Recurrent Group-switch Interactions in Heterogeneous Population Epidemic Modelling

Unexpected epidemic outcomes that show discrepancies between predicted and actual epidemics and the impacts of interventions are often attributed to factors such as biased data collection or pathogen evolution, while less attention is given to the realism of the epidemic models that inform public health decisions. Many models guiding policy still rely on simplified assumptions of homogeneous mixing, usually combining the effects of mobility and differing contact strengths within/between different sociodemographic groups at home and at work/school into a single transmission parameter, thereby overlooking the heterogeneity and structure of real human contact networks. Even in age-structured models, people within the same age group are considered to mix equally; moreover, differences in inter-group interaction that exist at home and work/school are often ignored. Recent studies have improved the realism of epidemic modelling by incorporating the concept of semi-random mixing and explicitly representing household and non-household interactions linked by daily mobility. Building on these developments, this study introduces a recurrent group-switch epidemic model that captures how individuals transition between socio-demographic groups at home and work or school, incorporating sociodemographic structure and semi-random mixing within a computationally efficient equation-based framework. Analytical derivations yield new formulations for the force of infection and group-specific metrics, including source, sink, and source-to-sink reproduction numbers. Model simulations using UK age-structured contact data underscore how, for COVID-19-like infections, individuals aged 6–24 years act as key drivers of transmission, while older adults serve primarily as infection recipients rather than infectors (sinks) with higher hospitalisation risks. Modelling non-pharmaceutical interventions shows that reducing inter-household/cluster connectivity among younger populations may substantially reduce transmission, whereas mobility restrictions alone can produce counterintuitive increases in epidemic size. By explicitly linking recurrent social behaviour, heterogeneity, and mobility, this model framework improves the realism of epidemic models and provides deeper insight into group-specific transmission dynamics, which could be used to guide more targeted and effective public health interventions.