基于ICD-10编码与机器学习的创伤相关死亡预测模型的推导
研究利用2015至2017年美国国家创伤数据库(National Trauma Data Bank)纳入创伤相关住院成人,将8,021个ICD-10编码归为1,495个唯一损伤变量,以院内死亡为主要结局,用XGBoost构建死亡预测模型,并与逻辑回归、ISS和TMPM-ICD10比较。共纳入1,611,063例患者,院内死亡54,870例(3.41%)。XGBoost的ROC曲线下面积为0.863(95%CI 0.862-0.864),高于逻辑回归0.845、ISS 0.828和TMPM-ICD10 0.861(均p<0.001),校准度也显著更优(R²=0.993)。该模型有望用于创伤严重度量化、死亡预后判断、质量改进和创伤研究。
要点
- XGBoost模型在1,611,063例创伤住院成人中预测院内死亡,ROC曲线下面积0.863(95%CI 0.862-0.864),优于ISS的0.828和TMPM-ICD10的0.861(均p<0.001)。
- 校准度方面XGBoost的R²=0.993,优于逻辑回归(0.981)、ISS(0.649)和TMPM-ICD10(0.830),提示预测概率与实际死亡率更吻合。
- 死亡患者较存活者更常见穿透性创伤且损伤数量更多,但原文未给出具体比例或阈值。
- 模型基于ICD-10编码,无需额外计算ISS所需的管理编码,但依赖完整编码数据,床旁实时应用需前瞻性验证。
- 证据等级为预后与流行病学level III,属回顾性数据库建模研究,不能直接作为床旁决策阈值。
适用人群与场景
- 2015至2017年美国国家创伤数据库(National Trauma Data Bank)中创伤相关住院的成人患者。
- 适用场景为创伤住院后的死亡预后量化、质量改进和创伤研究,非院前或抢救室实时决策场景。
诊断标准与关键阈值
- 主要结局为院内死亡(in-hospital mortality)。
- 损伤分类基于ICD-10编码:8,021个ICD-10编码归为1,495个唯一损伤变量。
- 原文未提及具体的评分界值或启动/停止阈值。
处置流程与药物剂量
- 原文未提及
红旗征与不能漏的情况
- 原文未提及
研究设计与关键结果
- 研究类型为回顾性数据库建模研究,使用2015至2017年国家创伤数据库,纳入1,611,063例创伤相关住院成人,院内死亡54,870例(3.41%)。
- 干预/方法:将8,021个ICD-10编码归为1,495个唯一损伤变量,用XGBoost构建死亡预测模型,并与逻辑回归、ISS和TMPM-ICD10比较。
- 主要终点:XGBoost的ROC曲线下面积0.863(95%CI 0.862-0.864),优于逻辑回归0.845(95%CI 0.844-0.846)、ISS 0.828(95%CI 0.827-0.829)和TMPM-ICD10 0.861(95%CI 0.860-0.862)(均p<0.001)。
- 校准度:XGBoost的R²=0.993,逻辑回归R²=0.981,ISS R²=0.649,TMPM-ICD10 R²=0.830。
- 死亡患者较存活者更常见穿透性创伤且损伤数量更多,原文未给出具体数值。
证据类型、等级与局限
- 证据类型:回顾性数据库建模研究(预后与流行病学),证据等级level III。
- 局限:基于国家创伤数据库的回顾性设计,依赖ICD-10编码完整性,模型需外部验证;原文未提及其他局限。
对床旁实践的改变
- 该模型可改进创伤严重度量化与死亡预后判断,但基于ICD-10编码,需编码完成后才能计算,不适用于抢救室或ICU床旁实时决策。
- 不应将模型输出的死亡概率直接作为床旁决策阈值,原文未给出决策阈值且模型需前瞻性验证。
- 适用于创伤质量改进和创伤研究中的风险调整,不适用于个体患者的治疗限制决策。
- 常见陷阱:ISS需额外管理编码而ICD-10模型无需,但编码不完整或错误会直接影响模型表现。
指南或共识出处
- 原文未提及指南或共识名称;研究数据来源为2015至2017年国家创伤数据库(National Trauma Data Bank)。
以上提炼依据论文摘要(全文未开放获取,或许可不允许翻译转载);全文请点「阅读原文」。
英文原文(摘要)· 点击展开
BACKGROUND: Existing mortality prediction models have attempted to quantify injury burden following trauma-related admissions with the most notable being the Injury Severity Score (ISS). Although easy to calculate, it requires additional administrative coding. International Classification of Diseases (ICD)-based models such as the Trauma Mortality Prediction Model (TMPM-ICD10) circumvent these limitations, but they use linear modeling, which may not adequately capture the intricate relationships of injuries on mortality. Using ICD-10 coding and machine learning (ML) algorithms, the present study used the National Trauma Data Bank to develop mortality prediction models whose performance was compared with logistic regression, ISS, and TMPM-ICD10.
METHODS: The 2015 to 2017 National Trauma Data Bank was used to identify adults following trauma-related admissions. Of 8,021 ICD-10 codes, injuries were categorized into 1,495 unique variables. The primary outcome was in-hospital mortality. eXtreme Gradient Boosting (XGBoost), a ML technique that uses iterations of decision trees, was used to develop mortality models. Model discrimination was compared with logistic regression, ISS, and TMPM-ICD10 using receiver operating characteristic curve and probabilistic accuracy with calibration curves.
RESULTS: Of 1,611,063 patients, 54,870 (3.41%) experienced in-hospital mortality. Compared with those who survived, those who died more frequently suffered from penetrating trauma and had a greater number of injuries. The XGBoost model exhibited superior receiver operating characteristic curve (0.863 [95% confidence interval (CI), 0.862-0.864]) compared with logistic regression (0.845 [95% CI, 0.844-0.846]), ISS (0.828 [95% CI, 0.827-0.829]), and TMPM-ICD10 (0.861 [95% CI, 0.860-0.862]) (all p < 0.001). Importantly, the ML model also had significantly improved calibration compared with other methodologies (XGBoost, coefficient of determination (R2) = 0.993; logistic regression, R2 = 0.981; ISS, R2 = 0.649; TMPM-ICD10, R2 = 0.830).
CONCLUSION: Machine learning models using XGBoost demonstrated superior performance and calibration compared with logistic regression, ISS, and TMPM-ICD10. Such approaches in quantifying injury severity may improve its utility in mortality prognostication, quality improvement, and trauma research.
LEVEL OF EVIDENCE: Prognostic and Epidemiologic; level III.
参考来源
- Tran Z, Zhang W, Verma A, et al. The derivation of an International Classification of Diseases, Tenth Revision-based trauma-related mortality model using machine learning. J Trauma Acute Care Surg. 2022. PMID: 34554135. doi:10.1097/TA.0000000000003416
供执业医师抢救与诊疗参考,不替代临床判断;剂量请结合体重、肝肾功能和本院药品规格复核,以指南原文为准。