Repository logo
 

Adaptive Prediction Timing for Electronic Health Records

Accepted version
Peer-reviewed

Type

Conference Object

Change log

Authors

Deasy, Jacob 
Liò, Pietro 

Abstract

In realistic scenarios, multivariate timeseries evolve over case-by-case time-scales. This is particularly clear in medicine, where the rate of clinical events varies by ward, patient, and application. Increasingly complex models have been shown to effectively predict patient outcomes, but have failed to adapt granularity to these inherent temporal resolutions. As such, we introduce a novel, more realistic, approach to generating patient outcome predictions at an adaptive rate based on uncertainty accumulation in Bayesian recurrent models. We use a Recurrent Neural Network (RNN) and a Bayesian embedding layer with a new aggregation method to demonstrate adaptive prediction timing. Our model predicts more frequently when events are dense or the model is certain of event latent representations, and less frequently when readings are sparse or the model is uncertain. At 48 hours after patient admission, our model achieves equal performance compared to its static-windowed counterparts, while generating patient- and event-specific prediction timings that lead to improved predictive performance over the crucial first 12 hours of the patient stay.

Description

Keywords

cs.LG, cs.LG, stat.ML

Journal Title

CoRR

Conference Name

International Conference on Learning Representations 2020

Journal ISSN

Volume Title

Publisher

Rights

All rights reserved