Data Science Institute, Southern Methodist University, Dallas, TX, USA.
Rocky Mountain Poison & Drug Safety, Denver Health and Hospital Authority, Denver, CO, USA.
BMC Pharmacol Toxicol. 2022 Jul 13;23(1):49. doi: 10.1186/s40360-022-00588-0.
With diabetes incidence growing globally and metformin still being the first-line for its treatment, metformin's toxicity and overdose have been increasing. Hence, its mortality rate is increasing. For the first time, we aimed to study the efficacy of machine learning algorithms in predicting the outcome of metformin poisoning using two well-known classification methods, including support vector machine (SVM) and decision tree (DT).
This study is a retrospective cohort study of National Poison Data System (NPDS) data, the largest data repository of poisoning cases in the United States. The SVM and DT algorithms were developed using training and test datasets. We also used precision-recall and ROC curves and Area Under the Curve value (AUC) for model evaluation.
Our model showed that acidosis, hypoglycemia, electrolyte abnormality, hypotension, elevated anion gap, elevated creatinine, tachycardia, and renal failure are the most important determinants in terms of outcome prediction of metformin poisoning. The average negative predictive value for the decision tree and SVM models was 92.30 and 93.30. The AUC of the ROC curve of the decision tree for major, minor, and moderate outcomes was 0.92, 0.92, and 0.89, respectively. While this figure of SVM model for major, minor, and moderate outcomes was 0.98, 0.90, and 0.82, respectively.
In order to predict the prognosis of metformin poisoning, machine learning algorithms might help clinicians in the management and follow-up of metformin poisoning cases.
随着全球糖尿病发病率的上升，以及二甲双胍仍然是治疗该病的一线药物，二甲双胍的毒性和过量使用也在增加。因此，其死亡率也在上升。我们首次旨在使用两种著名的分类方法，包括支持向量机（SVM）和决策树（DT），研究机器学习算法在预测二甲双胍中毒结果中的功效。
本研究是对美国最大的中毒病例数据库国家毒物数据系统（NPDS）数据进行的回顾性队列研究。使用训练和测试数据集开发了 SVM 和 DT 算法。我们还使用了精度-召回和 ROC 曲线以及曲线下面积值（AUC）来评估模型。
我们的模型表明，酸中毒、低血糖、电解质异常、低血压、阴离子间隙升高、肌酐升高、心动过速和肾衰竭是预测二甲双胍中毒结果的最重要决定因素。决策树和 SVM 模型的平均阴性预测值分别为 92.30 和 93.30。决策树对主要、次要和中度结果的 ROC 曲线的 AUC 分别为 0.92、0.92 和 0.89，而 SVM 模型对主要、次要和中度结果的 AUC 分别为 0.98、0.90 和 0.82。
为了预测二甲双胍中毒的预后，机器学习算法可能有助于临床医生管理和随访二甲双胍中毒病例。
