Clinical outcome prediction under hypothetical interventions -- a representation learning framework for counterfactual reasoning
This addresses the need for causal interpretation in clinical decision-making, allowing researchers and clinicians to predict outcomes under hypothetical interventions for personalized care, though it is incremental as it builds on existing representation learning methods.
The paper tackles the problem of enabling counterfactual reasoning in clinical outcome prediction by introducing a partial concept bottleneck framework that jointly optimizes for prediction accuracy and counterfactual explanations, with results showing comparable accuracy to prediction-only models.
Most machine learning (ML) models are developed for prediction only; offering no option for causal interpretation of their predictions or parameters/properties. This can hamper the health systems' ability to employ ML models in clinical decision-making processes, where the need and desire for predicting outcomes under hypothetical investigations (i.e., counterfactual reasoning/explanation) is high. In this research, we introduce a new representation learning framework (i.e., partial concept bottleneck), which considers the provision of counterfactual explanations as an embedded property of the risk model. Despite architectural changes necessary for jointly optimising for prediction accuracy and counterfactual reasoning, the accuracy of our approach is comparable to prediction-only models. Our results suggest that our proposed framework has the potential to help researchers and clinicians improve personalised care (e.g., by investigating the hypothetical differential effects of interventions)