Accounting for geographical inequalities in the assessment of equity in health care: a benefit incidence analysis

Abstract

BackgroundEquity in health expenditure in low-income and middle-income countries is commonly analysed using benefit incidence analysis (BIA). In BIA, the monetary value of the subsidy associated with public sector health-care utilisation (approximated by the cost of the service) is attributed to each individual according to their frequency and type of health-care utilisation. The benefit distribution is measured according to socioeconomic status. Despite widespread within-country geographical inequalities in health status and public expenditure, BIA has rarely accounted for such differences. We investigate how results would differ if geographical inequalities were taken into account. MethodsWe carry out four versions of BIA for outpatient public health-care expenditure in Manica Province, Mozambique and compare the results. First, following standard practice, we rank individuals by socioeconomic status (measured by their household consumption) and we use average expenditure across districts to estimate the individual benefit. Second, we use a disaggregated measure of expenditure across districts to calculate the benefit. Third, we rank individuals by a measure of need for health care based on socioeconomic status and average district health status (measured by child mortality). Fourth, we combine the second and third approaches. We use data from the Household Budget Survey 2008/09, Census 2007, Ministry of Health, and Ministry of Finance records. FindingsWe find that the gap in benefit from public expenditure between highest and lowest quintiles widens substantially if differences in health status and expenditure across districts are taken into account, increasing from a ratio of 1·2 to 2·0. InterpretationResults suggest that the methods currently used may underestimate inequities in public health expenditure in contexts where geographical inequalities exist. Refinement of BIA using disaggregated data available from local institutions may improve estimates, stimulate local information systems' strengthening, and ultimately provide insights for a more equitable and efficient allocation of resources. FundingML and KH are part of RESYST, a health systems research programme consortium funded by UKaid from the Department for International Development. The views expressed do not necessarily reflect the department's official policies.