The Effect of Different Algorithms on Prevalence of Attention Deficit Hyperactivity Disorder and Autism Spectrum...

To assess the effect on prevalence estimates of using different algorithms to identify children with attention deficit hyperactivity disorder (ADHD) and autism spectrum disorder (ASD) in healthcare data. Three algorithms were developed and run on administrative/research data in Finland, France (Haute Garonne), Italy (Emilia Romagna), Norway and Wales: (1) ≥ 1 ADHD or ASD diagnoses recorded in specialist settings, (2) ≥ 2 ADHD or ASD diagnoses recorded in primary care and (3) ≥ 1 prescription for medication to manage ADHD. Prevalence rates per 1000 children for each algorithm were calculated. 3,130,162 children (born 1996-2020) with 29,291,204 years of follow-up were included. ADHD prevalence per 1000 children in specialist settings ranged from 3.9 (Emilia Romagna) to 24.1 (Finland); and was 7.0 in primary care (Finland). Based on prescriptions, ADHD prevalence ranged from 0.1 (Emilia Romagna) to 9.9 (Haute Garonne). ASD prevalence in specialist settings ranged from 5.6 (Wales) to 9.7 (Finland), and in primary care from 1.0 (Finland) to 2.0 (Wales). Prevalence of ADHD and ASD was greater among children with longer follow-up. In Finland and Wales, 1.7% and 19.4% of children were diagnosed with ASD in primary care only respectively. The male:female ratio was 3-4:1. Whilst there was considerable geographical variation in the length of follow-up available, and prevalence of ADHD and ASD, specialist diagnoses recorded in healthcare data were key to identifying children with these disorders. These data sources can be complemented by using primary care diagnoses and prescription data to identify affected children more comprehensively.