longitudinal-ehr-qa

Construct visit-centric timelines from heterogeneous EHR data to answer longitudinal patient questions.

Updated Mar 24, 2026

Search Tags:#timeline #qa #provenance #harmonization #fhir #ehr #omop

Ready to run

npx skills add https://github.com/zengsihang/EHR-QA-Skill --skill longitudinal-ehr-qa

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: longitudinal-ehr-qa
Source: https://github.com/zengsihang/EHR-QA-Skill/tree/main
Command: npx skills add https://github.com/zengsihang/EHR-QA-Skill --skill longitudinal-ehr-qa

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires PyMuPDF>=1.23, and includes scripts (resource) and references (resource) components.

What problem does it solve?

The Longitudinal EHR QA skill enables analysts to answer patient-level longitudinal questions by constructing or reusing a harmonized, visit-centric timeline that preserves provenance and supports auditable decision-making.

Core Features & Use Cases

Builds or reuses harmonized timelines from heterogeneous sources (FHIR, OMOP, or mixed exports).
Produces per-question workspaces and artifacts (analysis plans, validation logs, and final reports) with explicit provenance.
Supports deterministic, code-assisted sub-tasks and reflection-driven validation to ensure trustworthiness.

Quick Start

Place the patient's data in the stable folder and run the detect, harmonization, and question-workflow steps to produce the final report and evidence table.

Frequently Asked Questions about longitudinal-ehr-qa

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I build a longitudinal EHR timeline from mixed FHIR and OMOP data?▼

To build a longitudinal EHR timeline, this Skill harmonizes heterogeneous FHIR, OMOP, and mixed CSV/JSON/XML/PDF exports into a visit-centric timeline. It processes your patient data to produce per-visit narratives and evidence-backed outputs.

Can I answer patient-level questions using a harmonized visit-centric timeline?▼

Yes, you can answer patient-level questions using a harmonized visit-centric timeline. The Skill creates question-specific workspaces with provenance-aware reasoning, ensuring agent-reviewed timing and deterministic sub-tasks for auditable decision-making.

What is the best way to perform EHR data harmonization and provenance tracking?▼

The best way to perform EHR harmonization and provenance tracking is using a Skill that constructs visit-centric timelines with evidence tables. It enforces agent-reviewed timing and produces analysis plans, validation logs, and JSON answers with explicit provenance.

Does this EHR timeline approach work with PDF patient records?▼

Yes, this EHR timeline approach works with PDF patient records by using the PyMuPDF dependency. It detects and harmonizes mixed CSV, JSON, XML, and PDF inputs into a unified timeline for longitudinal question answering.

What outputs do I get from patient-level longitudinal EHR question answering?▼

You get final deliverables including reports, JSON answers, and evidence tables from patient-level longitudinal EHR question answering. The Skill generates per-question workspaces containing analysis plans, validation logs, and final reports with explicit provenance.

How do I validate timing and events in an OMOP or FHIR timeline?▼

You validate timing and events in an OMOP or FHIR timeline through reflection-driven validation and deterministic Python sub-tasks. The Skill enforces agent-reviewed timing to ensure trustworthiness and produces validation logs within per-question workspaces.