机器学习模型预测全膝关节翻修术后出院去向的验证与可推广性
该研究基于全国数据库52,533例和本机构数据库1,628例全膝关节翻修术患者(非居家出院率分别为20.6%和19.4%),训练并内部验证五种机器学习模型,随后在本机构数据集中进行外部验证。所有模型在外部验证中均表现出良好至优秀的区分度、校准度和临床效用,其中人工神经网络预测性能最佳(AUC=0.78),证实了基于全国数据库开发的机器学习模型的可推广性。
要点
- 五种机器学习模型在全国数据库训练后,于本机构数据外部验证中均表现良好至优秀的区分度、校准度和临床效用。
- 人工神经网络是预测非居家出院的最佳模型(AUC=0.78),校准斜率0.93、截距0.02、Brier评分0.12。
- 非居家出院的最强预测因素为患者年龄、体重指数和手术指征。
核心结论
- 所有五种机器学习模型在外部验证中均表现出良好至优秀的区分度、校准度和临床效用,人工神经网络是预测全膝关节翻修术后出院去向的最佳模型。
- 研究结果证实了基于全国数据库数据开发的机器学习模型的可推广性。
- 将这些预测模型整合到临床工作流程中,可能有助于优化与全膝关节翻修术相关的出院计划、床位管理和成本控制。
研究设计与样本
- 全国队列52,533例患者,本机构队列1,628例患者,非居家出院率分别为20.6%和19.4%。
- 在全国大型数据集中训练并内部验证(五折交叉验证)五种机器学习模型,随后在本机构数据集上进行外部验证。
- 采用区分度、校准度和临床效用评估模型性能,并使用全局预测因子重要性图和局部替代模型进行解释。
主要发现
- 非居家出院的最强预测因素为患者年龄、体重指数和手术指征。
- 受试者工作特征曲线下面积从内部验证到外部验证有所增加,范围在0.77至0.79之间。
- 人工神经网络是识别非居家出院风险患者的最佳预测模型(AUC=0.78),也是最准确的模型(校准斜率=0.93,截距=0.02,Brier评分=0.12)。
关键数值与推荐
- 全国队列非居家出院率为20.6%,本机构队列非居家出院率为19.4%。
- 人工神经网络模型AUC=0.78,校准斜率=0.93,截距=0.02,Brier评分=0.12。
- 受试者工作特征曲线下面积范围在0.77至0.79之间。
临床怎么用
- 作者认为,将这些预测模型整合到临床工作流程中,可能有助于优化出院计划、床位管理和成本控制。
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: Postoperative discharge to facilities account for over 33% of the $ 2.7 billion revision total knee arthroplasty (TKA)-associated annual expenditures and are associated with increased complications when compared to home discharges. Prior studies predicting discharge disposition using advanced machine learning (ML) have been limited due to a lack of generalizability and validation. This study aimed to establish ML model generalizability by externally validating its prediction for nonhome discharge following revision TKA using national and institutional databases. METHODS: The national and institutional cohorts comprised 52,533 and 1,628 patients, respectively, with 20.6 and 19.4% nonhome discharge rates. Five ML models were trained and internally validated (five-fold cross-validation) on a large national dataset. Subsequently, external validation was performed on our institutional dataset. Model performance was assessed using discrimination, calibration, and clinical utility. Global predictor importance plots and local surrogate models were used for interpretation. RESULTS: The strongest predictors of nonhome discharge were patient age, body mass index, and surgical indication. The area under the receiver operating characteristic curve increased from internal to external validation and ranged between 0.77 and 0.79. Artificial neural network was the best predictive model for identifying patients at risk for nonhome discharge (area under the receiver operating characteristic curve = 0.78), and also the most accurate (calibration slope = 0.93, intercept = 0.02, and Brier score = 0.12). CONCLUSION: All five ML models demonstrated good-to-excellent discrimination, calibration, and clinical utility on external validation, with artificial neural network being the best model for predicting discharge disposition following revision TKA. Our findings establish the generalizability of ML models developed using data from a national database. The integration of these predictive models into clinical workflow may assist in optimizing discharge planning, bed management, and cost containment associated with revision TKA.
引用
- Buddhiraju A, Chen TL, Subih MA, et al. Validation and Generalizability of Machine Learning Models for the Prediction of Discharge Disposition Following Revision Total Knee Arthroplasty. J Arthroplasty. 2023. PMID: 36849013. doi:10.1016/j.arth.2023.02.054
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。