[Paper Review] EHRSummarizer: A Privacy-Aware, FHIR-Native Architecture for Structured Clinical Summarization of Electronic Health Records

Houman Kazemzadeh, Nima Minaifar+2 more|arXiv (Cornell University)|Jan 4, 2026

Electronic Health Records Systems0 citations

TL;DR

EHRSummarizer proposes a privacy-aware, FHIR-native architecture to produce source-grounded structured EHR summaries for chart review, with prototype demonstrations but no clinical outcomes reported.

ABSTRACT

Clinicians routinely navigate fragmented electronic health record (EHR) interfaces to assemble a coherent picture of a patient's problems, medications, recent encounters, and longitudinal trends. This work describes EHRSummarizer, a privacy-aware, FHIR-native reference architecture that retrieves a targeted set of high-yield FHIR R4 resources, normalizes them into a consistent clinical context package, and produces structured summaries intended to support structured chart review. The system can be configured for data minimization, stateless processing, and flexible deployment, including local inference within an organization's trust boundary. To mitigate the risk of unsupported or unsafe behavior, the summarization stage is constrained to evidence present in the retrieved context package, is intended to indicate missing or unavailable domains where feasible, and avoids diagnostic or treatment recommendations. Prototype demonstrations on synthetic and test FHIR environments illustrate end-to-end behavior and output formats; however, this manuscript does not report clinical outcomes or controlled workflow studies. We outline an evaluation plan centered on faithfulness, omission risk, temporal correctness, usability, and operational monitoring to guide future institutional assessments.

Motivation & Objective

• Motivate the need for coherent EHR summaries to aid clinicians amid fragmented EHR interfaces.
• Propose a privacy-aware, FHIR-native reference architecture for structured EHR summarization.
• Describe how to retrieve, normalize, and constrain summaries using high-yield FHIR resources.
• Address missing-data handling, medication-status ambiguity, and controlled use of narrative documents.
• Outline future source-to-summary traceability and evaluation plans.

Proposed method

• Retrieve a targeted set of high-yield HL7 FHIR R4 resources.
• Normalize retrieved data into a clinical context package.
• Apply a constrained summarization stage to produce source-grounded summaries.
• Clarify handling of missing-data status and medication-status ambiguity.
• Control use of narrative clinical documents when available.
• Outline source-to-summary traceability for future evaluation.

Experimental results

Research questions

• RQ1 How faithful are the generated summaries to the underlying source data?
• RQ2 How does the system manage missing data and ambiguity in medication status?
• RQ3 Can the summaries maintain temporal correctness and support chart review without including extraneous narratives?
• RQ4 What are the privacy, usability, and operational considerations for real-world institutional deployment?
• RQ5 What plans exist for source-to-summary traceability and longitudinal evaluation?

Key findings

• Prototype demonstrations on synthetic and test FHIR environments illustrate end-to-end behavior and output formats.
• The architecture focuses on privacy-aware, source-grounded summaries rather than clinical outcomes or decision support.
• The manuscript clarifies missing-data status handling, medication-status ambiguity, and controlled narrative-document handling in its v2 revision.
• The study explicitly notes that it does not report clinical outcomes, controlled workflow studies, or benchmark results.
• An evaluation plan is outlined to assess faithfulness, omission risk, temporal correctness, usability, privacy, and operational monitoring.