Amyotrophic Lateral Sclerosis in three Italian regions: prevalence and incidence based on health administrative databases

Abstract Background Studies of the epidemiology of amyotrophic lateral sclerosis (ALS) are numerous, with incidence and prevalence estimates varying widely depending on population characteristics, geographical areas, and available data sources. The aim of this study was to estimate prevalence and incidence of ALS in three Italian regions (Latium, Tuscany, and Umbria) using health administrative databases. Methods ALS patients residing in Latium, Umbria and Tuscany were identified through an algorithm based on data from three different health administrative databases: hospital discharges, exemptions from health care co-payment, and emergency departments (study period 2014-2019). We calculated crude, age- and gender specific prevalence estimates on December 31, 2019 and standardised incidence rates of ALS by region, year, and sex between 2014-2019. Moreover, using a clinical dataset available in the Lazio region, we calculated the proportion of individuals correctly identified as ALS cases by the algorithm in this region. Results A total of 1,031 persons affected by ALS, aged ≥ 18 years at the prevalence day were identified: 408 cases in Tuscany, 546 in Latium, and 77 in Umbria. Diseases specific co-payment exemptions accounted for the biggest contribution for case detection, ranging between 24.7% in Umbria and 30.6% in Latium. The algorithm produced standardised overall ALS prevalence rates similar between the three regions, varying between 12.31/100,000 in Tuscany, 11.52/100,000 in Latium and 9.90/100,000 in Umbria. The 5-year crude rates were higher in men, and in people aged 65-79 years. Among 310 patients included in the clinical dataset, 263 (84.8%) were identified by the algorithm based on health administrative databases. Conclusions ALS prevalence and incidence in three regions of Central Italy are rather similar but slightly higher than those previously reported. This finding is plausible given that previous results relate to at least ten years ago and evidenced of increasing trends. Overall, our results encourage the use of administrative data to produce occurrence estimates, useful to both, epidemiological surveillance and research and healthcare policies.