University of Michigan Medical School.
From the Departments of Bioinformatics.
Plast Reconstr Surg. 2024 Sep 1;154(3):573-580. doi: 10.1097/PRS.0000000000011099. Epub 2023 Sep 29.
The aim of this study was to evaluate the use of machine learning to predict persistent opioid use after hand surgery.
The authors trained 2 algorithms to predict persistent opioid use, first using a general surgery data set and then using a hand surgery data set, resulting in 4 trained models. Next, the authors tested each model's performance using hand surgery data. Participants included adult surgery patients enrolled in a cohort study at an academic center from 2015 to 2018. The first algorithm (Michigan Genomics Initiative model) was designed to accommodate patient-reported data and patients with or without prior opioid use. The second algorithm (claims model) was designed for insurance claims data from patients who were opioid-naive only. The main outcome was model discrimination, measured by area under the receiver operating curve (AUC).
Of 889 hand surgery patients, 49% were opioid-naive and 21% developed persistent opioid use. Most patients underwent soft-tissue procedures (55%) or fracture repair (20%). The Michigan Genomics Initiative model had AUCs of 0.84 when trained only on hand surgery data, and 0.85 when trained on the full cohort of surgery patients. The claims model had AUCs of 0.69 when trained only on hand surgery data, and 0.52 when trained on the opioid-naive cohort of surgery patients.
Opioid use is common after hand surgery. Machine learning has the potential to facilitate identification of patients who are at risk for prolonged opioid use, which can promote early interventions to prevent addiction.
本研究旨在评估机器学习在预测手部手术后持续使用阿片类药物方面的应用。
作者使用两种算法对手部手术后持续使用阿片类药物进行预测，首先使用一般手术数据集，然后使用手部手术数据集，从而训练出 4 种模型。然后，作者使用手部手术数据测试每种模型的性能。研究对象为 2015 年至 2018 年在学术中心参加队列研究的成年手术患者。第一种算法（密歇根基因组倡议模型）旨在适应有或没有阿片类药物使用史的患者的报告数据。第二种算法（理赔模型）仅适用于阿片类药物初治患者的理赔数据。主要结局是模型区分度，通过接受者操作特征曲线下面积（AUC）进行衡量。
889 例手部手术患者中，49%为阿片类药物初治患者，21%发生持续使用阿片类药物。大多数患者接受软组织手术（55%）或骨折修复（20%）。密歇根基因组倡议模型仅在手部手术数据上训练时的 AUC 为 0.84，在整个手术患者队列上训练时的 AUC 为 0.85。理赔模型仅在手部手术数据上训练时的 AUC 为 0.69，在阿片类药物初治手术患者队列上训练时的 AUC 为 0.52。
手部手术后阿片类药物的使用较为常见。机器学习有可能有助于识别有长期阿片类药物使用风险的患者，从而可以促进早期干预以预防成瘾。
