A machine learning framework for interpretable predictions in patient pathways: The case of predicting ICU admission for patients with symptoms of sepsis

TL;DR

This paper introduces PatWay-Net, an interpretable neural framework for predicting ICU admission from patient pathways in sepsis cases. It blends a static module of per-feature MLPs with an interpretable iLSTM for sequential data, connected through a simple fusion layer, and is augmented by a clinical dashboard that visualizes health trajectories and feature effects. Compared with interpretable baselines and several black-box models, PatWay-Net achieves competitive predictive performance while providing intrinsic interpretability via four explanation plots, and its clinical utility is validated through structured clinician interviews. The work demonstrates a practical path to reliable, transparent decision support in critical care, with potential to improve resource management and patient outcomes while avoiding opaque “black-box” reasoning.

Abstract

Proactive analysis of patient pathways helps healthcare providers anticipate treatment-related risks, identify outcomes, and allocate resources. Machine learning (ML) can leverage a patient's complete health history to make informed decisions about future events. However, previous work has mostly relied on so-called black-box models, which are unintelligible to humans, making it difficult for clinicians to apply such models. Our work introduces PatWay-Net, an ML framework designed for interpretable predictions of admission to the intensive care unit (ICU) for patients with symptoms of sepsis. We propose a novel type of recurrent neural network and combine it with multi-layer perceptrons to process the patient pathways and produce predictive yet interpretable results. We demonstrate its utility through a comprehensive dashboard that visualizes patient health trajectories, predictive outcomes, and associated risks. Our evaluation includes both predictive performance - where PatWay-Net outperforms standard models such as decision trees, random forests, and gradient-boosted decision trees - and clinical utility, validated through structured interviews with clinicians. By providing improved predictive accuracy along with interpretable and actionable insights, PatWay-Net serves as a valuable tool for healthcare decision support in the critical case of patients with symptoms of sepsis.

Figures (18)

• Figure 1: Illustration of underlying setting. Multiple tasks must be performed when a patient is transferred to a new department or receives a new treatment. Thus, early prediction of the various steps a patient goes through during their hospital stay leads to more efficient operations.
• Figure 2: Illustration of the architecture of PatWay-Net's dnn model consisting of a sequential, a static, and a connection module. Here, two static and two sequential features are shown, which run through their modules and are then connected.
• Figure 3: Medical dashboard with PatWay-Net's interpretation plots.
• Figure 4: Importance for static and sequential medical indicators.
• Figure 5: Interpretation plot for static medical indicator Age in the dashboard.

Theorems & Definitions (2)

• Definition 1: Patient Pathways
• Definition 2: Patient Pathway Prefix