 Our study demonstrated that light gradient boosting machine, LGBM, Logistic Regression, LR and Random Forest, RF models have high predictive power in the outcome prediction of diphenhydramine, DPH, poisoning. These models achieved an average precision recall area under the curve AUC of 0.84 which indicates that they are highly accurate in identifying patients at risk of experiencing adverse effects due to DPH ingestion. Furthermore, the specificity of these models was 87% indicating that they are able to accurately identify patients without any adverse effects. Additionally, the recall sensitivity of these models ranged between 73% and 75% with an F1 score of 75%. Overall, our study suggests that LGBM, LR and RF models could be used as effective tools for the early detection of DPH poisoning. This article was authored by Aamid Mahapur, Farhad Saidi, Jafar Abdullahi and others.