Key result
Explainable Boosting Machine model outperforms SAPS II for 3-day ICU readmission prediction with 0.119 PR-AUC.
Why the study?
Intensive care unit readmissions are linked to adverse outcomes, and although machine learning could aid discharge decisions, many models are black boxes with potentially dangerous properties.
Does an interpretable Explainable Boosting Machine model accurately predict 3-day ICU readmission or death in adult ICU patients compared to standard scoring systems and black-box machine learning models?
Cohort (n=15,589)
No
Does an interpretable Explainable Boosting Machine model accurately predict 3-day ICU readmission or death in adult ICU patients compared to standard scoring systems and black-box machine learning models?
Absolute Event Rate: 0.119% vs 0.084%
An inherently interpretable machine learning model using Explainable Boosting Machines can predict 3-day ICU readmissions with performance comparable to black-box models, allowing for clinical inspection and removal of problematic features.
Hypothesis-generating for interpretable ML in ICU readmission; prospective validation required before clinical use.
Intensive care unit readmissions are associated with mortality and bad outcomes. Machine learning could help to identify patients at risk to improve discharge decisions. However, many models are black boxes, so that dangerous properties might remain unnoticed. In this study, an inherently interpretable model for 3-day ICU readmission prediction was developed. We used a retrospective cohort of 15,589 ICU stays and 169 variables collected between 2006 and 2019. A team of doctors inspected the model, checked the plausibility of each component, and removed problematic parts. Qualitative feedback revealed several challenges for interpretable machine learning in healthcare. The resulting model used 67 features and showed an area under the precision-recall curve of 0.119±0.020 and an area under the receiver operating characteristic curve of 0.680±0.025. This is on par with state-of-the-art gradient boosting machines and outperforms the Simplified Acute Physiology Score II. External validation with the Medical Information Mart for Intensive Care database version IV confirmed our findings. Hence, a machine learning model for readmission prediction with a high level of human control is feasible without sacrificing performance.
No takes yet. Share an insight, caveat, or question.
Hegselmann et al. (2021) conducted a cohort in ICU patients at risk of 3-day readmission (n=15,589). Explainable Boosting Machines (EBM) model vs. Simplified Acute Physiology Score II (SAPS II) was evaluated on Area under the precision-recall curve (PR-AUC) for 3-day ICU readmission. An interpretable Explainable Boosting Machine model for 3-day ICU readmission prediction achieved a PR-AUC of 0.119, outperforming the Simplified Acute Physiology Score II.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: