Assessment of cardiovascular risk in a slum population in Kenya: use of World Health Organisation/International Society of Hypertension (WHO/ISH) risk prediction charts - secondary analyses of a household survey

Objectives Although cardiovascular disease (CVD) is of growing importance in low- and middle-income countries (LMICs), there are conflicting views regarding CVD as a major public health problem for the urban poor, including those living in slums. We examine multivariable risk prediction in a slum population and assess the number of cardiovascular related deaths within 10 years of application of the tool. Setting We use data from the Nairobi Urban Health and Demographic Surveillance System (NUHDSS) population (residents of two slum communities) between May 2008 and April 2009. Design This is a secondary data analysis from a cross-sectional survey. We use the WHO/International Society of Hypertension (WHO/ISH) cardiovascular risk prediction tool to examine 10-year risk of major CVD events in a slum population. CVD deaths in the cohort, reported up until June 2018 and identified through verbal autopsy are also presented. Participants 3063 men and women aged over 40 years with complete data for variables needed for the WHO/ISH risk prediction tool were eligible to take part. Results The majority of study members (2895, 94.5%) were predicted to have 'low' risk (<10%) of a cardiovascular event over the next 10 years and just 51 (1.7%) to have 'high' CVD risk (>= 20%). 91 CVD deaths were reported for the cohort up until June 2018. Of individuals classified as low risk, 74 (2.6%) were identified as having died of CVD. Nine (7.7%) individuals classified at 10% to 20% risk and eight (15.9%) classified at >20% were identified as dying of CVD. Conclusions This study shows that there is a low risk profile of CVD in this slum population in Nairobi, Kenya, in comparison to results from application of multivariable risk prediction tools in other LMIC populations. This has implications for health service planning in these contexts.