Social capital and self-rated health of residents of Gauteng province: Does area-level deprivation influence the relationship?

Although social capital has been linked to population health, there is a dearth of studies on the phenomenon especially in sub-Saharan Africa. We investigated the individual and contextual effects of social capital indicators (group membership, registered to vote, perception towards safety in community and generalised trust) on the self-rated health (SRH) of the residents of Gauteng province. We used data from the 2015 Quality of Life (QoL) survey, which included a random representative sample of 27476 residents (level 1) in 508 administrative wards (level 2). We employed a multilevel logistic regression to examine the association of social capital and SRH (good vs poor). After adjusting for individual and area-level factors, no main effect of group membership (Adjusted OR: 0.93: 95% CI: 0.85-1.02), generalised trust (Adjusted OR: 1.01: 95% CI: 0.89-1.49) and registered to vote (Adjusted OR: 0.95; 95% CI: 0.82-1.10) was observed. However, if respondents were positive in their perception towards safety in community, there was a positive association with good SRH (Adjusted OR: 1.15; 95% CI: 1.01-1.31); while if residents reported a negative perception towards safety in community, a strong negative association with good SRH (Adjusted OR: 0.70; 95% CI: 0.62-0.79) was observed. Both ward variance and median odds ratio (MOR) indicate significant differences in good SRH by wards. A strong positive joint effect on the multiplicative scale was observed between satisfied with safety and the ward-level South African Multiple Deprivation Poverty Index (SAMPI), while a strong negative joint effect was also observed on a multiplicative scale between dissatisfied with safety and the SAMPI. Perception of safety in community is the core domain of social capital that significantly impacts the SRH of residents of Gauteng. Although the effect of perception towards safety in community on good SRH is influenced by ward deprivation,the effect is not dependent on the level of deprivation. Contextual factors as evidenced by the persistent MOR, in addition to individual factors, explain variation in reporting good SRH in the study area.