利用机器学习模型更准确地预测初次全关节置换术的手术时长
该回顾性研究纳入2017年7月至2022年12月间的10,590例全髋关节置换术(THA)和12,179例全膝关节置换术(TKA),开发了线性岭回归、随机森林、XGBoost和可解释提升机四种机器学习模型预测手术时间。XGBoost模型表现最优,预测TKA和THA手术时间均优于其他模型。模型使TKA手术排程改善181个时间块(2,715分钟),THA改善138个时间块(2,070分钟)。作者认为随着机器学习模型改进,外科医生可考虑在手术排程中利用机器学习,但既往30例手术时间中位数可作为足够的替代方案。
要点
- 最具预测力的模型输入是「既往30例手术时间中位数」。
- XGBoost模型在预测TKA和THA手术时间上均优于其他模型。
- 模型使TKA排程改善181个时间块(2,715分钟),THA改善138个时间块(2,070分钟)。
核心结论
- 机器学习在预测全关节置换术手术时间方面优于传统排程方法。
- 既往30例手术时间中位数对排程时间准确性影响最大。
- 随着机器学习模型改进,外科医生可考虑在手术排程中利用机器学习,但既往30例手术时间中位数可作为足够的替代方案。
研究设计与样本
- 回顾性识别2017年7月至2022年12月间的10,590例THA和12,179例TKA。
- 病例按时间顺序分为训练集、验证集和测试集,测试集包括1,588例TKA和1,204例THA。
- 开发了四种机器学习算法:线性岭回归(LR)、随机森林、XGBoost和可解释提升机,并将每个模型的病例时间估计与以15分钟为单位的「等待」时间块(「预订不足」)和「超额」时间块(「预订过多」)衡量的计划估计进行比较。
- 记录手术病例时间,并为患者特征、手术信息和患者医疗状况分配SHAP值以理解特征重要性。
主要发现
- XGBoost模型在预测TKA和THA手术时间上均优于其他模型。
- 模型使TKA「超额时间块」减少85个(P < 0.001),「等待时间块」减少96个(P < 0.001)。
- 模型未显著减少THA的「超额时间块」(P = 0.89),但显著减少「等待时间块」134个(P < 0.001)。
- 模型总共改善TKA手术排程181个时间块(2,715分钟),改善THA手术排程138个时间块(2,070分钟)。
临床怎么用
- 作者认为,随着机器学习模型改进,外科医生可考虑在手术排程中利用机器学习;但既往30例手术时间中位数可作为足够的替代方案。
局限性
- 原文未报告局限性
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
原文
BACKGROUND: Accurate operative scheduling is essential for the appropriation of operating room esources. We sought to implement a machine learning model to predict primary total hip arthroplasty (THA) and total knee arthroplasty (TKA) case time. METHODS: A total of 10,590 THAs and 12,179 TKAs between July 2017 and December 2022 were retrospectively identified. Cases were chronologically divided into training, validation, and test sets. The test set cohort included 1,588 TKAs and 1,204 THAs. There were four ML algorithms developed: linear ridge regression (LR), random forest, XGBoost, and explainable boosting machine. Each model's case time estimate was compared to the scheduled estimate measured in 15-minute "wait" time blocks ("underbooking") and "excess" time blocks ("overbooking"). Surgical case time was recorded, and SHAP values were assigned to patient characteristics, surgical information, and the patient's medical condition to understand feature importance. RESULTS: The most predictive model input was "median previous 30 procedure case times." The XGBoost model outperformed the other models in predicting both TKA and THA case times. The model reduced TKA 'excess time blocks' by 85 blocks (P < 0.001) and 'wait time blocks' by 96 blocks (P < 0.001). The model did not significantly reduce 'excess time blocks' in THA (P = 0.89) but did significantly reduce 'wait time blocks' by 134 blocks (P < 0.001). In total, the model improved TKA operative booking by 181 blocks (2,715 minutes) and THA operative booking by 138 blocks (2,070 minutes). CONCLUSIONS: Machine learning outperformed a traditional method of scheduling total joint arthroplasty cases. The median time of the prior 30 surgical cases was the most influential on scheduling case time accuracy. As ML models improve, surgeons should consider ML utilization in case scheduling; however, prior 30 surgical cases may serve as an adequate alternative.
引用
- DelliCarpini G, Passano B, Yang J, et al. Utilization of Machine Learning Models to More Accurately Predict Case Duration in Primary Total Joint Arthroplasty. J Arthroplasty. 2024. PMID: 39477036. doi:10.1016/j.arth.2024.10.100
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。