Skip to main content
Back to Library
Prompt Engineering Guide

Mastering Medical report summary
on Claude 3.5 Sonnet

Stop guessing. See how professional prompt engineering transforms Claude 3.5 Sonnet's output for specific technical tasks.

The "Vibe" Prompt

"Summarize this medical report for me. Make it easy to understand."
Low specificity, inconsistent output

Optimized Version

STABLE
You are a highly skilled medical summarization AI. Your task is to extract and synthesize key information from the provided medical report into a concise, easy-to-understand summary for a non-medical professional. Follow these steps: 1. **Identify Patient Demographics:** Extract the patient's name, age, gender (if available), and relevant identifiers. 2. **Highlight Chief Complaint:** What is the primary reason the patient sought medical attention? 3. **Summarize History of Present Illness (HPI):** Briefly describe the onset, duration, character, aggravating/alleviating factors, and associated symptoms. 4. **List Relevant Medical History:** Include significant past medical conditions, surgeries, and allergies. 5. **Note Key Physical Exam Findings:** Focus on abnormal or particularly relevant findings. 6. **Synthesize Diagnostic Test Results:** Summarize crucial lab results, imaging findings, or other diagnostic reports. 7. **Outline Assessment and Plan:** What is the diagnosis (differential or definitive) and the proposed treatment plan, including medications, follow-up, and lifestyle recommendations? 8. **Structure the Summary:** Present the information clearly using bullet points or short paragraphs under clear headings (e.g., 'Patient Information', 'Reason for Visit', 'Key Medical History', 'Findings', 'Diagnosis & Plan'). Ensure the language is accessible to a layperson, avoiding complex medical jargon where possible, or explaining it simply. Maintain accuracy and completeness of essential medical information. Focus on conciseness without sacrificing clarity. Medical Report: [INSERT MEDICAL REPORT HERE]
Structured, task-focused, reduced hallucinations

Engineering Rationale

The optimized prompt leverages several best practices for LLM interaction: 1. **Role-playing:** 'You are a highly skilled medical summarization AI.' sets a clear persona and expectation. 2. **Chain-of-thought (CoT):** The numbered steps break down the complex task into manageable sub-tasks, guiding the model's processing and ensuring comprehensive coverage. 3. **Structured Output Requirements:** Explicitly asking for bullet points/headings ('Structure the Summary') helps organize the output. 4. **Constraint-based Generation:** 'Ensure the language is accessible to a layperson,' 'avoiding complex medical jargon,' 'Maintain accuracy and completeness,' and 'Focus on conciseness' provide clear boundaries and quality expectations. 5. **Explicit Input Placeholder:** '[INSERT MEDICAL REPORT HERE]' makes it clear where the actual report should go. This structured approach drastically reduces ambiguity and provides the model with a clear roadmap for generating a high-quality summary, leading to more consistent and accurate results compared to the vague 'vibe_prompt'.

0%
Token Efficiency Gain
The optimized prompt explicitly asks for a summary for a 'non-medical professional'.
The optimized prompt uses numbered steps for a 'chain-of-thought' approach.
The optimized prompt specifies the desired output structure (e.g., 'bullet points or short paragraphs under clear headings').

Ready to stop burning tokens?

Join 5,000+ developers using Prompt Optimizer to slash costs and boost LLM reliability.

Optimize My Prompts