Directed Acyclic Graphs and causal thinking in clinical risk prediction modeling

Abstract

Background: In epidemiology, causal inference and prediction modeling methodologies have been historically distinct. Directed Acyclic Graphs (DAGs) are used to model a priori causal assumptions and inform variable selection strategies for causal questions. Although tools originally designed for prediction are finding applications in causal inference, the counterpart has remained largely unexplored. The aim of this theoretical and simulation-based study is to assess the potential benefit of using DAGs in clinical risk prediction modeling. Methods and Findings: We explore how incorporating knowledge about the underlying causal structure can provide insights about the transportability of diagnostic clinical risk prediction models to different settings. A single-predictor model in the causal direction is likely to have better transportability than one in the anticausal direction. We further probe whether causal knowledge can be used to improve predictor selection. We empirically show that the Markov Blanket, the set of variables including the parents, children, and parents of the children of the outcome node in a DAG, is the optimal set of predictors for that outcome. Conclusions: Our findings challenge the generally accepted notion that a change in the distribution of the predictors does not affect diagnostic clinical risk prediction model calibration if the predictors are properly included in the model. Furthermore, using DAGs to identify Markov Blanket variables may be a useful, efficient strategy to select predictors in clinical risk prediction models if strong knowledge of the underlying causal structure exists or can be learned.