ABSTRACT
AIM: To estimate the treated incidence of individuals with first-episode psychosis (FEP) who contacted the Emilia-Romagna public mental healthcare system (Italy); to examine the variability of incidence and user characteristics across centres and years. METHODS: We computed the raw treated incidence in 2013-2019, based on FEP users aged 18-35, seen within or outside the regional program for FEP. We modelled FEP incidence across 10 catchment areas and 7 years using Bayesian Poisson and Negative Binomial Generalized Linear Models of varying complexity. We explored associations between user characteristics, study centre and year comparing variables and socioclinical clusters of subjects. RESULTS: Thousand three hundred and eighteen individuals were treated for FEP (raw incidence: 25.3 / 100.000 inhabitant year, IQR: 15.3). A Negative Binomial location-scale model with area, population density and year as predictors found that incidence and its variability changed across centres (Bologna: 36.55; 95% CrI: 30.39-43.86; Imola: 3.07; 95% CrI: 1.61-4.99) but did not follow linear temporal trends or density. Centers were associated with different user age, gender, migrant status, occupation, living conditions and cluster distribution. Year was associated negatively with HoNOS score (R = -0.09, p < .001), duration of untreated psychosis (R = -0.12, p < .001) and referral type. CONCLUSIONS: The Emilia-Romagna region presents a relatively high but variable incidence of FEP across areas, but not in time. More granular information on social, ethnic and cultural factors may increase the level of explanation and prediction of FEP incidence and characteristics, shedding light on social and healthcare factors influencing FEP.
