🇺🇸 United States · OpenEvidence — AI for Clinical Decision Support
Status: 🟩 COMPLETE 🟦 LIVING Section: 10 — AI and LLMs
| Vendor | OpenEvidence |
| Country/origin | 🇺🇸 United States (founded by Daniel Nadler at Harvard) |
| Recommended for AUS? | ✅ Yes for clinicians — US-based; evidence-based; HIPAA-aware; growing Australian medical user base |
| Privacy summary | Designed for healthcare professionals; HIPAA compliant; standard SaaS data handling; clinical content not used for training without consent |
| Free tier | Yes — free for verified medical professionals |
| Paid tiers | Free for verified clinicians; enterprise tiers for institutions |
| First released | Founded 2023; rapid growth 2023–2024; full launch 2024 |
| Last reviewed | June 2026 |
| Official site | https://openevidence.com |
What it is
OpenEvidence is an AI-powered clinical decision support tool designed specifically for medical professionals — doctors, nurses, pharmacists, and other healthcare practitioners. It synthesises the latest medical research, clinical guidelines, drug information, and case-based evidence to answer clinical questions in seconds.
Think of it as: a clinical research assistant that has read the entire current medical literature and can answer your clinical question with citations.
Founded by Daniel Nadler (the entrepreneur behind Kensho, the financial AI acquired by S&P Global for $700M), OpenEvidence is one of the most-used AI tools in US healthcare. It’s particularly popular for:
- Differential diagnosis support: “47yo male with chest pain, normal ECG, mildly elevated troponin — what should I be considering?”
- Treatment recommendations: “First-line treatment for resistant hypertension in a Type 2 diabetic”
- Drug interactions and dosing: “Is fluoxetine safe in pregnancy? What are alternatives?”
- Guidelines synthesis: “What does the latest evidence say about colorectal cancer screening intervals?”
- Rare disease information: Answers questions about uncommon conditions where individual clinicians may have limited experience
How it’s different from ChatGPT for medical use
OpenEvidence and general AI assistants (ChatGPT, Claude, Gemini) for medical questions differ critically:
| Aspect | OpenEvidence | ChatGPT/Claude/Gemini |
|---|---|---|
| Source citation | ✅ Every claim cited to specific paper/guideline | 🔸 Limited or inconsistent |
| Current evidence | ✅ Continuously updated medical literature | ❌ Training data has cutoff |
| Hallucination risk | 🟡 Lower (grounded in citations) | 🔴 High for medical specifics |
| Audience | Medical professionals only | General public |
| Safety scope | Clinical decision support | Generic disclaimers |
| Designed by | Medical informaticists | Generalist AI labs |
Critically: OpenEvidence shows you the actual research papers behind its answers. You can click through and read the primary source. This is essential for clinical work where evidence quality matters.
How to access from Australia
OpenEvidence is free for verified medical professionals. To access:
- Go to https://openevidence.com
- Sign up with your work email (medical institution domain preferred)
- Verify your medical professional status (AHPRA registration in Australia, or equivalent international medical credentials)
- Once verified: full access to the tool
Verification typically takes 1-2 business days. Without verification, access is limited.
What you’d use it for as an Australian clinician
- At point of care: Quick lookups during patient consultations for unusual presentations
- Treatment planning: Synthesizing recent evidence for treatment options
- Continuing professional development: Staying current on guideline updates
- Patient education preparation: Finding evidence to share with patients
- Research support: Initial literature review for clinical research
- Junior clinician support: Education and guideline lookup for trainees
Australian medical context
- AHPRA professional responsibility: AI tools support clinical decision-making but don’t replace clinical judgment. The medical practitioner remains responsible for diagnosis and treatment decisions.
- MBS/PBS context: OpenEvidence’s recommendations may reference US-specific medications, dosing, and guidelines. Australian clinicians need to verify availability via PBS, applicability under MBS, and any Australian-specific guideline differences (RACGP, RACP, etc.).
- Therapeutic Goods Administration (TGA): Medication recommendations may need cross-checking with current TGA approvals and PBS listings.
- Indigenous health considerations: Few AI medical tools include specific evidence on Aboriginal and Torres Strait Islander health considerations. Use OpenEvidence as a starting point; reference Indigenous-specific guidelines (e.g., NACCHO resources) for relevant patients.
What it costs
OpenEvidence is free for verified medical professionals. The business model is enterprise institutional licensing and selective advertising (controlled).
Enterprise tiers for hospitals and health systems include:
- Institutional verification
- Custom integration with EHR systems
- Audit logs for clinical governance
- Custom medical knowledge bases
How it compares to other clinical AI
| Tool | Country | Focus | Free for clinicians |
|---|---|---|---|
| OpenEvidence | 🇺🇸 | Clinical decision support; latest evidence | ✅ |
| UpToDate | 🇺🇸 | Curated clinical guidance; subscription | ❌ (institutional) |
| Wolters Kluwer Lexicomp | 🇺🇸 | Drug information | ❌ |
| eMIMS | 🇦🇺 | Australian drug reference | Subscription |
| Hippocratic AI | 🇺🇸 | Patient engagement (not clinician) | Enterprise |
| General AI (ChatGPT, etc.) | 🇺🇸 | Not clinical-specific | Various |
OpenEvidence’s unique position: Free, evidence-grounded, citation-rich AI assistance — without the institutional subscription cost of UpToDate.
Privacy considerations
- HIPAA-compliant infrastructure
- For Australian clinicians: ensure use complies with state-specific medical practice regulations and AHPRA codes of conduct
- Patient information: Do not enter specific patient identifiers when querying OpenEvidence — describe cases in general terms (age, gender, condition, relevant history) without identifying details. This is the same principle for using any AI for clinical work.
- Practice notes and patient documentation should not be uploaded without appropriate institutional agreements
Gotchas
- AI is not a substitute for clinical judgment. OpenEvidence is a decision support tool — the clinician retains all clinical responsibility.
- US guideline orientation. Many recommendations reflect US guidelines (USPSTF, ACOG, etc.). Australian-specific guidelines (RACGP, RACP, RACS) may differ. Always cross-check with Australian guidelines for definitive recommendations.
- Drug names and availability differ. A medication recommended by OpenEvidence may not be PBS-listed in Australia, or may be available under a different brand name. Verify with eMIMS, MIMS Australia, or PBS.
- Evidence quality varies. OpenEvidence cites the literature it finds — but the quality of that literature varies. Apply your own critical appraisal.
- Not a substitute for specialist consultation. For complex cases outside your specialty, refer rather than rely on AI synthesis.
- Stay current on TGA safety alerts. AI may not surface very recent regulatory actions (TGA black triangle warnings, drug recalls). Check TGA notifications independently.
See also
- hippocratic-ai — patient-facing clinical AI (different use case)
- hallucinations — why citation-grounded AI matters for medicine
- deep-research-mode — research-grounded AI search
- ai-document-generation — for clinical documentation
Sources
- OpenEvidence official: openevidence.com
- Daniel Nadler interviews and OpenEvidence launch coverage (2023–2024)
- STAT News, Healthcare IT News coverage of clinical AI tools
- AHPRA Code of Conduct for medical practitioners
- RACGP guidance on AI in general practice (2024)
- TGA — Therapeutic Goods Administration on AI as medical device guidance