Income in epidemiological research: a guide to measurement and analytical treatment with a case study on mental disorders and mortality.

Abstract

Background: Income is one of the most widely used indicators of socioeconomic position in studies of health inequalities. Despite its frequent use, no empirical-driven guidelines exist on how to operationalise and analytically handle income. In this study, we develop straightforward step-by-step graphical guidelines based on a comprehensive comparative examination.

Methods: We followed the entire population aged 30-65 residing in Denmark in 2010 (n=2 699 296) until death, diagnosis of either severe mental disorder or neurotic, stress-related somatoform or depressive disorders, respectively, or censoring. Adjusted HRs were estimated for four tax-reported income measures, examining variations due to the handling of extreme values, income categorisation and reference categories, both overall and stratified by sex.

Results: Across all income measures and handling strategies, lower baseline incomes were consistently associated with an elevated risk of experiencing adverse health outcomes with reversed patterns at the lowest end of the income spectrum. At these low-income levels, lower income was associated with a steep increase in corresponding household assets, with very high asset values observed at both ends of the income spectrum. Associations varied mainly for the lowest income categories with some differences in these patterns between males and females depending on the handling of extreme values, income categorisation and choice of reference category.

Conclusion: The findings indicate a misclassification of financial capability for the lowest tax-reported incomes and emphasise the need for methodological clarity in income-based health research. The step-by-step graphical guidelines, informed by the findings, offer a methodological framework aiming to enhance the precision, comparability and relevance of future studies.