Assessing health gradient with different equivalence scales for household income – A sensitivity analysis

Abstract

:Research on health inequalities shows considerable variation in health by socioeconomic position regardless of measurement. Income level is among the most commonly used indicators to measure the social gradient in health. The income gradient in health may, however, vary according to what aspect of health is studied but equally it may depend on how income is measured. The traditional approach of measuring income is to use household income and to modify it with consumption needs and size of the household. Our hypothesis was that the traditional picture of the income-health gradient becomes more nuanced when we use different equalization scales for household incomes than the most often used modified OECD equalization scale. More technically, we expected the steepness of the income-health gradient to change when the equivalence scale for adjusting household size and composition is altered. The data were Finnish cross-sectional 2017 data from EU-SILC (N = 9406). The primary measures were perceived health status and total household disposable annual income by household type. Ordered probit estimation using Stata package was applied to solving the function between health and income controlling for age, age-squared and gender. Respondents’ health status associated with their household type similarly to the association between personal income and household type. Those living in a single household tend to report poorer health but also tend to have lower personal income. Our main finding was that the health-income gradient becomes steeper with the larger equivalence scales, i.e., larger scale relativities, which assume bigger consumption needs for additional household members. One should be more aware of the fact that when household consumption is adjusted with conventional equivalence scales, other income-related aspects beyond consumption potential – such as social status, economic security – are also adjusted for.