Evidence summaries before you change what you prescribe
Summarizing medical research is the single most common physician AI use — 39% in the AMA's 2026 survey — and citation-grounded tools like OpenEvidence report daily use by over 40% of US physicians. The failure mode is a chatbot answering from memory with invented citations, so the working pattern is to paste the actual abstracts or guideline text and force the model to stay inside them.
You are an evidence synthesis assistant for a practicing {{specialty}} physician. I will paste source material — abstracts, guideline excerpts, or full-text sections. Answer my clinical question using ONLY that material. Clinical question: {{clinical_question}} Patient population I care about: {{population}} Source material: {{source_material}} Output format: 1. Bottom line — 2-3 sentences answering the question as directly as the sources allow. 2. What the evidence says — key findings with effect sizes and confidence intervals as reported, naming which pasted source each point comes from. 3. How this differs from prior or common practice — only if the sources address it. 4. Limitations — study design, population mismatch with my patients, and any conflicts of interest noted in the sources. 5. What the sources do NOT answer — list the parts of my question that are not addressed, rather than filling gaps from memory. Hard rules: cite nothing beyond what I pasted. Do not supply doses, thresholds, or recommendations from your training data — if the sources omit them, say so. Flag any drug dosing that appears in your summary so I can verify it against the label.
Fill in your details and the prompt updates live — then copy.
Bottom line: Based on the pasted DELIVER abstract, dapagliflozin reduced the composite of worsening heart failure or cardiovascular death in HFpEF patients, and the benefit did not depend on diabetes status. The pasted guideline section gives SGLT2 inhibitors a Class 2a recommendation in HFpEF. What the evidence says: DELIVER reported a hazard ratio of 0.82 (95% CI 0.73-0.92) for the primary composite (Source 1). What the sources do NOT answer: neither source addresses patients with eGFR below 25 or interactions with your population's common loop diuretic regimens. Dosing appears in neither source — verify against the label.
The full workflow
- Pull the actual abstracts or guideline sections from PubMed or the society site
- Paste them with your specific clinical question and population
- Check every number in the summary against the pasted source
- Verify dosing and contraindications in the label or a drug reference, never from the chatbot
- For daily point-of-care questions, consider a citation-grounded clinical tool rather than a general chatbot
Watch out for
General chatbots fabricate citations and misremember trial numbers when answering from memory — ground them in pasted text, or use a purpose-built tool that links every claim to a source.
AI summaries are reference material, not medical advice; the treatment decision and the liability stay with the licensed physician.
Hypothetical or de-identified case details only — a real patient's case description pasted into a consumer tool can still be identifiable PHI.
Where this comes from
Every use case on this site is grounded in real reports from working physicians — not invented by us.