Methods: Community key informants were trained to identify and refer people with probable SMD in Sodo district, south-central Ethiopia, using vignettes of typical presentations. Records of those referred to the new PHC-based service were linked to healthcare records to identify people who engaged with care and non-engagers over a six month period. Standardised interviews by psychiatric nurses were used to confirm the diagnosis in those attending PHC. Non-engagers were visited in their homes and administered the Psychosis Symptom Questionnaire. Socio-economic status, discrimination, disability, substance use, social support and distance to the nearest health facility were measured..
Results: Contact coverage for the new service was estimated to be 81.7% (300 engaged out of 367 probable cases of SMD identified). Reimbursement for transport and time may have elevated coverage estimates. In the fully adjusted multivariable model, rural residents had 3.81 increased odds (95% CI; 1.22, 11.89) of not accessing care, in part due to geographical distance from the health facility (Odds Ratio 3.37 (1.12, 10.12) for people living more than 180 minutes away. There was no association with lower socioeconomic status, female gender or physical impairment. Higher levels of functional impairment were associated with increased odds of engagement. Amongst non-engagers, the most frequently endorsed barriers were thinking the problem would get better by itself and concerns about the cost of treatment.
Conclusion and implications: Integrating mental healthcare into primary care can achieve high levels of coverage in a rural African setting, which is equitable with respect to gender and socio-economic status. Service outreach into the community may be needed to achieve better contact coverage for rural residents.