文献速递关节外科

强化质量测量以预测全膝关节置换术成功:来自全国代表性全膝关节置换术队列的结果

Strengthening Quality Measurement to Predict Success for Total Knee Arthroplasty: Results from a Nationally Representative Total Knee Arthroplasty Cohort.
PubMed 文献 · J Bone Joint Surg Am · 2024-01-25

基于全国代表性初次单侧全膝关节置换术队列共5958例患者,研究者在紧密遵循美国医疗保险和医疗补助服务中心(CMS)提出的质量测量指标所构建的原始模型基础上,仅增加基线KOOS, JR一个预测因子构建增强模型,以预测术后1年达到KOOS, JR至少20分改善的成功结局。增强模型将AUC从0.58提高至0.73,且在预测成功率最高十分位中94%达到成功、最低十分位中34%达到成功,而原始模型对应为77%与57%。

要点

  • 增强模型仅增加基线KOOS, JR一个预测因子,AUC由0.58升至0.73。
  • 增强模型预测成功率最高十分位94%达到成功,最低十分位34%达到成功。
  • 原始模型区分度较低:最高十分位77%成功,最低十分位57%成功。

核心结论

  • 在具有全国代表性的队列中,将基线膝关节特异性PROM评分加入质量测量模型可大幅提高其预测能力,消除天花板效应和地板效应以及对可识别患者亚组的错误预测。
  • 增强模型既不偏向也不阻碍对膝关节功能较差患者的诊疗,且无需采集新数据。

研究设计与样本

  • 研究使用接受初次单侧全膝关节置换术的全国队列患者数据,开发了一个紧密遵循CMS提出的质量测量指标的原始模型,以及仅增加基线KOOS, JR一个预测因子的增强模型。
  • 研究纳入5958例患者,平均年龄67岁;63%为女性,93%为白人,87%为超重或肥胖。
  • 成功定义为术后1年KOOS, JR达到至少20分改善(即达到实质性临床获益)。
  • 采用受试者工作特征曲线下面积(AUC)和观察结局与预期(模型预测)结局比值(O:E比值)评估各模型性能。
  • 证据等级为预后Level II。

主要发现

  • 加入基线KOOS, JR使AUC从0.58提高到0.73。
  • 增强模型预测成功率最高十分位中94%达到成功,最低十分位中34%达到成功。
  • 原始模型对应数值区分度较低:最高十分位77%达到成功,最低十分位57%达到成功。
  • 仅增强模型能在全部基线评分范围内准确预测成功。
  • 仅对年龄≥65岁患者重复分析时,结果几乎完全相同。

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: When performed well on appropriate patients, total knee arthroplasty (TKA) can dramatically improve quality of life. Patient-reported outcome measures (PROMs) are increasingly used to measure outcome following TKA. Accurate prediction of improvement in PROMs after TKA potentially plays an important role in judging the surgical quality of the health-care institutions as well as informing preoperative shared decision-making. Starting in 2027, the U.S. Centers for Medicare & Medicaid Services (CMS) will begin mandating PROM reporting to assess the quality of TKAs. METHODS: Using data from a national cohort of patients undergoing primary unilateral TKA, we developed an original model that closely followed a CMS-proposed measure to predict success, defined as achieving substantial clinical benefit, specifically at least a 20-point improvement on the Knee injury and Osteoarthritis Outcome Score, Joint Arthroplasty (KOOS, JR) at 1 year, and an enhanced model with just 1 additional predictor: the baseline KOOS, JR. We evaluated each model's performance using the area under the receiver operator characteristic curve (AUC) and the ratio of observed to expected (model-predicted) outcomes (O:E ratio). RESULTS: We studied 5,958 patients with a mean age of 67 years; 63% were women, 93% were White, and 87% were overweight or obese. Adding the baseline KOOS, JR improved the AUC from 0.58 to 0.73. Ninety-four percent of those in the top decile of predicted probability of success under the enhanced model achieved success, compared with 34% in its bottom decile. Analogous numbers for the original model were less discriminating: 77% compared with 57%. Only the enhanced model predicted success accurately across the spectrum of baseline scores. The findings were virtually identical when we replicated these analyses on only patients ≥65 years of age. CONCLUSIONS: Adding a baseline knee-specific PROM score to a quality measurement model in a nationally representative cohort dramatically improved its predictive power, eliminating ceiling and floor effects and mispredictions for readily identifiable patient subgroups. The enhanced model neither favors nor discourages care for those with greater knee dysfunction and requires no new data collection. LEVEL OF EVIDENCE: Prognostic Level II . See Instructions for Authors for a complete description of levels of evidence.

来源:J Bone Joint Surg Am,查看原文。版权归原作者所有。

引用

  1. Zheng H, Ash AS, Yang W, et al. Strengthening Quality Measurement to Predict Success for Total Knee Arthroplasty: Results from a Nationally Representative Total Knee Arthroplasty Cohort. J Bone Joint Surg Am. 2024. PMID: 38271493. doi:10.2106/JBJS.23.00602

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。