Skip to main content
Back to Library
Prompt Engineering Guide

Mastering Medical report summary
on Claude 3.5 Haiku

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

The "Vibe" Prompt

"Summarize this medical report in a way that a doctor can quickly understand the key information. Make it concise but include all important details. Ensure accuracy."
Low specificity, inconsistent output

Optimized Version

STABLE
You are a highly experienced medical professional assisting other doctors. Your task is to extract and synthesize the most critical information from a patient's medical report into a concise summary. This summary should enable a consulting doctor to grasp the patient's current status and history without needing to read the full report. Follow these steps: 1.IDENTIFY PATIENT DEMOGRAPHICS: Extract patient name, age, primary identifiers (if available). 2.HIGHLIGHT CHIEF COMPLAINT: State the primary reason for the medical visit or report. 3.SUMMARIZE CLINICAL HISTORY: Condense relevant past medical history, surgical history, family history, and social history directly pertinent to the chief complaint or current status. 4.LIST PERTINENT FINDINGS: Detail significant physical exam findings, lab results (with units if critical), imaging results, and specialist consultations. 5.OUTLINE DIAGNOSIS/IMPRESSION: State the confirmed or suspected diagnoses. 6.DETAIL TREATMENT PLAN & RECOMMENDATIONS: Describe current medications (dose, frequency), procedures, follow-up plans, and any specific recommendations. 7.FLAG CRITICAL ALERTS: Note any allergies, critical lab values, or urgent concerns. 8.SYNTHESIZE: Combine the above points into a coherent, flowing summary, prioritizing clarity and brevity. Ensure all medical terminology is accurate and appropriate. Aim for a summary readable in under 2 minutes. MEDICAL REPORT: [Insert Medical Report Text Here]
Structured, task-focused, reduced hallucinations

Engineering Rationale

The optimized prompt leverages chain-of-thought by breaking down the complex task of summarization into discrete, logical steps (1-7) before synthesizing (step 8). It establishes a clear persona ('highly experienced medical professional') and defines the target audience ('consulting doctor'), which guides tone and content selection. The explicit structure minimizes ambiguity, ensures consistent output, and forces the model to systematically cover all crucial aspects of a medical report, reducing factual omissions or hallucinations. It also sets an implicit constraint on conciseness ('readable in under 2 minutes') and explicitly demands accuracy ('all medical terminology is accurate and appropriate'). The naive prompt is vague, offering little guidance on what 'key information' or 'important details' means, which can lead to inconsistent and less comprehensive summaries.

0%
Token Efficiency Gain
The optimized prompt explicitly asks for 'critical alerts' (e.g., allergies), which is missing from the naive prompt.
The optimized prompt provides a structured output format, ensuring consistency across summaries.
The optimized prompt outlines specific categories of information to include (e.g., 'patient demographics', 'clinical history', 'treatment plan'), which the naive prompt lacks.

Ready to stop burning tokens?

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

Optimize My Prompts