Multilevel analysis of individual heterogeneity and discriminatory accuracy (MAIHDA) within an intersectional...

Analyzing Body Mass Index as a didactical example, the study by Evans, Williams, Onnela, and Subramanian (EWOS study) introduce a novel methodology for the investigation of socioeconomic disparities in health. By using multilevel analysis to model health inequalities within and between strata defined by the intersection of multiple social and demographic dimensions, the authors provide a better understanding of the health heterogeneity existing in the population. Their innovative methodology allows for gathering inductive information on a large number of stratum-specific interactions of effects and, simultaneously, informs on the discriminatory accuracy of such strata for predicting individual health. Their study provides an excellent answer to the call for suitable quantitative methodologies within the intersectionality framework.

The EWOS study is a well-written tutorial; thus, in this commentary, I will not repeat the explanation of the statistical/epidemiological concepts. Instead, I will share with the reader a number of thoughts on the theoretical consequences derived from the application of multilevel analysis of individual heterogeneity and discriminatory accuracy (MAIHDA) in (social) epidemiology in general, and within the intersectional framework in particular. MAIHDA is a reorganization of concepts that allows for a better understanding of the distribution and determinants of individual health and disease risk in the population.

By applying MAIHD within an intersectional framework, the EWOS study provides a superior theoretical and quantitative instrument for documenting health disparities and it should become the new gold standard for investigating health disparities in (social) epidemiology. This approach is more appropriate for eco-social perspectives than the habitual probabilistic strategy based on differences between group average risks. However, both, the translation of intersectionality theory into (social) epidemiology and the intersectional quantitative methodology (especially for generalized linear models) are still under development.