老年风险评分(GeRi-Score)用于老年股骨近端骨折患者死亡率预测——来自老年创伤登记处(ATR-DGU)的开发与验证研究
本研究基于老年创伤登记处(ATR-DGU)将患者分为开发组与验证组,开发并验证了主要依赖常规数据的简易死亡预测工具GeRi-Score。最终模型AUC为0.727(95%CI 0.711–0.742),Hosmer-Lemeshow检验显示无显著拟合不足(p=0.07)。GeRi-Score能区分轻、中、高风险组,可能有助于分配医疗护理强度。
要点
- GeRi-Score为简易死亡预测工具,主要依赖常规数据,区分度可接受且无显著拟合不足。
- 开发集预测院内死亡率5.3% vs 观察5.3%,验证集预测5.4% vs 观察5.7%。
- GeRi-Score可区分轻、中、高风险组,可能用于围手术期医疗护理强度分配及质量管理基准工具。
核心结论
- GeRi-Score是简易死亡预测工具,区分度可接受且无显著拟合不足,可能有助于分配髋部骨折手术围手术期医疗护理强度,并可作为质量管理项目的基准工具。
研究设计与样本
- 基于老年创伤登记处(ATR-DGU)的患者,分为开发组和验证组,使用逻辑回归模型构建院内死亡模型并得出评分,用AIC和似然比检验比较候选模型,用AUC和Hosmer-Lemeshow检验评估模型质量。
主要发现
- 共纳入38,570例患者,几乎平均分布于开发集和验证集。
- 最终模型AUC为0.727(95%CI 0.711–0.742),AIC显示较基础模型偏差显著降低,Hosmer-Lemeshow检验无显著拟合不足(p=0.07)。
- GeRi-Score在开发集预测院内死亡率为5.3%,观察死亡率为5.3%;在验证集预测为5.4%,观察为5.7%。
- GeRi-Score能够区分轻度、中度和高风险组。
临床怎么用
- 作者认为GeRi-Score可能有助于分配髋部骨折手术的围手术期医疗护理强度,并可在质量管理项目中作为基准工具使用。
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
UNLABELLED: This study developed an easy-to-use mortality prediction tool, which showed an acceptable discrimination and no significant lack of fit. The GeRi-Score was able to predict mortality and could distinguish between mild, moderate and high risk groups. Therefore, the GeRi-Score might have the potential to distribute the intensity of medical care. PURPOSE: Several mortality-predicting tools for hip fracture patients are available, but all consist of a high number of variables, require a time-consuming evaluation and/or are difficult to calculate. The aim of this study was to develop and validate an easy-to-use score, which depends mostly on routine data. METHODS: Patients from the Registry for Geriatric Trauma were divided into a development and a validation group. Logistic regression models were used to build a model for in-house mortality and to obtain a score. Candidate models were compared using Akaike information criteria (AIC) and likelihood ratio tests. The quality of the model was tested using the area under the curve (AUC) and the Hosmer-Lemeshow test. RESULTS: 38,570 patients were included, almost equal distributed to the development and to the validation dataset. The AUC was 0.727 (95% CI 0.711 - 0.742) for the final model, AIC resulted in a significant reduction in deviance compared to the basic model, and the Hosmer-Lemeshow test showed no significant lack of fit (p = 0.07). The GeRi-Score predicted an in-house mortality of 5.3% vs. 5.3% observed mortality in the development dataset and 5.4% vs. 5.7% in the validation dataset. The GeRi-Score was able to distinguish between mild, moderate and high risk groups. CONCLUSIONS: The GeRi-Score is an easy-to-use mortality-predicting tool with an acceptable discrimination and no significant lack of fit. The GeRi-Score might have the potential to distribute the intensity of perioperative medical care in hip fracture surgery and can be used in quality management programs as benchmark tool.
引用
- Schoeneberg C, Heuser L, Rascher K, et al. The Geriatrics at Risk Score (GeRi-Score) for mortality prediction in geriatric patients with proximal femur fracture - a development and validation study from the Registry for Geriatric Trauma (ATR-DGU). Osteoporos Int. 2023. PMID: 36892634. doi:10.1007/s00198-023-06719-5
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。