Bridging the mental health treatment gap: Effects of a collaborative care intervention (matrix support) in the detection and treatment of mental disorders in a Brazilian city

OBJECTIVE: To analyse temporal trends in diagnosis and treatment of mental disorders in primary care following implementation of a collaborative care intervention (matrix support). DESIGN: Dynamic cohort design with retrospective time-series analysis. Structured secondary data on medical visits to general practitioners of all study clinics were extracted from the municipal electronic records database. Annual changes in the odds of mental disorders diagnoses and antidepressants prescriptions were estimated by multiple logistic regression at visit and patient-year levels with diagnoses or prescriptions as outcomes. Annual changes during two distinct stages of the intervention (stage 1 when it was restricted to mental health (2005-2009), and stage 2 when it was expanded to other areas (2010-2015)) were compared by adding year-period interaction terms to each model. SETTING: 49 primary care clinics in the city of Florianópolis, Brazil. PARTICIPANTS: All adults attending primary care clinics of the study setting between 2005 and 2015. RESULTS: 3 131 983 visits representing 322 100 patients were analysed. At visit level, the odds of mental disorder diagnosis increased by 13% per year during stage 1 (OR 1.13, 95% CI 1.11 to 1.14, p<0.001) and decreased by 5% thereafter (OR 0.95, 95% CI 0.94 to 0.95, p<0.001). The odds of incident mental disorder diagnoses decreased by 1% per year during stage 1 (OR 0.99, 95% CI 0.98 to 1.00, p=0.012) and decreased by 7% per year during stage 2 (OR 0.93, 95% CI 0.92 to 0.93, p<0.001). The odds of antidepressant prescriptions in patients with a mental disorder diagnosis increased by 7% per year during stage 1 (OR 1.07, 95% CI 1.05 to 1.20, p<0.001); this was driven by selective serotonin reuptake inhibitor prescriptions which increased 14% per year during stage 1 (OR 1.14, 95% CI 1.12 to 1.18, p<0.001) and 9% during stage 2 (OR 1.09, 95% CI 1.08 to 1.10, p<0.001). The odds of incident antidepressant prescriptions did not increase during stage 1 (OR 1.00, 95% CI 0.97 to 1.02, p=0.665) and increased by 3% during stage 2 (OR 1.03, 95% CI 1.00 to 1.04, p<0.001). Changes per year were all significantly greater during stage 1 than stage 2 (p values for interaction terms