综合关节置换捆绑支付模型对花费与照护终点影响的系统综述
该系统综述纳入23项研究,评估美国医疗保险CJR捆绑支付模型对全髋/膝关节置换的影响。多数研究发现90天花费显著下降(每例469至1322美元),主要源于减少机构性急性后期照护,临床结局未普遍恶化,但存在患者选择与差异扩大证据。
要点
- CJR与90天花费下降相关,降幅主要来自减少 skilled-nursing facility 等机构性急性后期照护。
- 多数研究未发现90天并发症或死亡率增加,部分报告再入院小幅下降。
- 多项分析提示患者选择偏向低风险、年轻及白人患者,黑人受益者和社会经济弱势地区手术率下降。
核心结论
- CJR实现了有意义的、主要由急性后期照护驱动的 episode 花费下降,且临床结局未广泛恶化。
- 但患者选择与差异扩大的证据提示该政策产生了不良分布效应。
- 未来 episode-based 改革应纳入社会风险调整与保障措施,在控费同时维护可及性与公平获益。
研究设计与样本
- 遵循PRISMA指南(PROSPERO CRD42024596111)的系统综述。
- 检索PubMed、Embase和Web of Science(2015年1月至2024年9月)的英文研究,比较CJR实施前后结局。
- 共纳入23项研究,主要为准实验性医疗保险索赔分析(2011至2020年),涵盖数百万THA/TKA episodes。
- 由两名评审独立筛选、提取预设变量并按领域综合 findings,采用 validated risk of bias tool 评估非随机研究的偏倚风险。
主要发现
- 16项评估90天花费的研究中,10项报告显著节省,每例平均估计花费降低范围为469至1322美元。
- 节省主要来自减少机构性急性后期照护,尤其是 skilled-nursing facilities。
- 部分分析提示节省随时间递减,且 reconciliation 后净节省有限。
- 多数研究未发现90天并发症或死亡率增加,若干研究报告再入院小幅减少,部分场景住院时长适度下降。
- 8项研究报告机构性急性后期照护出院减少、居家或居家健康使用增加,常伴随更少访视;多项分析提示患者选择偏向低风险、年轻及白人患者,黑人受益者和社会经济弱势地区手术率下降。
关键数值与推荐
- CJR捆绑支付模型实施期为2016至2024年,旨在提升THA/TKA的价值。
- 16项评估90天花费的研究中10项报告显著节省,每例平均估计花费降低范围为469至1322美元。
- 总体偏倚风险因观察性设计常见为低至中等,证据等级为Level III。
临床怎么用
- 骨科智能体医生可参考CJR证据,关注减少不必要的机构性急性后期照护以控制 episode 花费。
- 在临床决策中需警惕患者选择偏倚,避免因社会经济或种族因素影响THA/TKA可及性。
- 未来政策设计应纳入社会风险调整与保障措施,平衡控费与公平获益。
局限性
- 纳入研究主要为准实验性观察设计,总体偏倚风险常见为低至中等。
- 部分分析提示节省随时间递减及 reconciliation 后净节省有限,证据存在异质性。
- 存在患者选择与差异扩大的证据,提示政策不良分布效应。
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: The Centers for Medicare & Medicaid Services implemented the Comprehensive Care for Joint Replacement (CJR) bundled-payment model (2016 to 2024) to improve value for total hip arthroplasty and knee arthroplasty (THA/TKA). We systematically reviewed the model's impact on episode spending, clinical outcomes, discharge destination, and total joint arthroplasty utilization. METHODS: Following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (PROSPERO CRD42024596111), we searched PubMed, Embase, and Web of Science (January 2015 to September 2024) for English language studies comparing outcomes before versus after CJR. There were two reviewers who independently screened records, extracted prespecified variables, and synthesized findings by domain. Risk of bias for nonrandomized studies was assessed with a validated risk of bias tool. RESULTS: There were 23 studies included, predominantly quasi-experimental Medicare claims analyses (2011 to 2020) encompassing millions of THA/TKA episodes. Of 16 studies that assessed 90-days costs, 10 reported major savings (mean estimated spending reductions ranged from $469 to $1,322 per episode), largely from reduced institutional postacute care (PAC), especially skilled-nursing facilities. Some analyses suggested diminishing savings over time and limited net savings after reconciliation. Most studies found no increase in 90-days complications or mortality, with several reporting small reductions in readmissions. Lengths of stay decreased modestly in some settings. There were eight studies that had lower discharge to institutional PAC and greater home or home-health usage, often with fewer visits. Multiple analyses suggested patient selection favoring lower-risk, younger, and White patients; procedure rates declined among Black beneficiaries and in socioeconomically disadvantaged areas. Overall risk of bias was commonly low to moderate given observational designs. CONCLUSIONS: The CJR achieved meaningful, primarily PAC-driven reductions in episode spending without broad deterioration in clinical outcomes. Evidence of patient selection and widening disparities, however, suggests that the policy engendered adverse distributional effects. Future episode-based reforms should incorporate social risk adjustment and guardrails to preserve access and equitable benefit alongside cost containment. LEVEL OF EVIDENCE: Level III.
引用
- Sontag-Milobsky I, Knebel R, Edelstein AI, et al. Impact of the Comprehensive Care for Joint Replacement Model on Spending and Care Endpoints: A Systematic Review. J Arthroplasty. 2026. PMID: 42456787. doi:10.1016/j.arth.2026.07.005
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。