Abstract
BackgroundMachine learning (ML) holds the promise of becoming an essential tool for utilising the increasing amount of clinical data available for analysis and clinical decision support. However, the lack of trust in the models has limited the acceptance of this technology in healthcare. This mistrust is often credited to the shortage of model explainability and interpretability, where the relationship between the input and output of the models is unclear. Improving trust requires the development of more transparent ML methods.MethodsIn this paper, we use the publicly available eICU database to construct a number of ML models before examining their internal behaviour with SHapley Additive exPlanations (SHAP) values. Our four models predicted hospital mortality in ICU patients using a selection of the same features used to calculate the APACHE IV score and were based on random forest, logistic regression, naive Bayes, and adaptive boosting algorithms.ResultsThe results showed the models had similar discriminative abilities and mostly agreed on feature importance while calibration and impact of individual features differed considerably and did in multiple cases not correspond to common medical theory.ConclusionsWe already know that ML models treat data differently depending on the underlying algorithm. Our comparative analysis visualises implications of these differences and their importance in a healthcare setting. SHAP value analysis is a promising method for incorporating explainability in model development and usage and might yield better and more trustworthy ML models in the future.
Highlights
Machine learning (ML) holds the promise of becoming an essential tool for utilising the increasing amount of clinical data available for analysis and clinical decision support
The importance of different features on a prediction can for some ML models be explained using coefficients, or by tracing or visualising the steps taken by the algorithm
This study shows differences between what the ML models consider important factors for mortality compared to medical theory
Summary
Machine learning (ML) holds the promise of becoming an essential tool for utilising the increasing amount of clinical data available for analysis and clinical decision support. With the increasing availability and use of digital aid in health care, such as sensors and electronic health records, patients generate large amounts of data that can be used in treatment and analysis Some of this information is not necessarily informative on its own but can give insight into complex medical problems when combined. The importance of different features on a prediction can for some ML models be explained using coefficients, or by tracing or visualising the steps taken by the algorithm. Still, this is not always sufficient for obtaining a model that is understood by humans for development and use. The need for explainable ML and models that can be understood by humans is becoming increasingly apparent [4, 5]
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call