A mixed methods evaluation of an integrated adult mental health service model

Background The Floresco integrated service model was designed to address the fragmentation of community mental health treatment and support services. Floresco was established in Queensland, Australia, by a consortium of non-government organisations that sought to partner with general practitioners (GPs), private mental health providers and public mental health services to operate a 'one-stop' mental health service hub. Methods We conducted an independent mixed-methods evaluation of client outcomes following engagement with Floresco (outcome evaluation) and factors influencing service integration (process evaluation). The main data sources were: (1) routinely-collected Recovery Assessment Scale - Domains and Stages (RAS-DS) scores at intake and review (n = 108); (2) RAS-DS scores, mental health inpatient admissions and emergency department (ED) presentations among clients prospectively assessed at intake and six-month follow-up (n = 37); (3) semi-structured interviews with staff from Floresco, consortium partners, private practitioners and the local public mental health service (n = 20); and (4) program documentation. Results Interviews identified staff commitment, co-location of services, flexibility in problem-solving, and anecdotal evidence of positive client outcomes as important enablers of service integration. Barriers to integration included different organisational practices, difficulties in information-sharing and in attracting and retaining GPs and private practitioners, and systemic constraints on integration with public mental health services. Of 1129 client records, 108 (9.6%) included two RAS-DS measurements, averaging 5 months apart. RAS-DS 'total recovery' scores improved significantly (M = 63.3%, SD = 15.6 vs. M = 69.2%, SD = 16.1; p < 0.001), as did scores on three of the four RAS-DS domains ('Looking forward', p < 0.001; 'Mastering my illness', p < 0.001; and 'Connecting and belonging', p = 0.001). Corresponding improvements, except in 'Connecting and belonging', were seen in the 37 follow-up study participants. Decreases in inpatient admissions (20.9% vs. 7.0%), median length of inpatient stay (8 vs. 3 days), ED presentations (34.8% vs. 6.3%) and median duration of ED visits (187 vs. 147 min) were not statistically significant. Conclusions Despite the lack of a control group and small follow-up sample size, Floresco's integrated service model showed potential to improve client outcomes and reduce burden on the public mental health system. Horizontal integration of non-government and private services was achieved, and meaningful progress made towards integration with public mental health services.