Rural‒urban disparities in household catastrophic health expenditure in Bangladesh: a multivariate decomposition analysis

Rural‒urban disparities in household catastrophic health expenditure in Bangladesh: a multivariate decomposition analysis

Background: Rural‒urban disparity in catastrophic healthcare expenditure (CHE) is a well-documented challenge in low- and middle-income countries, including Bangladesh, limiting financial protection and hindering the achievement of the Universal Health Coverage target of the United Nations Sustainable Development Goals. However, the factors driving this divide remain poorly understood. Therefore, this study aims to identify the key determinants of the rural‒urban disparity in CHE incidence in Bangladesh and their changes over time.
Methods: We used nationally representative data from the latest three rounds of the Bangladesh Household Income and Expenditure Survey (2005, 2010, and 2016). CHE incidence among households seeking healthcare was measured using the normative food, housing, and utilities method. To quantify covariate contributions to the rural‒urban CHE gap, we employed the Oaxaca-Blinder multivariate decomposition approach, adapted by Powers et al. for nonlinear response models.
Results: CHE incidence among rural households increased persistently during the study period (2005: 24.85%, 2010: 25.74%, 2016: 27.91%) along with a significant (p-value ≤ 0.01) rural‒urban gap (2005: 9.74%-points, 2010: 13.94%-points, 2016: 12.90%-points). Despite declining over time, substantial proportions of CHE disparities (2005: 87.93%, 2010: 60.44%, 2016: 61.33%) are significantly (p-value ≤ 0.01) attributable to endowment differences between rural and urban households. The leading (three) covariate categories consistently contributing significantly (p-value ≤ 0.01) to the CHE gaps were composition disparities in the lowest consumption quintile (2005: 49.82%, 2010: 36.16%, 2016: 33.61%), highest consumption quintile (2005: 32.35%, 2010: 15.32%, 2016: 18.39%), and exclusive reliance on informal healthcare sources (2005: -36.46%, 2010: -10.17%, 2016: -12.58%). Distinctively, the presence of chronic illnesses in households emerged as a significant factor in 2016 (9.14%, p-value ≤ 0.01), superseding the contributions of composition differences in household heads with no education (4.40%, p-value ≤ 0.01) and secondary or higher education (7.44%, p-value ≤ 0.01), which were the fourth and fifth significant contributors in 2005 and 2010.
Conclusions: Rural‒urban differences in household economic status, educational attainment of household heads, and healthcare sources were the key contributors to the rural‒urban CHE disparity between 2005 and 2016 in Bangladesh, with chronic illness emerging as a significant factor in the latest period. Closing the rural‒urban CHE gap necessitates strategies that carefully address rural‒urban variations in the characteristics identified above.