Skip to main content
Back to Library
Prompt Engineering Guide

Mastering Medical report summary
on Qwen 2.5 72B

Stop guessing. See how professional prompt engineering transforms Qwen 2.5 72B's output for specific technical tasks.

The "Vibe" Prompt

"Summarize this medical report: [MEDICAL REPORT TEXT]"
Low specificity, inconsistent output

Optimized Version

STABLE
You are an AI assistant specialized in medical summarization. Your goal is to provide a concise, accurate, and clinically relevant summary of the following medical report. Focus on identifying key patient information, primary diagnosis, significant findings, treatment plan, and follow-up recommendations. Extract only the most pertinent information. Do not include extraneous details or interpret findings beyond what is explicitly stated. Here's the medical report: [MEDICAL REPORT TEXT] Follow this chain of thought to generate the summary: 1. **Identify Patient Demographics:** Extract patient's age, gender, and relevant medical record identifiers if present. 2. **Identify Chief Complaint/Reason for Visit:** Determine the primary reason the patient sought medical attention. 3. **Extract Pertinent Medical History:** Note any significant past medical history, surgical history, allergies, or medications that are directly relevant to the current presentation. 4. **Summarize Key Physical Examination Findings:** Focus on abnormal or noteworthy findings. 5. **Summarize Diagnostic Test Results:** Include results from labs, imaging (e.g., X-ray, CT, MRI), or other tests that contribute to the diagnosis. 6. **State Primary Diagnosis(es):** Clearly list the main medical diagnosis or diagnoses. 7. **Outline Treatment Plan:** Detail interventions, medications prescribed, procedures performed, or referrals made. 8. **List Follow-up Recommendations:** Include instructions for patient care after the visit, such as follow-up appointments, lifestyle advice, or monitoring. 9. **Synthesize into a Concise Summary:** Combine the extracted information into a coherent, flowing paragraph or a bulleted list if more appropriate for clarity, ensuring clinical relevance and brevity. Avoid repetition. Summary:
Structured, task-focused, reduced hallucinations

Engineering Rationale

The optimized prompt leverages several techniques to improve performance for Qwen 2.5 72B. Firstly, it establishes a clear 'role' for the AI ('AI assistant specialized in medical summarization'), which guides the model's tone and focus. Secondly, it explicitly states the 'goal' and 'focus' of the summary, preventing the inclusion of irrelevant information. Thirdly, the 'chain-of-thought' (CoT) prompting provides a structured, step-by-step reasoning process. This breaks down the complex task into manageable sub-tasks, guiding the model to systematically extract and process information before generating the final summary. This reduces hallucination and improves accuracy by forcing the model to consider specific categories of information. Finally, the explicit instruction to 'avoid repetition' and 'ensure clinical relevance and brevity' refines the output quality.

0%
Token Efficiency Gain
The optimized prompt significantly increases the accuracy of identifying primary diagnoses compared to the naive version.
The optimized prompt leads to a more structured and clinically relevant summary, with fewer extraneous details.
Summaries generated by the optimized prompt consistently include key treatment plans and follow-up recommendations.

Ready to stop burning tokens?

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

Optimize My Prompts