X-FEMR: A Token-level Explainable Approach for Electronic Health Records Foundation Models using Transformer-based Models

Huang, Jie; Yin, Pengfei; Xu, Zihan; Capurro, Daniel; Conway, Mike; Dang, Ting

doi:10.24963/ijcai.2026/749

X-FEMR: A Token-level Explainable Approach for Electronic Health Records Foundation Models using Transformer-based Models

Jie Huang, Pengfei Yin, Zihan Xu, Daniel Capurro, Mike Conway, Ting Dang

Download BibTeX Cite

Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence. AI and Health. Pages 6734-6742.

https://doi.org/10.24963/ijcai.2026/749

Keywords

Health data mining: Health data mining
EHR analysis: EHR analysis
Multimodal data: Multimodal data
Explainable AI: Explainable AI
Clinical decision support system: Clinical decision support system

Abstract

Foundation Models for Electronic Health Records (FEMRs) are pretrained on large-scale structured patient data, enabling them to convert longitudinal patient trajectories into generalizable representations for diverse clinical prediction tasks. Despite their effectiveness, FEMRs remain black-box models, raising concerns about bias, interpretability, and clinical trust. To address this, we propose the first token-level explainability approach for FEMRs. We train a Transformer-based surrogate model on input-output pairs from the FEMR across two prediction tasks, approximating its behavior while preserving temporal dynamics. We identify the most influential tokens, providing insights into how FEMRs leverage different aspects of patient history for predictions. To evaluate clinical relevance, we introduce a novel clinical alignment metric that quantifies the correspondence between the surrogate model’s key tokens and clinically validated features. Our results demonstrate that the surrogate closely approximates FEMR predictions and that token-level explanations align well with clinical knowledge, offering a practical framework for interpretable and trustworthy clinical AI.

Jie Huang, Pengfei Yin, Zihan Xu, Daniel Capurro, Mike Conway, and Ting Dang. 2026. X-FEMR: A Token-level Explainable Approach for Electronic Health Records Foundation Models using Transformer-based Models. In Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence, IJCAI-26. IJCAI, 6734–6742. https://doi.org/10.24963/ijcai.2026/749

@inproceedings{ijcai2026p749,
  title     = {X-FEMR: A Token-level Explainable Approach for Electronic Health Records Foundation Models using Transformer-based Models},
  author    = {Huang, Jie and Yin, Pengfei and Xu, Zihan and Capurro, Daniel and Conway, Mike and Dang, Ting},
  booktitle = {Proceedings of the Thirty-Fifth International Joint Conference on
               Artificial Intelligence, {IJCAI-26}},
  publisher = {International Joint Conferences on Artificial Intelligence Organization},
  editor    = {Diego Calvanese},
  pages     = {6734--6742},
  year      = {2026},
  month     = {8},
  note      = {AI and Health},
  doi       = {10.24963/ijcai.2026/749},
  url       = {https://doi.org/10.24963/ijcai.2026/749},
}

TY  - CPAPER
TI  - X-FEMR: A Token-level Explainable Approach for Electronic Health Records Foundation Models using Transformer-based Models
AU  - Huang, Jie
AU  - Yin, Pengfei
AU  - Xu, Zihan
AU  - Capurro, Daniel
AU  - Conway, Mike
AU  - Dang, Ting
T2  - Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence, IJCAI-26
PB  - International Joint Conferences on Artificial Intelligence Organization
PY  - 2026
SP  - 6734
EP  - 6742
DO  - 10.24963/ijcai.2026/749
UR  - https://doi.org/10.24963/ijcai.2026/749
ER  -