From Basic to Extra Features: Hypergraph Transformer Pretrain-then-Finetuning for Balanced Clinical Predictions on EHR
This work addresses the challenge of handling incomplete feature sets in EHR data for clinical prediction, which is incremental as it builds on existing pretrain-then-finetune approaches with specific enhancements.
The paper tackled the problem of deep learning models relying on massive features in Electronic Health Records (EHRs) that may not be available for all patients, proposing HTP-Star with a hypergraph transformer pretrain-then-finetune framework and techniques like Smoothness-inducing Regularization and Group-balanced Reweighting, resulting in consistent outperformance of baselines on two real EHR datasets while balancing patients with basic and extra features.
Electronic Health Records (EHRs) contain rich patient information and are crucial for clinical research and practice. In recent years, deep learning models have been applied to EHRs, but they often rely on massive features, which may not be readily available for all patients. We propose HTP-Star, which leverages hypergraph structures with a pretrain-then-finetune framework for modeling EHR data, enabling seamless integration of additional features. Additionally, we design two techniques, namely (1) Smoothness-inducing Regularization and (2) Group-balanced Reweighting, to enhance the model's robustness during fine-tuning. Through experiments conducted on two real EHR datasets, we demonstrate that HTP-Star consistently outperforms various baselines while striking a balance between patients with basic and extra features.