Real-time imputation of missing predictor values in clinical practice

TL;DR

This paper introduces a joint modeling imputation method for real-time handling of missing predictor data in clinical prediction models, demonstrating improved accuracy over mean imputation in cardiovascular cohorts.

Contribution

It presents a novel joint multivariate normal imputation approach that personalizes missing data handling using auxiliary variables, enhancing prediction performance in clinical practice.

Findings

JMI improves mean squared error and discrimination over mean imputation.

Incorporating auxiliary variables enhances imputation accuracy.

Calibration deteriorates when applying external cohort models, but discrimination remains stable.

Abstract

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.