标题： 临床实践中缺失预测值的实时插补 显示英文标题

标题： Real-time imputation of missing predictor values in clinical practice

摘要： 预测模型的使用在临床指南中被广泛推荐，但通常需要所有预测因子的完整信息，而在日常实践中这些信息并不总是可获得的。 我们描述了在实际应用预测模型时，对缺失预测值进行实时处理的两种方法。 我们将广泛使用的均值插补方法（M-imp）与一种利用观察到的患者特征来个性化插补的方法进行了比较。 这些特征可能包括预测模型变量和其他特征（辅助变量）。 该方法通过从患者特征的联合多元正态模型进行插补来实现（联合建模插补；JMI）。 使用两个不同的心血管队列数据，其中包含心血管预测因子和结果，以评估实时插补方法。 我们量化了预测模型的整体性能（线性预测器的均方误差（MSE））、区分度（c指数）、校准（截距和斜率）以及净收益（决策曲线分析）。 与均值插补相比，JMI显著改善了MSE（0.10 vs. 0.13）、c指数（0.70 vs 0.68）和校准（整体校准：0.04 vs. 0.06；校准斜率：1.01 vs. 0.92），尤其是在纳入辅助变量时。 当插补方法基于外部队列时，校准有所下降，但区分度保持相似。 我们建议在实时插补缺失值时使用包含辅助变量的JMI，并在在新环境或（亚）人群中实施时更新插补模型。 显示英文摘要

摘要： Use of prediction models is widely recommended by clinical guidelines, but usually requires complete information on all predictors that is not always available in daily practice. We describe two methods for real-time handling of missing predictor values when using prediction models in practice. We compare the widely used method of mean imputation (M-imp) to a method that personalizes the imputations by taking advantage of the observed patient characteristics. These characteristics may include both prediction model variables and other characteristics (auxiliary variables). The method was implemented using imputation from a joint multivariate normal model of the patient characteristics (joint modeling imputation; JMI). Data from two different cardiovascular cohorts with cardiovascular predictors and outcome were used to evaluate the real-time imputation methods. We quantified the prediction model's overall performance (mean squared error (MSE) of linear predictor), discrimination (c-index), calibration (intercept and slope) and net benefit (decision curve analysis). When compared with mean imputation, JMI substantially improved the MSE (0.10 vs. 0.13), c-index (0.70 vs 0.68) and calibration (calibration-in-the-large: 0.04 vs. 0.06; calibration slope: 1.01 vs. 0.92), especially when incorporating auxiliary variables. When the imputation method was based on an external cohort, calibration deteriorated, but discrimination remained similar. We recommend JMI with auxiliary variables for real-time imputation of missing values, and to update imputation models when implementing them in new settings or (sub)populations.