Informal Food Vendors and the Obesogenic Food Environment of an Informal Settlement in Nairobi, Kenya: A Descriptive and Spatial Analysis

OBJECTIVES: The objective of this study was to characterize the food environment of an informal settlement in Nairobi, Kenya according to the obesogenic properties and spatial distribution of informal food vendors. METHODS: From July 15 to August 9, 2019, we identified 524 vendors in the informal settlement of Kibera. Guided by the NOVA classification system for food processing, we dichotomized foods sold into categories of high- vs. low-risk and protective vs. non-protective with respect to their propensity to promote weight gain. Then, we classified vendors into obesogenic risk groups based on a 1/3 cutpoint for the proportion of high-risk and protective foods sold: 1) low-risk, protective, 2) low-risk, non-protective, 3) high-risk, protective, 4) high-risk, non-protective. We calculated descriptive statistics of the types of foods sold according to vendor type and obesogenic risk group. To assess clustering by obesogenic risk, for each group, we calculated the mean distance to the nearest vendor of the same group, and used Ripley's K function and 100 Monte Carlo simulations to identify significant clustering. RESULTS: The 456 vendors in the analytic sample included approximately 30% stands, 29% kiosks, 17% market stalls, 13% hawkers, and 12% restaurants. Foods most commonly sold were sweets/confectionary (29% of vendors), raw vegetables (28%), fried starches (23%), and fruits (21%). Forty-four % of vendors were classified as low-risk, protective, 34% as high-risk, non-protective, 16% as low-risk, non-protective, and 6% as high-risk, protective. The mean distance (95% CI) to the nearest vendor of the same group was 26 m (21, 31) for vendors in the low-risk, protective group, 29 m (25, 33) in the high-risk, non-protective group, 114 m (88,139) in the high-risk, protective group, and 43 m (30, 56) in the low-risk, non-protective group. Clustering was significant for all obesogenic risk groups except high-risk, protective. CONCLUSIONS: Our findings demonstrate a duality of obesogenic and non-obesogenic foods in this environment. Clustering of vendors selling obesogenic foods highlights the need for strategies to ensure consistent access to health-promoting foods throughout informal settlements. FUNDING SOURCES: UNC-Chapel Hill Department of Nutrition; Robertson Scholars Leadership Program at Duke University; NHLBI predoctoral traineeship.