Johannesburg's ‘poor housing, good health’ paradox: the role of health status assessment, statistical modelling, residential context and migrant status

Abstract

ObjectivesThe ‘poor housing, good health’ paradox observed by De Wet et al. (2011) across eight of Johannesburg's poorest Wards (neighbourhoods), was re-examined using: a more generic measure of self-reported health; better-specified adjustment for measured confounders; household data from a range of Wards and income strata across Johannesburg and Gauteng; and adjustment for migrant status. Study designThe present study examined the relationship between (formal vs. informal) housing and self-reported health (-limited work and/or social activities) across four subsamples of respondents to the 2013 Quality of Life survey undertaken by the Gauteng City Regional Observatory: n = 1494 from households in the eight Wards examined by De Wet et al. (2011); n = 3059 from households with the lowest income tertile in Johannesburg; n = 8263 from households throughout Johannesburg; and n = 24,727 from households throughout Gauteng Province, irrespective of Ward or income. MethodsThe relationship between housing and self-reported health in each of these subsamples was examined before and after adjustment for measured confounders identified using a temporally determined causal path diagram in the form of a directed acyclic graph. ResultsFollowing adjustment for measured confounders, ‘informal housing’ was only associated with ‘good’ self-reported health in Johannesburg's poorest Wards (odds ratio [OR]: 1.39; 95% confidence interval [CI]: 1.07, 1.79) and Johannesburg as a whole (OR: 1.15; 95%CI: 1.00, 1.31). These associations were substantially attenuated following adjustment for migrant status (OR: 1.26; 95%CI: 0.97, 1.65; OR: 1.07; 95%CI: 0.93, 1.22, respectively). ConclusionsWhile the present study found that Johannesburg's ‘poor housing, good health’ paradox was still evident when using a more generic/subjective (self-reported) measure of health, the apparent paradox appears to reflect, at least in part: the differential aggregation of migrant-headed households in Johannesburg neighbourhoods exhibiting a high concentration of informal housing; and the likely impact of the health-related selection on the health of migrant-headed households.