Hospital discharge summaries generated with linked clinical evidence
Toward Complete Hospital Discharge Summarization with Abstract Meaning Representation
Computation and Language
Summary
Writing hospital discharge summaries can take doctors a lot of time, so the authors developed a method to create these summaries automatically. Their approach carefully links each sentence in the summary to exact parts of the original medical notes to avoid mistakes. They tested this on public and local hospital data and shared their code and trained models for others to use. This helps make sure summaries are accurate and trustworthy.
What this means in practice
- •For hospital data teams: Generate hospital discharge summaries with verified links to original notes to reduce clinician documentation time safely.
- •For healthcare software developers: Build clinical documentation tools that produce evidence-backed summaries minimizing errors caused by AI hallucinations.
Authors
Paul Landes, Sitara Rao, Aaron Jeremy Chaise, Barbara Di Eugenio
Abstract
Discharge summaries are lengthy medical documents that summarize a hospital in-patient visit. Automatically generating them can reduce documentation burden and return clinician time to patient care. Whereas Large Language Model (LLMs) could be used for this task, their Achilles heel is hallucinations, which can have drastic consequences for clinical documentation. We present an evidence-driven alignment framework for discharge summarization at the clinical encounter level, that treats provenance as a first-class constraint, using semantic graphs and deep learning models. Each summary sentence is selected and organized via cross-document semantic alignment and is accompanied by explicit evidence links to its source spans. We show our results on two corpora: a publicly available corpus (MIMIC-III) and clinical notes written by physicians at the University of Illinois Hospital (UIC Health). Additionally, we make source code and trained models available.