Details of multilevel regression analyses.

Introduction

Snakebite is an important public health concern, especially in tropical areas, but the true burden remains unclear due to sub-optimal reporting and over-reliance on health facility-based data.

Methods

A community-based cross-sectional survey was conducted in Samburu County, Kenya from December 2019 to March 2020. Geospatial techniques were used to create a sampling frame of all households in Samburu County and a multistage cluster sampling strategy to select households and recruit study participants. Five year prevalence and mortality rates were estimated, the characteristics and circumstances of snakebite were described, and multilevel logistic regression models were built to identify independent risk factors for snakebite.

Results

We recruited 3,610 individuals living in 875 households from 30 clusters. The 5-year prevalence of snakebite was 2.2% (95% CI 1.4%–3.4%), and the 5-year mortality rate was 138 (95% CI 44–322) deaths per 100,000 inhabitants, resulting in an estimated 1,406 snakebites and 88 deaths from snakebites per year in Samburu County. Snakebite incidents often occurred at night between 9pm and 6 am (44%, n = 36), and the participants were mostly walking/playing outdoors (51%, n = 41) or sleeping (32%, n = 27) when they were bitten. Lower household socioeconomic status and smaller numbers of people per house were significant independent risk factors.

Conclusion

Samburu County has a high snakebite burden and the most victims are bitten while sleeping or walking outdoors at night. Snakebite prevention and health promotion programs in Samburu County, and other endemic regions, need to be contextualised and consider the geographic, seasonal, and temporal specificities found in our study. Our findings also have implications for health care delivery, especially identification of the need for night-time staffing with expertise in snakebite management and antivenom availability to better manage patients and thereby improve outcomes.