Rates and predictors of three-year mortality in older people in rural Tanzania

Citation: 
William K. Gray, Felicity Dewhurst, Matthew J. Dewhurst, Golda Orega, John Kissima, Paul Chaote, Richard W. Walker
Publication year: 
2015

Background

There are few data on mortality rates in the general elderly living in sub-Saharan Africa. We aimed to detail three-year mortality rates in a population of rural community-dwelling older adults in northern Tanzania.

Methods

We performed a community-based study of 2232 people aged 70 years and over living in Hai district, Tanzania. At baseline, participants underwent clinical assessment for disability, neurological disorders, hypertension, atrial fibrillation and memory problems. At three-year follow-up mortality data were collected. Mortality rates were compared to UK estimates.

Results

At follow-up, data were available for 1873 subjects (83.9%). Of those, 208 (11.1%, 95% CI 9.7–12.5) had died. The age-standardised mortality rate was 10.2% (95% CI 8.8–11.6). Age-standardised mortality rates were lower than estimated for the UK (13.9%). In Cox regression analysis, greater age, higher levels of functional disability, use of a walking aid, subjective report of memory problems, being severely underweight and being normotensive were significant predictors of mortality.

Conclusions

Those who survive to old age in Tanzania appear to have relatively low mortality rates. Physical and cognitive disabilities were strongly associated with mortality risk in this elderly community-dwelling population. The association between blood pressure and mortality merits further study.