A Narrative-Driven Computational Framework for Clinician Burnout Surveillance

2025-09-01

AI Summary

Current monitoring of clinician burnout in intensive care units (ICUs) relies on retrospective self-report questionnaires or coarse-grained electronic health record (EHR) metadata, overlooking rich, unstructured clinical narratives. Method: We propose the first narrative-driven computational framework for burnout detection from discharge summaries, integrating clinically fine-tuned BioBERT sentiment embeddings, a burnout-specific lexicon, and LDA topic modeling to extract psychological stress signals from free-text notes; we further design a multimodal analytical pipeline combining semantic representations, topic features, workload proxy variables, and logistic regression. Results: Evaluated on MIMIC-IV, our approach achieves an F1-score of 0.84—significantly outperforming baselines—and reveals, for the first time, elevated burnout risk among radiologists, psychiatrists, and neurologists. This work establishes a novel paradigm for real-time, proactive occupational health surveillance in clinical settings.

Technology Category

• Natural Language Processing: Fact-Checking / Misinformation Detection (NLP Focus)
• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Web Mining and Content Analysis: Large pretrained models with web data
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Clinician burnout poses a substantial threat to patient safety, particularly in high-acuity intensive care units (ICUs). Existing research predominantly relies on retrospective survey tools or broad electronic health record (EHR) metadata, often overlooking the valuable narrative information embedded in clinical notes. In this study, we analyze 10,000 ICU discharge summaries from MIMIC-IV, a publicly available database derived from the electronic health records of Beth Israel Deaconess Medical Center. The dataset encompasses diverse patient data, including vital signs, medical orders, diagnoses, procedures, treatments, and deidentified free-text clinical notes. We introduce a hybrid pipeline that combines BioBERT sentiment embeddings fine-tuned for clinical narratives, a lexical stress lexicon tailored for clinician burnout surveillance, and five-topic latent Dirichlet allocation (LDA) with workload proxies. A provider-level logistic regression classifier achieves a precision of 0.80, a recall of 0.89, and an F1 score of 0.84 on a stratified hold-out set, surpassing metadata-only baselines by greater than or equal to 0.17 F1 score. Specialty-specific analysis indicates elevated burnout risk among providers in Radiology, Psychiatry, and Neurology. Our findings demonstrate that ICU clinical narratives contain actionable signals for proactive well-being monitoring.

Problem

• Detecting clinician burnout using narrative clinical notes
• Overcoming limitations of retrospective surveys and EHR metadata
• Proactive monitoring through computational analysis of ICU narratives

Innovation

• Hybrid pipeline combining BioBERT embeddings and LDA
• Lexical stress lexicon tailored for burnout surveillance
• Provider-level logistic regression classifier achieving high precision