Exploring social determinants of disability among older filipinos: insights from a polysocial score approach

The rapid aging in the Philippines presents significant challenges, including high rates of activities of daily living (ADL) disability among older adults. While research has identified various social determinants of ADL disability, there is a gap in understanding how their joint impact on disability among older Filipinos. We adopted the polysocial score approach to assess these joint associations of multiple social factors with disability among older adults in the Philippines.

Individuals included in the analysis were from the Longitudinal Study of Ageing and Health in the Philippines. Twenty-nine social factors from five domains were included. The Light Gradient Boosting Machine (LightGBM) model was employed to identify and rank key social determinants, which were then incorporated into a logistic regression model to derive the polysocial score, both continuously and in categories and assessed its association with ADL disability. Model performances were assessed by discrimination, calibration, and reclassification (compared to a reference model). All analyses were separated for men and women.

We included 5,000 participants (37.0% men) with an average age of 68.3 years. The polysocial score included 19 social factors for men and 18 for women. Among men, the most significant predictors were the frequency of engaging in social activities and the number of friends they had contact with. Attending religious services outside the home and the frequency of attending social activities were the most important factors for women. Eleven social factors overlapped between men and women. Individuals with moderate or high polysocial scores exhibited reduced likelihood of experiencing ADL disability relative to those with low scores. We observed satisfactory model performance among men and women.

We identified important social factors for men and women and their joint association with ADL disability. The polysocial score could be used to design person-centered social interventions that promote health and support independence.