Making health continuous: implications of different methods on the measurement of inequality.

In most national surveys, health is measured as a categorical variable. However, in order to be able to calculate socio-economic inequalities in health, a continuous variable is needed. The recently developed interval regression approach was shown to outperform to other approaches like ordered probit. In this research we investigate the impact of using different sets of external data to estimate health inequalities: the EQ-index and the Canadian HUI. We found that the concentration index differs, but the income-related health mobility index and its decomposition are highly similar. Copyright (c) 2005 John Wiley & Sons, Ltd.