美国外科医师学会NSQIP儿科按需风险调整平滑率报告的开发与评估
该研究开发了NSQIP儿科按需报告方法,利用5年历史数据构建 logistic 模型预测方程,实时生成平滑和风险调整的基准报告。比较2022年全年数据,按需报告与半年度报告在死亡率、并发症和手术部位感染上的基准结果几乎完全一致,加权 kappa 平均0.865,相关系数平均0.978。结论认为按需应用可补充半年度报告,更早提供医院性能信息。
要点
- 按需报告利用5年历史数据实时生成风险调整和平滑的基准报告。
- 与半年度报告比较,基准结果几乎完全一致,加权 kappa 平均0.865。
- 按需报告可更早提供医院性能信息,辅助质量改进。
核心结论
- 新引入的NSQIP儿科按需应用将补充半年度报告,更早提供基于模型的医院性能信息以改善手术结局。
研究设计与样本
- 比较所有手术、胃肠手术和普外科手术的死亡率、并发症和手术部位感染。
主要发现
- 按需与半年度报告的基准结果几乎完全一致,加权 kappa 统计量(除死亡率外)平均0.865,相关系数平均0.978。
关键数值与推荐
- 按需报告使用5年历史数据构建预测方程,支持实时生成平滑和风险调整的基准报告。
- 自定义过滤器允许用户为特定时间段和患者亚组构建模型,与固定人群结局率比较。
- 基于率的指标(风险调整率)及置信区间以图形描述医院性能。
临床怎么用
- NSQIP儿科医院可使用按需报告更早获取医院性能信息,补充半年度报告用于质量改进。
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 NSQIP Pediatric Semiannual Report (NSQIP Ped SAR) provides hospitals with risk-adjusted benchmarked results for comparative performance based on 1 year of data. These data are 6 to 18 months old due to requirements for data processing and modeling, and this delay potentially limits their usefulness for hospital surgical quality improvement efforts. A timelier reporting mechanism is needed. STUDY DESIGN: The "On-Demand" methodology uses 5 years of historical data to construct logistic model prediction equations for outcomes, which support the construction of smoothed and risk-adjusted benchmarking reports in real time. In addition, custom filters allow users to build models for time periods and unique patient subgroups of interest, allowing comparisons of hospital performance across different time periods to a fixed population outcome rate. A rate-based metric (risk-adjusted rates) along with CIs graphically describe hospital performance.SAR and On-Demand statistical benchmarking methodologies were compared for mortality, morbidity, and surgical site infection across all surgeries, gastrointestinal surgeries, and general surgeries, for patients with operation dates from January 1, 2022, to December 31, 2022. Differences in outlier detections and logged odds ratios from benchmarked results were evaluated using weighted kappa statistics and Pearson correlation coefficients. RESULTS: On-Demand predictive equations were built using data from January 1, 2018, to December 31, 2022. Comparisons of benchmarking results from the SAR and On-Demand from January 1, 2022, to December 31, 2022, demonstrated near-perfect agreement; weighted kappa statistics (excluding mortality) averaged 0.865, and correlations averaged 0.978. CONCLUSIONS: The newly introduced NSQIP Ped On-Demand application will augment SAR reporting, providing earlier information, on a model-by-model basis, for assessing hospital performance to improve surgical outcomes at NSQIP Ped hospitals.
引用
- Coleman M, Liu Y, Thompson V, et al. Development and Evaluation of American College of Surgeons NSQIP Pediatric On-Demand Reporting of Risk-Adjusted and Smoothed Rates. J Am Coll Surg. 2026. PMID: 40919788. doi:10.1097/XCS.0000000000001619
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。