Clinical outcomes of a primary care mental health implementation program scale-up in the Eastern Province of Rwanda: a prospective cohort study.

BACKGROUND: The Mentoring and Enhanced Supervision at Health Centers for Mental Health (MESH-MH) program supports mental health care delivery by non-specialist, primary care providers in Rwanda to improve access to quality mental health care. After an initial pilot in one rural district, the program was scaled to support service delivery across two additional districts. We aimed to assess changes in symptoms and functioning outcomes among patients who received treatment at selected health centers supported by MESH-MH. We compare them with patient outcomes from the original pilot district. METHODS: We conducted a prospective cohort study among patients with mental health conditions and epilepsy who started treatment at ten health centers in the Rwinkwavu and Kirehe Hospital catchment areas between November 2020 and December 2021. We assessed patients' symptoms using the 12-item General Health Questionnaire (GHQ-12) and daily functioning using the 12-item World Health Organization Disability Assessment Scale (WHO-DAS 2.0 Brief). We collected data at baseline, midline (on average, 71 days from treatment initiation), and endline (on average, 197 days from treatment initiation). We measured changes over time in outcomes and associations with the duration of treatment using Wilcoxon signed rank tests and linear mixed models. We compared findings with those of our previous study in the Burera district using linear mixed models and adjusting for possible confounding factors. RESULTS: Of 151 participants enrolled, primary diagnoses were epilepsy (n = 45, 29.8%), depression (n = 36, 23.8%), brief psychosis (n = 26, 17.2%), and schizophrenia (n = 20, 13.2%). The median GHQ-12 score improved from 24 (IQR: 18-30) at baseline to 11 (IQR: 5-17) at endline (median change: -12 [IQR: -19, -6], p < 0.001). The median WHO-DAS Brief score improved from 21.8 (IQR: 13-30) to 6 (IQR: 3-15) (median change: -13 [IQR: -22, -4], p < 0.001). The median number of days with having difficulties to carry out regular activities declined from 20 (IQR: 7-20) at baseline to 5 (IQR: 2-15) at endline (median change: -10 [IQR: -21, 0], p < 0.001) with comparable significant reductions in the number of days where patients were less able to engage in usual activities. We consistently observed significant improvements when using linear mixed models and across patient diagnoses. The scale-up sites revealed slightly greater reductions in symptoms relative to Burera, the pilot site study, while they showed comparable improvements in functioning. CONCLUSIONS: Our findings indicate that patients with mental health conditions and epilepsy who received treatment at health centers newly supported by MESH-MH experienced significant improvements in symptoms and functioning. These improvements were comparable to the outcomes seen in the initial pilot of MESH-MH. Scaling-up interventions like the MESH-MH model could help to decentralize and increase access to mental health services in Rwanda and other similar settings.