An agent-based model to investigate the effects of urban segregation around the clock on inequalities in health behaviour.

Abstract

Social segregation in cities refers to the uneven spatial distribution of individuals from unequal social groups, such as affluent and economically vulnerable people. Social segregation may, in turn, produce social inequalities through contextual effects, since neighbourhood mixing or concentration plays a role in shaping individuals' opinions and behaviours in multiple life domains, including health. Because segregation and contextual effects occur at the places of residence as well as throughout the day, as people move between locations in a city, we aim to understand the social effect of urban segregation 'around the clock' on health behaviours (such as the choice of a healthy diet), using an empirical agent-based model initialised on the Paris region with a synthetic population. We built this synthetic population by pulling together data from two health & nutrition surveys conducted 6 years apart, data from the French census and data from an origin-destination survey. We then combined scenarios of residential patterns (random allocation vs. census-based allocation reflecting the empirical level of residential segregation) with scenarios of daily mobility (no daily moves, random moves or survey-based daily moves reflecting the empirical level of daytime segregation in Paris) to assess the effect of spatio-temporal segregation on the diffusion of health behaviours. While the same upward trend of healthy behaviours is obtained in all scenarios simulated, we find contrasted results with respect to social inequalities: 1/ when the agents' residence is allocated at random, social inequalities of health decrease in the long run; 2/ randomizing daily mobility can mitigate the increase in social inequalities in dietary behaviours induced by effective residential segregation, with this mitigation effect appearing as soon as a small proportion of daily moves are random; 3/ daytime segregation as it exists in Paris slightly reinforces the unequal distribution of health behaviours between the most and least educated groups compared with the sole effect of residential segregation.

Supplementary information: The online version contains supplementary material available at 10.1140/epjds/s13688-025-00603-4.