Search

Documentation burden is a leading contributor to physician burnout, yet detection often depends on periodic self-report surveys. Electronic health record audit logs provide an objective and continuous record of clinical work patterns that may reveal burden before physicians formally report distress. Moving from reactive survey assessment to proactive detection requires transforming complex, high-dimensional audit log signals into meaningful burden classifications. These signals must be modeled in a way that reflects workload intensity, temporal accumulation, and the interaction of documentation, inbox, and order-entry demands. This article proposes a conceptual deep neural network model for detecting physicians with high documentation burden. The model uses note length, time in the electronic health record, after-hours charting activity, inbox volume, and order entry patterns as core input domains. The proposed model uses a multi-input neural architecture that fuses aggregated and temporally aware features derived from raw electronic health record audit logs. The model would generate a burden probability score for each physician over a defined weekly period without requiring direct linkage to individual patient content. Conceptually, the model could identify physicians with rising documentation burden earlier than survey-based approaches. It would also be expected to reveal the dominant burden component, such as excessive inbox work, prolonged after-hours charting, unusually long notes, or high-complexity order entry. A deep learning model for documentation burden detection could help health systems move from burnout treatment to prevention. By connecting objective workload signals to targeted operational interventions, such a model could support physician well-being while preserving privacy and professional trust.

Foundation models pre-trained on massive, multimodal data have transformed several areas of artificial intelligence. Their adaptation to healthcare operations analytics is an emerging frontier because hospital workflows generate dense streams of structured, textual, temporal, and administrative data. This systematic review examined applications of foundation models to patient flow prediction, documentation burden estimation, resource allocation, and operational risk forecasting from 2017 to 2026. The review focused on multimodal pretraining, transfer learning, downstream adaptation, validation, and implementation maturity. A PRISMA 2020-compliant search was designed for PubMed, Scopus, IEEE Xplore, and Web of Science. Records were screened by two reviewers, and eligible studies were synthesized narratively by operational domain, model architecture, data modality, and maturity level. A small but rapidly growing body of work suggests that pre-trained multimodal models may support operational prediction tasks across healthcare systems. Evidence was concentrated in patient flow and resource allocation, while documentation burden estimation and operational risk forecasting were less frequently studied. Foundation models show potential to unify operational analytics across hospital systems. The field remains immature, with limited external validation, few prospective implementation studies, and no widely adopted benchmarks for operational foundation models.