Marginal effects analysis of correlates of enrollment in a community–based health insurance scheme in Ethiopia

Community–based health insurance (CBHI) is advocated as an alternative financing scheme to cater for the unexpected nature of healthcare expenditure for the vulnerable segments of the population in low-income countries. The general objective of this study was to assess households’ willingness to enroll in  a proposed CBHI scheme in Adama city, Ethiopia. Two-stage cluster sampling was used to collect data from a random sample of members of mutual aid associations (called Iddirs) using structured questionnaires. The impact of covariates on respondents’ willingness to enroll in the proposed scheme was explored using average marginal effects and marginal effects at representative values from the fitted binary logistic regression model. The results revealed that the probability of willingness to join the proposed scheme was higher for male respondents, respondents with at least primary education, married respondents, the elderly and respondents who have serious difficulty in finding money in times of ill-health. Marginal effects analysis also revealed that the impact of gender and marital status diminishes as age and level of education increase. Joint marginal effects analysis of covariates indicated that the probability of willingness to enroll in the scheme was significantly lower for unmarried and divorced/widowed female respondents with no education. Thus, there should be selective targeting of the vulnerable groups within the communities in setting up CBHI schemes.