Differentiation of bites of venomous snakes in Taiwan using the lens model approach

Abstract

Because the signs and symptoms exhibited after bites of Taiwan's six most common venomous snakes are confusing and inconsistent, rapid and accurate judgment of these snake bites is difficult. We developed a decision-support system using the lens model approach to distinguish snake bites. The occurrence rates of the nine most commonly exhibited signs and symptoms of seven snake bite categories reported in the literature were utilized to generate simulated snake bite cases for training and testing the models. The developed models effectively predicted the bites of Vipera russelli formosensis (discriminability index (d′) = 4.84) and Bungarus multicinctus (d′ = 4.01), but poorly predicted the bites of Naja naja atra, Trimeresurus mucrosquamatus, and undefined snake (d′ < 1.4). This study demonstrates the potential of the lens model approach in differentiating bites among the six venomous snakes. Future research should include empirical data with more clinical information to enhance model performance.