Non-Invariance in Nested Prediction Models under Selective Predictor Availability

Abstract

Selective measurement of predictors is common in routinely collected health data. We used nested prediction models as a framework for characterising the consequences of a selectively measured predictor, with a restricted model defined in the target population and an extended model including the selectively measured predictor defined in a selected population. We show that non-invariance in the restricted model between selected patients and the target population decomposes into components due to omission of the additional predictor, potential residual non-invariance, and their interaction. We extended this framework to predictors measured through multiple routes of selection resulting in collider structures. We further show that imputation based on the conditional distribution of the additional predictor in the selected population transfers restricted-model non-invariance to the imputed extended model as imputation bias. We illustrate the framework using the Kidney Failure Risk Equation, where albumin-to-creatinine ratio is selectively measured in routine clinical practice. The framework provides a formal basis for understanding how selective predictor measurement may affect the development and validation of clinical prediction models using routinely collected health data.