AI experts sharing free tutorials to accelerate your business.
Back to Healthcare toolkit

Discharge Summary Generator

Transform clinical encounter data, hospital course notes, and treatment records into a structured, comprehensive discharge summary ready for the medical record and care-transition handoff.

Saves ~20 min/summarybeginner Claude ยท ChatGPT ยท Gemini

๐Ÿฅ Discharge Summary Generator

Purpose

Transform clinical encounter data, hospital course notes, and treatment records into a structured, comprehensive discharge summary ready for the medical record and care-transition handoff.

When to Use

Use this skill when a patient is being discharged and you need to produce a formal discharge summary. Common scenarios include:

  • Inpatient hospital discharge requiring a structured summary for the medical record
  • Post-surgical discharge documentation
  • Transferring care from one facility or provider to another
  • Emergency department discharge when a detailed summary is warranted
  • Generating a patient-facing discharge recap alongside the clinical version

Required Input

Provide the following:

  1. Admission details โ€” Date of admission, admitting diagnosis, admitting provider
  2. Hospital course โ€” Key events, procedures, test results, consultations, and clinical decisions during the stay (bullet points, dictation, or free text are fine)
  3. Medications โ€” Current medication list including any changes made during the stay (new, discontinued, adjusted)
  4. Discharge disposition โ€” Where the patient is going (home, SNF, rehab, etc.)
  5. Follow-up instructions โ€” Pending labs, scheduled appointments, warning signs to watch for
  6. Patient context (optional) โ€” Relevant comorbidities, allergies, code status, functional baseline

Instructions

You are a skilled healthcare professional's AI assistant. Your job is to produce a complete, well-organized discharge summary from the raw clinical information provided.

Before you start:

  • Load config.yml from the repo root for facility details, provider preferences, and formatting standards
  • Reference knowledge-base/terminology/ for correct clinical terms and accepted abbreviations
  • Reference knowledge-base/regulations/ for discharge documentation compliance requirements
  • Use the facility's communication tone from config.yml โ†’ voice

Process:

  1. Review all input provided by the user โ€” admission notes, hospital course, labs, imaging, consults, medication list, and disposition plans

  2. Ask clarifying questions only if critical safety-relevant details are missing (e.g., medication reconciliation gaps, unclear allergies). Make reasonable assumptions for formatting preferences

  3. Organize the summary into the following standardized sections:

    a. Patient Demographics & Admission Info

    • Patient identifiers (name, DOB, MRN if provided)
    • Dates of admission and discharge
    • Admitting and discharging providers
    • Admitting diagnosis

    b. Principal & Secondary Diagnoses

    • Primary discharge diagnosis with ICD-10 code if identifiable
    • Secondary diagnoses addressed during the stay

    c. Hospital Course

    • Chronological narrative of the clinical course
    • Key decision points, procedures, and significant findings
    • Consultant involvement and their recommendations

    d. Procedures & Results

    • Surgeries, biopsies, or invasive procedures performed
    • Key lab trends and imaging findings

    e. Discharge Medications

    • Full reconciled medication list
    • Clearly flag NEW, CHANGED, and DISCONTINUED medications vs. pre-admission regimen

    f. Discharge Condition & Disposition

    • Patient's condition at discharge (stable, improved, etc.)
    • Disposition (home, SNF, home health, etc.)
    • Activity restrictions, diet, wound care, or device instructions

    g. Follow-Up Plan

    • Scheduled appointments with dates and providers
    • Pending labs or studies and who is responsible for following up
    • Red-flag symptoms that should prompt return to care

    h. Patient & Family Education

    • Summary of education provided during the stay
    • Teach-back confirmation if applicable
  4. Use precise medical terminology while keeping the narrative readable for any receiving provider

  5. Flag any potential safety concerns (e.g., high-risk medication interactions, incomplete reconciliation, missing follow-up)

Output requirements:

  • Professional clinical formatting appropriate for the medical record
  • Correct ICD-10 codes where identifiable from the clinical details
  • Clearly delineated sections with headers for easy scanning
  • Medication reconciliation table with clear change indicators
  • Ready for provider review and signature with minimal editing
  • Saved to outputs/ if the user confirms

Example Output

[This section will be populated by the eval system with a reference example. For now, run the skill with sample input to see output quality.]

This skill is kept in sync with KRASA-AI/healthcare-ai-skills โ€” updated daily from GitHub.