Improving Patient Subtyping on Longitudinal Data using Representations from Mamba-based Architecture

2026-06-26

AI Summary

Electronic health record (EHR) data exhibit irregular temporal sampling and sparsity, posing significant challenges for patient subtyping. This work proposes a novel self-supervised temporal modeling approach that, for the first time, leverages the Mamba state space model for EHR representation learning. The method effectively captures long-range dependencies while robustly handling missing data and asynchronous sampling inherent in real-world EHRs. Using the learned high-quality representations combined with multiple clustering algorithms, the proposed framework consistently outperforms existing baselines across several real-world EHR datasets. It not only achieves superior accuracy in unsupervised patient subtyping but also demonstrates enhanced performance on downstream predictive tasks, highlighting the clinical utility of the learned representations.

Abstract

Effective sub-typing (also known as grouping or clustering) of patients using their electronic health record (EHR) data can greatly inform precision medicine efforts. However, subtyping temporal EHR datasets is known to be challenging due to inherent EHR issues, including complexity and irregularity. In this study, we propose a self-supervised Mamba-based model that learns effective EHR representations and enables enhanced patient subtyping. We evaluate the proposed model on public and private real-world EHR datasets to classify the data based on the available labels and subtype patients based on the representations learned from the model. Through an extensive set of experiments, we demonstrate that our model's design choices lead to better performance compared to competitive baseline models for prediction. Moreover, we evaluate several clustering techniques to demonstrate that our findings offer valuable insights into subtyping patients based on temporal records from EHR models\footnote{Our implementations are available at https://github.com/healthylaife/triplet_mamba.

Problem

• patient subtyping
• longitudinal data
• electronic health records
• temporal EHR
• precision medicine

Innovation

• Mamba-based architecture
• self-supervised learning
• patient subtyping
• electronic health records (EHR)
• longitudinal data