腹壁疝修补结局报告应用ORACLE的开发与验证
基于美洲疝学会质量协作组中接受择期腹壁疝修补且有30天随访数据的患者,开发并验证了面向医患沟通的ORACLE预测工具,针对术后伤口事件、住院时长、30天非计划再入院及1年复发等5项质量指标建立预测模型。共纳入10690例患者,各模型经校准曲线比较显示具有较强预测能力。
要点
- ORACLE基于10690例择期腹壁疝修补患者数据开发与验证。
- 工具预测5项结局:伤口事件、住院时长、30天再入院、1年复发。
- 模型校准与预测强度经Brier评分、c-index及R2评估。
核心结论
- ORACLE可用于促进腹壁疝修补术前医患沟通与个体化干预决策。
- 模型基于全国数据集开发,并计划随协作组数据持续更新再评估。
研究设计与样本
- 纳入美洲疝学会质量协作组中所有接受择期腹壁疝修补且有30天随访数据的患者。
- 采用bootstrapping与偏倚校正校准生成并验证5项术后质量指标的预测模型。
- 共10690例患者符合纳入标准。
主要发现
- 5项结局模型的理论最佳校准曲线与观察校准曲线比较显示较强预测强度。
- 预测强度通过手术部位感染、需操作干预的手术部位事件及30天再入院的Brier评分体现。
- 预测强度通过1年疝复发的c-index及预测住院时长的R2值体现。
关键数值与推荐
- 预测结局包括术后伤口事件、估计住院时长、30天非计划再入院及术后1年疝复发风险。
- 原文未报告Brier评分、c-index及R2的具体数值。
临床怎么用
- 可在术前使用ORACLE向腹壁疝患者展示个体化结局风险以辅助共同决策。
- 可依据患者独特因素 tailor 干预措施,但需随协作组数据更新持续再评估模型。
局限性
- 研究基于美洲疝学会质量协作组数据,外部人群适用性需进一步验证。
- 原文未报告模型具体预测性能数值及外部验证结果。
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: Patient engagement is an increasingly important component of surgical decision making. Given the many factors associated with successful ventral hernia repair (VHR), we developed and validated the Outcomes Reporting App for Clinical and Patient Engagement (ORACLE) tool to help facilitate preoperative surgeon-patient discussions about VHR. METHODS: All patients undergoing elective, VHR with 30-day follow-up data available within the Americas Hernia Society Quality Collaborative were eligible for study inclusion. Using bootstrapping and bias-corrected calibration, predictive models were generated and validated for 5 quality measures after VHR, including postoperative wound events, estimated length of hospital stay, unplanned 30-day readmission to the hospital, and risk of hernia recurrence at 1 year postoperatively. RESULTS: A total of 10,690 patients met inclusion criteria. Based on comparison of the theoretical best to the observed calibration curves, the models for each of the 5 outcomes of interest have strong predictive strength as reflected in the Brier score for surgical site infection, surgical site occurrence requiring procedural intervention, and 30-day hospital readmission, the c-index for 1-year hernia recurrence and the R2 value for the model for predicted hospital length of stay. CONCLUSIONS: Using a national data set for development, ORACLE can be used to facilitate patient engagement, with the goal of tailoring interventions for VHR given each patient's unique factors. With ongoing data input into the Americas Hernia Society Quality Collaborative and a continuous re-evaluation of these risk models, it is our intention that this tool will serve as an up-to-date resource for hernia surgeons and ventral hernia patients.
引用
- Haskins IN, Olson MA, Stewart TG, et al. Development and Validation of the Ventral Hernia Repair Outcomes Reporting App for Clinician and Patient Engagement (ORACLE). J Am Coll Surg. 2019. PMID: 31054912. doi:10.1016/j.jamcollsurg.2019.03.014
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。