Social heterogeneity in self-reported health status and measurement of inequalities in health

This study aims to analyse the impact of the measurement of health status on socioeconomic inequalities in health. A MIMIC model with structural equations is used to create a latent variable of health status from four health indicators: self-assessed health, report of chronic diseases, report of activity limitations and mental health. Then, we disentangle the impact of sociodemographic characteristics on latent health from their direct impact on each heath indicator and discuss their effects on the assessment of socioeconomic inequalities in health. This study emphasises differences in inequalities in health according to latent health. In addition, it suggests the existence of reporting heterogeneity biases. For a given latent health status, women and old people are more likely to report chronic diseases. Mental health problems are over-reported by women and isolated people and under-reported by the oldest people. Active and retired people as well as non manual workers in the top of the social hierarchy more often report activity limitations. Finally, highly educated and socially advantaged people more often report chronic diseases whereas less educated people under-report a poor self-assessed health. To conclude, the four health indicators suffer from reporting heterogeneity biases and the report of chronic diseases is the indicator which biases the most the measurement of socioeconomic inequalities in health.