Trends and risk factors for non-communicable diseases mortality in Nairobi slums (2008–2017)

Abstract

Introduction: Tracking progress in reaching global targets for reducing premature mortality from non-communicable diseases (NCDs) requires accurately collected population based longitudinal data. However, most African countries lack such data because of weak or non-existent civil registration systems. We used data from the Nairobi Urban Health and Demographic Surveillance System (NUDSS) to estimate NCD mortality trends over time and to explore the determinants of NCD mortality. Methods: Deaths identified in the NUHDSS were followed up with a verbal autopsy to determine the signs and symptoms preceding the death. Causes of death were then assigned using InSilicoVA algorithm. We calculated the rates of NCD mortality in the whole NUHDSS population between 2008 and 2017, looking at how these changed over time. We then merged NCD survey data collected in 2008, which contains information on potential determinants of NCD mortality in a sub-sample of the NUHDSS population, with follow up information from the full NUHDSS including whether any of the participants died of an NCD or non-NCD cause. Poisson regression models were used to identify independent risk factors (broadly categorized as socio-demographic, behavioural and physiological) for NCD mortality, as well as non-NCD mortality. Results: In the total NUHDSS population of adults age 18 and over, 23% were assigned an NCD as the most likely cause of death. There was evidence that NCD mortality decreased over the study period, with rates of NCD mortality dropping from 1.32 per 1000 person years in 2008–10 (95% CI: 1.13–1.54) to 0.93 per 1000 person years in 2014–17 (95% CI: 0.80–1.08). Of 5115 individuals who participated in the NCD survey in 2008, 421 died during the follow-up period of which 43% were attributed to NCDs. Increasing age, lower education levels, ever smoking and having high blood pressure were identified as independent determinants of NCD mortality in multivariate analyses. Conclusion: We found that NCDs account for one-quarter of mortality in Nairobi slums, although we document a reduction in the rate of NCD mortality over time. This may be attributed to increased surveillance and introduction of population-wide NCD interventions and health system improvements from research activities in the slums. To achieve further decline there is a need to strengthen health systems to respond to NCD care and prevention along with addressing social factors such as education.

Keywords

• NCD mortality trends
• Risk factors
• Informal settlements
• Kenya