Revisiting socio-economic inequalities in sedentary leisure time in Sweden: An intersectional analysis of individual heterogeneity and discriminatory accuracy (AIHDA)

Abstract

Aims: Swedish public health reports have repeatedly provided information about socio-economic inequalities in sedentary leisure time, despite that, in the interest of health equity, physical activity should be equally distributed in the population. Such public health reports, however, neither consider the intersection of multiple socio-demographic factors nor the individual heterogeneity around group averages. Drawing on intersectionality theory, this study aimed to revisit previous findings on sedentary leisure time from Swedish public health surveys and demonstrate how the analysis of individual heterogeneity and discriminatory accuracy (AIHDA) can be used for analysing complex health inequalities.

Methods: Using data from Swedish national public health surveys (2004-2015), we applied the AIHDA to define 72 intersectional groups by categories of age, gender, educational achievement, migration status and household composition. We then calculated (a) the absolute and relative risk of sedentary leisure time and (b) the discriminatory accuracy (DA) of the intersectional grouping.

Results: The average risk of sedentary leisure time ranged from 5.8% among native-born, highly educated, young women living alone to 41.0% among immigrated young men, living alone, with low education. The risk was higher in strata comprising immigrated people with low education and lower in strata including native-born, highly educated people. However, the DA of the grouping was poor, indicating a substantial overlap of individual risk between groups.

Conclusions: Using the AIHDA and drawing on intersectionality, this study provides an improved mapping of the socio-economic distribution of sedentary leisure time in Sweden, with the poor DA suggesting universal rather than targeted physical activity interventions.

Keywords: Health inequalities; Sweden; intersectionality; physical activity; public health; sedentarism.

Figure 1.

Flow chart describing the National Public Health Surveys study sample, as well as the inclusion and exclusion criteria and missing data.

Figure 2.

Prevalence of sedentary leisure time with 95% confidence intervals (vertical lines) for the 72 intersectional strata defined by combining the three categories of age, two categories of gender, three categories of educational achievement, two categories of migration status and two categories of household composition (A: living alone; C: cohabiting). The lines connecting the prevalence of sedentarism across groups with different levels of educational achievement show clear socio-economic gradients. The figures were obtained after weighting and imputation for missing values on educational achievement.