Abstract

Heart failure (HF) is one of the leading causes of hospital admissions in the US. Readmission within 30 days after a HF hospitalization is both a recognized indicator for disease progression and a source of considerable financial burden to the healthcare system. Consequently, the identification of patients at risk for readmission is a key step in improving disease management and patient outcome. In this work, we used a large administrative claims dataset to (1) explore the systematic application of neural network-based models versus logistic regression for predicting 30 days all-cause readmission after discharge from a HF admission, and (2) to examine the additive value of patients’ hospitalization timelines on prediction performance. Based on data from 272,778 (49% female) patients with a mean (SD) age of 73 years (14) and 343,328 HF admissions (67% of total admissions), we trained and tested our predictive readmission models following a stratified 5-fold cross-validation scheme. Among the deep learning approaches, a recurrent neural network (RNN) combined with conditional random fields (CRF) model (RNNCRF) achieved the best performance in readmission prediction with 0.642 AUC (95% CI, 0.640–0.645). Other models, such as those based on RNN, convolutional neural networks and CRF alone had lower performance, with a non-timeline based model (MLP) performing worst. A competitive model based on logistic regression with LASSO achieved a performance of 0.643 AUC (95% CI, 0.640–0.646). We conclude that data from patient timelines improve 30 day readmission prediction, that a logistic regression with LASSO has equal performance to the best neural network model and that the use of administrative data result in competitive performance compared to published approaches based on richer clinical datasets.

Highlights

  • Heart failure (HF) is one of the leading causes for hospital admissions in the US1–4 with high numbers of readmissions within 30 days of discharge[2,3,4]

  • Starting from recurrent neural networks (RNN), the models trained with loss functions incorporating/emphasizing the loss from last HF event (i.e LastHF and Convex_HF_LastHF) achieved higher performance 0.636 and 0.635 area under the ROC curve (AUC) respectively compared to other loss function definitions

  • For the best neural model (RNNCRF), we report the analysis of feature importance using a similar approach to the one in[12]

Read more

Summary

Introduction

Heart failure (HF) is one of the leading causes for hospital admissions in the US1–4 with high numbers of readmissions within 30 days of discharge[2,3,4]. One specific aim of this study is to examine the value of including a patient’s trajectory data in a 30 day readmission prediction model To this end, we examine three approaches for modeling the problem of which two use the temporal information encoded in the patients’ trajectories (sequence labeling and sequence classification), and one that does not (index event classification). We implemented multiple neural network models with varying architectures and objective functions such as recurrent neural networks (RNN), and convolutional neural networks (CNN) as examples of sequence labeling and classification approaches, and multilayer perceptron (MLP) along with logistic regression as baseline models representing the index event classification approach We conducted these studies with a large administrative claims dataset, which lacks the detailed clinical information found in datasets typically used for this problem. As claims data are readily available and can be robustly harmonized, they pose less privacy concerns and are ideally suited for tacking the HF readmission problem

Methods
Results
Conclusion

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

Disclaimer: 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.