence of epilepsy ranges from 4 to 10 cases per 1000. Italy lacks recent epidemiological studies on large populations. In the present study, prevalence of epilepsy has been assessed in Tuscany, an Italian Region with 3,750,000 habitants, implementing an algorithm based on administrative data from the Regional Information Health System. To identify patients with epilepsy, we used at least one the following criteria: (a) at least one EEG and at least two dispensations of any antiepileptic drug (AEDS) at a minimum distance of 12 months; (b) at least two dispensations of one "specific" AED (authorized for use only for patients with epilepsy) at a minimum distance of 12 months; and (c) hospital admission for epilepsy or recurrent relapses (cod. ICD-IX-345.*). This algorithm was validated through comparison with lists of true patients with epilepsy and subjects without neurological disorders (gold standard). 35,950 cases were identified. Total crude prevalence was 9.6/1000. Prevalence increased in older patients up to 16/1000 without gender differences. Overall sensitivity of the algorithm was 87.3%, and specificity was 99.9%. This algorithm identifies patients with epilepsy with acceptable sensitivity and specificity and can be used to assess the burden of disease and for monitoring health services.