Cadernos de Saude Publica

Combined social factors linked to death risk in older Brazilians

Differences between groups defined by sex, race or skin color, education and income accounted for 7% of mortality variation in an unadjusted model.

Women stand among white silhouettes beneath an open book, beside a denser group of men beneath scattered coins.
Cohort study of 8,918 Brazilians aged 50 years or older, median 4.8 yearsPopulations

Limit: Starting at 50 years excludes early deaths disproportionately affecting non-white people, potentially understating racial inequalities

Edited by Boris Djordjevic · AgingResearch.News

Follow

A prospective observational cohort study followed 8,918 Brazilian adults aged 50 or older between 2015 and 2021. Researchers grouped participants into 32 categories combining sex, self-reported race or skin color, education and household income per person. They linked participants’ records to Brazil’s mortality records. During a median follow-up of 4.8 years, 817 participants died.

After accounting for age, health, behavior and other factors, some groups of women with one to four years of education still had higher death risk than the model expected. Some groups of high-income non-white men had lower risk than expected. These included men with no formal education and men with five to ten years of education.

Why it matters

The authors called for policies addressing the social structures behind unequal mortality in later life, rather than considering each social disadvantage separately.

Caveats

This observational study identified associations, not causes. It excluded 494 people with missing social-factor data, and its findings concern Brazilian adults aged 50 or older.

From the paper

Follow-up

The final sample of 8,918 participants had a median follow-up period of 4.8 years (interquartile range: 4.1-5.4 years).

Results

Result

… but presented a statistically significant association as protective factors in the frailty model (FHR = 0.82; 95% confidence interval - 95%CI: 0.72-0.93 for rural area), although with overlap of the confidence intervals between the estimates (HR and FHR).

Results

Quoted word for word from the full text on PubMed Central.