Abstract
The ability to perform accurate prognosis is crucial for proactive clinical decision making, informed resource management and personalised care. Existing outcome prediction models suffer from a low recall of infrequent positive outcomes. We present a highly-scalable and robust machine learning framework to automatically predict adversity represented by mortality and ICU admission and readmission from time-series of vital signs and laboratory results obtained within the first 24 hours of hospital admission. The stacked ensemble platform comprises two components: a) an unsupervised LSTM Autoencoder that learns an optimal representation of the time-series, using it to differentiate the less frequent patterns which conclude with an adverse event from the majority patterns that do not, and b) a gradient boosting model, which relies on the constructed representation to refine prediction by incorporating static features. The model is used to assess a patient's risk of adversity and provides visual justifications of its prediction. Results of three case studies show that the model outperforms existing platforms in ICU and general ward settings, achieving average Precision-Recall Areas Under the Curve (PR-AUCs) of 0.891 (95% CI: 0.878-0.939) for mortality and 0.908 (95% CI: 0.870-0.935) in predicting ICU admission and readmission.
Highlights
The secondary re-use of routinely collected patient data has been a facilitator of innovations aiming to improve patient care
Despite the promising results of Machine Learning early warning systems, we find that existing approaches bear several shortcomings that adversely affect model performance and adoption potential
This paper presents KD-OP (Knowledge Distillation Outcome Predictor), an ensemble Machine Learning framework designed to overcome the current difficulties in predicting adverse clinical outcomes from electronic health records data
Summary
The secondary re-use of routinely collected patient data has been a facilitator of innovations aiming to improve patient care. A prominent example is the development of early warning systems that predict adversity from patient physiological measurements. The majority of early warning models take the form of ad-hoc scoring tools [40] such as the National Early Warning Score (NEWS2) widely used in the United Kingdom [54]. Such tools estimate a patient’s risk of adversity using aggregates of physiological measurements [27]. Machine Learning models have been developed to overcome the limitations of scoring tools via sophisticated architectures that capture non-linearities within the multivariate temporal patient data [12]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
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 callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
