Geographic Variation in Household and Catastrophic Health Spending in India: Assessing the Relative Importance of Villages, Districts, and States, 2011-2012
- Per-capita household health spending was higher in economically developed states and was associated with ability to pay, but catastrophic health spending (CHS) was equally high in both poorer and more developed states in India.
- Based on multilevel modeling, we found that the largest geographic variation in health spending and CHS was at the state and village levels, reflecting wide inequality in the accessibility to and cost of health care at these levels.
- Contextual factors at macro and micro political units are important to reduce health spending and CHS in India.
Context: In India, health care is a local good, and households are the major source of financing it. Earlier studies have examined diverse determinants of health care spending, but no attempt has been made to understand the geographical variation in household and catastrophic health spending. We used multilevel modeling to assess the relative importance of villages, districts, and states to health spending in India.
Methods: We used data on the health expenditures of 101,576 households collected in the consumption expenditure schedule (68th round) carried out by the National Sample Survey in 2011-2012. We examined 4 dependent variables: per-capita health spending (PHS), per-capita institutional health spending (PIHS), per-capita noninstitutional health spending (PNHS), and catastrophic health spending (CHS). CHS was defined as household health spending exceeding 40% of its capacity to pay. We used multilevel linear regression and logistic models to decompose the variation in each outcome by state, region, district, village, and household levels.
Findings: The average PHS was 1,331 Indian rupees (INR), which varied by state-level economic development. About one-fourth of Indian households incurred CHS, which was equally high in both the economically developed and poorer states. After controlling for household level factors, 77.1% of the total variation in PHS was attributable to households, 10.1% to states, 9.5% to villages, 2.6% to districts, and 0.7% to regions. The pattern in variance partitioning was similar for PNHS. The largest interstate variation was found for CHS (15.9%), while the opposite was true for PIHS (3.2%).
Conclusions: We observed substantial variations in household health spending at the state and village levels compared with India’s districts and regions. The large variation in CHS attributable to states indicates interstate inequality in the accessibility to and cost of health care. Our findings suggest that contextual factors at the macro and micro political units are important to reduce India’s household health spending and CHS.
Keywords: household health spending, catastrophic health spending, geographic variation, multilevel modeling, India.
Volume 96, Issue 1 (pages 167-206)
Published in 2018