Wealth, education, and sex split chronic disease risk in rural South Africa
In rural South Africa’s DIMAMO surveillance site, socioeconomic status did not line up neatly with a single pattern of chronic disease risk. The clearest divide emerged in harmful alcohol use, which was more common in the richer and richest quintiles, while physical inactivity fell as wealth increased. Tobacco use tracked in a different direction, clustering less often among women and among adults with secondary or tertiary education. These were not abstract social gradients; they were measurable differences in the distribution of everyday exposures that feed into non-communicable disease burden over time.
The survey covered 398 adults aged 18 years and older attending primary healthcare facilities in the DIMAMO Health and Demographic Surveillance System catchment area. Socioeconomic status was built from an asset-based wealth index using principal component analysis and then split into quintiles. Investigators used adapted WHO STEPS and SAGE questionnaires to collect information on tobacco use, harmful alcohol use, physical inactivity, low fruit and vegetable intake, high salt intake, self-reported hypertension, and diabetes. The analytic approach relied on modified Poisson regression with robust standard errors, a practical choice when prevalence ratios are the goal rather than odds ratios that can exaggerate common outcomes.
A single blood pressure or glucose result would not have captured the same pattern. The study was designed instead to show how social position maps onto risk behavior and self-reported chronic disease in a setting where the epidemiology of NCDs is already shifting under pressure from demographic change, nutrition transition, and uneven access to prevention.
Harmful alcohol use rose with wealth while inactivity fell
The strongest socioeconomic gradient involved alcohol. Adults in the richer quintile had a higher prevalence of harmful alcohol use than those in the poorest group, with an adjusted prevalence ratio of 1.58 and a 95% confidence interval of 1.03 to 2.41. The richest quintile showed an even higher prevalence, with an adjusted prevalence ratio of 1.78 and a 95% confidence interval of 1.11 to 2.85. These estimates point to a meaningful association rather than a weak signal drifting across the confidence interval.
Physical inactivity moved in the opposite direction. It was less prevalent among participants in the middle, richer, and richest quintiles than among those in the poorest group. In a rural setting, that pattern may reflect differences in the types of work, mobility, household resources, or leisure-time structure that accompany higher wealth. It also shows why a single assumption about “better” socioeconomic position can mislead prevention planning. Higher wealth can protect against some risks while concentrating others, especially when social life, alcohol availability, and consumption patterns change faster than health behavior.
Tobacco use tracked with sex and schooling rather than wealth alone
Tobacco use was less common among females and among participants with secondary or tertiary education. This finding fits a broader social pattern in which smoking and other forms of tobacco exposure often cluster in groups with fewer educational opportunities and less access to health information or cessation support. Education likely acts as more than a marker of knowledge; it shapes employment, social networks, and exposure to prevention messages, all of which can influence whether smoking becomes established and sustained.
No single wealth gradient captured tobacco behavior in the same way it captured alcohol use. That separation matters for local prevention strategy. A clinic or community program that assumes one socioeconomic lever will shift all behavioral risks is likely to miss the distinct social drivers behind each exposure.
Hypertension and diabetes did not follow the same wealth pattern
Socioeconomic status was not significantly associated with hypertension after adjustment. Self-reported diabetes tended to be more prevalent in the richest quintile, but the association did not reach statistical significance. Even without statistical separation, the direction of travel is clinically familiar in transitional settings: diabetes may appear more often where diagnosis is more likely, where screening is more accessible, and where people live long enough to encounter detection through the health system. Self-reported disease, by definition, reflects both biology and access to recognition. In communities with uneven contact with care, prevalence can look like a disease map when it is also a map of diagnosis.
Age, sex, educational attainment, and employment status were associated with several NCD risk factors across the analysis. That clustering reinforces a central point of the study: socioeconomic inequality does not act through a single pathway. It works through behavior, opportunity, access, and the way daily life is organized around work, mobility, and health services.
What the pattern means for prevention in rural settings
The results fit the broader understanding that low- and middle-income countries rarely show a simple rich-poor gradient for all chronic disease risks. Some exposures remain concentrated in disadvantaged groups, while others rise with affluence during periods of epidemiologic transition. In the DIMAMO cohort, higher wealth was linked to more harmful alcohol use and lower physical inactivity, while lower education and female sex were associated with less tobacco use. The mix is uneven, but it is not random. It reflects a population in transition, where economic change can reduce some forms of risk while intensifying others.
For primary care and public health practice, the implication is straightforward: prevention programs in rural South Africa need to be selective, not generic. Alcohol control, physical activity promotion, tobacco cessation, and chronic disease screening do not belong in the same social box just because they all contribute to NCD burden. Each risk factor sits in a different social pattern, and each needs a different entry point.
The study is cross-sectional, so it cannot establish causality or show how these inequalities evolved over time. It also relied on self-reported hypertension and diabetes, which may undercount undiagnosed disease or reflect different access to screening across socioeconomic groups. Even so, the evidence is strong enough to show that chronic disease prevention in this rural South African setting cannot rely on a one-size-fits-all model. The next step is longitudinal work that can separate enduring social risk from the changing visibility of disease in the health system.