PLoS ONE (Jan 2020)

Outcomes of a primary care mental health implementation program in rural Rwanda: A quasi-experimental implementation-effectiveness study.

  • Stephanie L Smith,
  • Molly F Franke,
  • Christian Rusangwa,
  • Hildegarde Mukasakindi,
  • Beatha Nyirandagijimana,
  • Robert Bienvenu,
  • Eugenie Uwimana,
  • Clemence Uwamaliya,
  • Jean Sauveur Ndikubwimana,
  • Sifa Dorcas,
  • Tharcisse Mpunga,
  • C Nancy Misago,
  • Jean Damascene Iyamuremye,
  • Jeanne d'Arc Dusabeyezu,
  • Achour A Mohand,
  • Sidney Atwood,
  • Robyn A Osrow,
  • Rajen Aldis,
  • Shinichi Daimyo,
  • Alexandra Rose,
  • Sarah Coleman,
  • Anatole Manzi,
  • Yvonne Kayiteshonga,
  • Giuseppe J Raviola

DOI
https://doi.org/10.1371/journal.pone.0228854
Journal volume & issue
Vol. 15, no. 2
p. e0228854

Abstract

Read online

INTRODUCTION:To address the know-do gap in the integration of mental health care into primary care in resource-limited settings, a multi-faceted implementation program initially designed to integrate HIV/AIDS care into primary care was adapted for severe mental disorders and epilepsy in Burera District, Rwanda. The Mentoring and Enhanced Supervision at Health Centers (MESH MH) program supported primary care-delivered mental health service delivery scale-up from 6 to 19 government-run health centers over two years. This quasi-experimental study assessed implementation reach, fidelity, and clinical outcomes at health centers supported by MESH MH during the scale up period. METHODS:MESH MH consisted of four strategies to ensure the delivery of the priority care packages at health centers: training; supervision and mentorship; audit and feedback; and systems-based quality improvement (QI). Implementation reach (service use) across the 19 health centers supported by MESH MH during the two year scale-up period was described using routine service data. Implementation fidelity was measured at four select health centers by comparing total clinical supervisory visits and checklists to target goals, and by tracking clinical observation checklist item completion rates over a nine month period. A prospective before and after evaluation measured clinical outcomes in consecutive adults presenting to four select health centers over a nine month period. Primary outcome assessments at baseline, 2 and 6 months included symptoms and functioning, measured by the General Health Questionnaire (GHQ-12) and the World Health Organization Disability Assessment Scale (WHO-DAS Brief), respectively. Secondary outcome assessments included engagement in income generating work and caregiver burden using a quantitative scale adapted to context. RESULTS:A total of 2239 mental health service users completed 15,744 visits during the scale up period. MESH MH facilitated 70% and 76% of supervisory visit and clinical checklist utilization target goals, respectively. Checklist item completion rates significantly improved overall, and for three of five checklist item subgroups examined. 121 of 146 consecutive service users completed outcome measurements six months after entry into care. Scores improved significantly over six months on both the GHQ-12, with median score improving from 26 to 10 (mean within-person change 12.5 [95% CI: 10.9-14.0] p< 0.0001), and the WHO-DAS Brief, with median score improving from 26.5 to 7 (mean within-person change 16.9 [95% CI: 14.9-18.8] p< 0.0001). Over the same period, the percentage of surveyed service users reporting an inability to work decreased significantly (51% to 6% (p < 0.001)), and the proportion of households reporting that a caregiver had left income-generating work decreased significantly (41% to 4% (p < 0.001)). CONCLUSION:MESH MH was associated with high service use, improvements in mental health care delivery by primary care nurses, and significant improvements in clinical symptoms and functional disability of service users receiving care at health centers supported by the program. Multifaceted implementation programs such as MESH MH can reduce the evidence to practice gap for mental health care delivery by nonspecialists in resource-limited settings. The primary limitation of this study is the lack of a control condition, consistent with the implementation science approach of the study. STUDY REGISTRATION:ISRCTN #37231.