Effectiveness of community mental health nursing on psychiatric hospitalisations in Switzerland: a mirror-image analysis

BACKGROUND: Community mental health nursing has gained importance in Switzerland as part of a broader shift towards community-based psychiatric care. While community mental health nursing services are increasingly implemented, empirical evidence on their effectiveness in reducing psychiatric hospitalisations remains limited. This study aims to evaluate the effect of community mental health nursing on psychiatric hospitalisations and to identify risk factors for hospitalisation during community mental health nursing care. METHODS: We conducted a naturalistic observational study using data from two linked national databases: the HomeCareData register and the Medical Statistics of Hospitals. The study included adults who were admitted to community mental health nursing services between January 2020 and December 2022 and whose data was documented in HomeCareData. With mirror-image methods, we compared psychiatric hospitalisation rates in equally long pre-index and post-index periods, defined by each individual's duration of community mental health nursing care, with each participant serving as their own control. The start of community mental health nursing care was defined as the index date. For a sensitivity analysis, individuals hospitalised within 60 days before the start of community mental health nursing were excluded to address potential regression to the mean effects. Uni-and multivariable logistic regression were used to identify risk factors of hospitalisation during community mental health nursing care. RESULTS: The study sample comprised 5252 individuals with a mean age of 52 years; 66% were female. The primary mirror-image analysis revealed a significant reduction in psychiatric hospitalisations following the start of community mental health nursing (unadjusted Incidence Rate Ratio: 0.60; 95% confidence interval: 0.55-0.65). No significant effect was observed in the sensitivity analysis (unadjusted Incidence Rate Ratio: 0.97; 95% confidence interval: 0.87-1.09). Multivariable logistic regression identified previous psychiatric hospitalisations and compulsory admissions as significant risk factors of hospitalisation during community mental health nursing care. Sociodemographic and diagnostic variables showed no significant associations. CONCLUSIONS: This is the first study to show a reduction in psychiatric hospitalisations after participation in community mental health nursing services in Switzerland. However, this effect was not confirmed in the sensitivity analysis which may suggest a regression to the mean effect. The study demonstrates the feasibility of using linked routine data to evaluate real-world mental health interventions and underscores the need for broader data coverage and improved documentation to strengthen future research.