Mathematical Model of Snakebite Dynamics in Turkana County and Outcome Based on Hospital Seeking Behavior

Abstract

Snakebite envenoming (SBE) is a public health concern of the poor tropical populations that results when venomous snakes introduce lethal toxins to victims either through a bite or spit in the eyes. Similar to other neglected tropical diseases (NTDs), snakebite programs in LMICs face challenges of inadequate data and statistics therefore making it difficult to make evidence based decisions towards public health interventions and resource mobilization/allocation. The currently available data from KHIS platform has greatly improved the situation by capturing snakebite cases reported at health facilities. However, the reported cases gravely under-estimates snakebite burden by failing to capture snakebite victims presenting to traditional healers and those seeking no treatment at all. This study investigate dynamics of snakebite disease and outcomes given access to treatment by incorporating both snake and human factors in a single model. Applying a mathematical modeling approach, the study borrows from household surveys to obtain a non-reporting rates for adjusting the KHIS data. The adjusted data is used to calibrate a logistic growth model estimation population of snakes at the human-snake interface. Logistic growth model fitness is enhanced through trajectory matching using maximum likelihood estimates of simulated data. Simulated data on snake population is introduced into a SIRS model that simulates snakebite occurrence for a period of 365 days, estimates snakebite incidences and outcome given access to treatment. The effects of varying model parameters is investigated to understand dynamics of snakebite and advice on potential interventions. Finally, the study investigates the impact of hospital seeking behavior in averting snakebite-associated deaths. Present study estimated 3,371 annual incidence of snakebite in Turkana County, amounting to 1,315 missed cases. The burden of snakebite from simulated data - 368/100,000 person-year (95% CI: 356-381) agrees with incidence rates from reported (adjusted) data - 364.99/100,000 person-year (95% CI: 353-378). Activities associated with increased exposure showed to half snakebite burden when the force of infection is reduced by 50%. The case fatality rate is estimated at 5.6 (4.2-7.4) deaths per 100,000 person-years with 52 deaths reported at the end of the simulation period (t = 365 days). There were 3.5% averted deaths for every 10% increase in the rate of hospital seeking amongst the snakebite victims. The model predicts 32.4% averted deaths if all victims were to present to health facility following snakebite, regardless of whether they receive antivenom or not. It is important therefore that, public health awareness interventions be instituted in snakebite endemic communities to encourage victims to seek hospital treatment following snakebite. The model is transferrable in similar settings with limited data and no statistics on the burden of snakebite disease.