Systematic comparison of household income, consumption, and assets to measure health inequalities in low- and middle-income countries.

There has been no systematic comparison of how the three most common measures to quantify household SES-income, consumption, and asset indices-could impact the magnitude of health inequalities. Microdata from 22 Living Standards Measurement Study surveys were compiled and concentration indices, relative indices of inequality, and slope indices of inequality were calculated for underweight, stunting, and child deaths using income, consumption, asset indices, and hybrid predicted income. Meta-analyses of survey year subgroups (pre-1995, 1995-2004, and post-2004), outcomes (child deaths, stunting, and underweight), and World Bank country-income status (low, low-middle, and upper-middle) were then conducted. Asset indices and the related hybrid income proxy result in the largest magnitudes of health inequalities for all 12 overall outcomes, as well as most country-income and survey year subgroupings. There is no clear trend of health inequality magnitudes changing over time, but magnitudes of health inequality may increase as country-income levels increase. There is no significant difference between relative and absolute inequality measures, but the hybrid predicted income measure behaves more similarly to asset indices than the household income it is supposed to model. Health inequality magnitudes may be affected by the choice of household SES measure and should be studied in further detail.