Learning Objectives

1. Explain why format specification is critical for clinical workflow reliability.

2. Apply at least four output formats to hospital documentation tasks.

3. Understand why unspecified format breaks downstream automation in EHR systems.

4. Select the right format for the right clinical audience.

Format Dimension

Options

Healthcare Example

Structure

Specify the prompt how would you like the output. e.g., 

• Table
• Numbered list
• JSON
• Paragraph

‘Return as a table, numbered list, JSON or some other format’

Length

Provide more details of the output e.g., 

• Word count
• Sentence count
• Max fields

‘Assessment section: 2 sentences maximum’

Required Fields

• Mandatory sections
• Headers
• Data elements

‘Include: Drug Name | Dose | Frequency | Purpose | Side Effects’

Excluded Content

What NOT to include in the output

‘Do not include patient name, DOB, or MRN’

Output Type

• Text
• Table
• JSON
• Structured data

‘Return as JSON for EHR import’

Maria Gonzalez (Patient Thread — Specify Format): Admission Documentation

Maria has been admitted. The Internal Medicine team needs three documents from this visit — each for a different audience and system.

Document 1 —  Numbered List (for the EHR and physician review):

Prompt: ‘Generate a numbered list for a 62-year-old female with uncontrolled T2DM (HbA1c 10.2%) and CKD Stage 2.

Format:

• S — Chief complaint and patient-reported symptoms
• O — Vital signs, lab values (BMP, HbA1c, eGFR)
• A — Primary diagnosis and differential
• P — Treatment plan, medications ordered, consults, follow-up.’

Document 2 — Medication Table (for the pharmacy and nursing team):

Prompt: ‘Generate a medication reconciliation table. Columns: Drug Name | Dose | Route | Frequency | Purpose | Key Side Effects.’

Output: Clean table — directly usable for nursing medication administration and pharmacy review.

Document 3 — Patient Education Paragraph (for Maria and her family):

Prompt: ‘Write a 100-word plain-language paragraph explaining Maria’s diagnosis and what will happen today. Use simple words. Avoid medical jargon. Write as if speaking directly to the patient.’