Modelling healthy ageing in South Africa: a network analysis of intrinsic capacities

The rapidly increasing older adult population in sub-Saharan Africa necessitates an understanding of how health domains interact to shape healthy ageing. Existing approaches overlook complex interdependencies between cognitive, physical, emotional, and functional capacities. This study applies network analysis to model these relationships in a large cohort of older adults in rural South Africa.

We analysed data from 4783 participants (aged ≥40) in the 2021-2022 wave of the HAALSI study. Mixed Graphical Models (MGMs) were estimated from a mixed correlation matrix to assess conditional associations among 12 variables spanning cognitive function, depressive symptoms, physical capacity, self-rated health, and socioeconomic status. Central and bridge nodes were identified, and gender-stratified networks were compared.

The network revealed strong interconnectivity among intrinsic capacity domains. Cognitive function (delayed recall) and functional ability (ADL limitations) emerged as the most central nodes, with self-rated health and depressive symptoms as key bridges. While global network strength and structure did not differ significantly by gender, men's networks were anchored around cognitive function, whereas women's were more centred on physical functioning. Depressive symptom clustering was denser among women.

Intrinsic capacity domains are tightly coupled in rural South African older adults; cognition and ADL limitations are pivotal. Subtle gender differences in network structure suggest a need for tailored, multidomain interventions. These findings support the WHO Healthy Ageing framework and highlight the utility of network models for identifying intervention leverage points in low-resource settings.