Health service costs and their association with functional impairment among adults receiving integrated mental health care in five low- and middle-income countries: the PRIME cohort study

Abstract

Abstract This study examines the level and distribution of service costs—and their association with functional impairment at baseline and over time—for persons with mental disorder receiving integrated primary mental health care. The study was conducted over a 12-month follow-up period in five low- and middle-income countries participating in the Programme for Improving Mental health carE study (Ethiopia, India, Nepal, South Africa and Uganda). Data were drawn from a multi-country intervention cohort study, made up of adults identified by primary care providers as having alcohol use disorders, depression, psychosis and, in the three low-income countries, epilepsy. Health service, travel and time costs, including any out-of-pocket (OOP) expenditures by households, were calculated (in US dollars for the year 2015) and assessed at baseline as well as prospectively using linear regression for their association with functional impairment. Cohort samples were characterized by low levels of educational attainment (Ethiopia and Uganda) and/or high levels of unemployment (Nepal, South Africa and Uganda). Total health service costs per case for the 3 months preceding baseline assessment averaged more than US$20 in South Africa, $10 in Nepal and US$3–7 in Ethiopia, India and Uganda; OOP expenditures ranged from $2 per case in India to $16 in Ethiopia. Higher service costs and OOP expenditure were found to be associated with greater functional impairment in all five sites, but differences only reached statistical significance in Ethiopia and India for service costs and India and Uganda for OOP expenditure. At the 12-month assessment, following initiation of treatment, service costs and OOP expenditure were found to be lower in Ethiopia, South Africa and Uganda, but higher in India and Nepal. There was a pattern of greater reduction in service costs and OOP spending for those whose functional status had improved in all five sites, but this was only statistically significant in Nepal.