Estimating real snakebite incidence in Colombia by using mathematical modelling and statistical inference

Abstract

Snakebite envenoming is a Neglected Tropical Disease affecting mainly deprived populations. Its burden is normally underestimated because patients prefer to seek for traditional medicine. Thus, applying strategies to optimize disease' management and treatment delivery is difficult. We propose a framework to estimate snakebite incidence at a fine political scale based on available data, testing it in Colombia. First, we produced snakebite fine-scale risk maps based on the most medically important venomous snake species (Bothrops asper and B. atrox). We validated them with reported data in the country. Then, we proposed a generalized mixed effect model that estimates total incidence based on produced risk maps, poverty indexes, and an accessibility score that reflects the struggle to reach a medical center. Finally, we calibrated our model with national snakebite reported data from 2010 to 2019 using a Markov chain Monte Carlo (MCMC) algorithm and estimated underreporting based on the total incidence estimation. Our results suggest that 10.3% of total snakebite cases are not reported in Colombia and do not seek medical attention. The Orinoco and Amazonian regions (east of Colombia) share a high snakebite risk with a high underreporting. Our work highlights the importance of multidisciplinary approaches to face snakebite.