Open Evidence — AU & China Guidelines Strategy

Framework for referencing local clinical guidelines · March 2026 · Confidential

Recommendation: Start with Australia, License & Display. Clearest path, English-language, known counterparties, and mirrors what UpToDate and MIMS have already done. China should be Phase 2 with a dedicated local partner.

📋 Bottom Line Summary

	🇦🇺 Australia	🇨🇳 China
License & Display	✅ Viable — hardest part is eTG negotiation	⚠️ Complex — need local entity + legal
Train on Content	⚠️ Possible — premium cost + harder negotiation	🔴 High Risk — limited legal precedent, regulatory exposure

🇦🇺 Australia

Primary Sources

eTG Complete (Therapeutic Guidelines Ltd) — gold standard for bedside prescribing, used in virtually every Australian hospital
NHMRC — National Health and Medical Research Council, evidence-based guidelines across specialties
Specialty college guidelines — RACGP (GPs), RACP, RACS, etc.

Dimension	License & Display	Train on Content
AI Approved?	No blanket approval — eTG license explicitly restricts redistribution and automated processing. Negotiation required.	Harder ask — training rights are a separate, premium negotiation from display rights.
Contract Type	Commercial display license (eTG Ltd) + data use agreement (NHMRC). Specialty colleges negotiated separately.	Same + IP indemnification clause + explicit training rights addendum.
Steps	1. Map which guidelines to include by specialty 2. Contact eTG Ltd licensing team 3. Negotiate display/reference license 4. Agree on update cadence & versioning (eTG updates ~3x/year) 5. Build content integration pipeline 6. Legal sign-off	Steps 1–4 same as display, then: 5. Negotiate separate training rights (expect pushback) 6. Legal review of AI output liability 7. Define retraining governance & audit trail
Cost	Negotiated — likely mid-to-high five figures AUD/yr minimum. No public pricing.	No public precedent — expect 2–3x display license cost.
Access Method	No public API. Structured content export via negotiated data delivery agreement.	Same bulk export — you build the ingestion pipeline.
Key Risks	eTG may include AI-specific restrictions in license terms. Must be reviewed carefully.	IP liability if model outputs are misattributed to guidelines.
Staff Needed	Australian clinical editor, legal counsel, content ops.	Same + ML/data engineer + clinical safety reviewer.
Timeline	3–6 months	6–12+ months

🇨🇳 China

Primary Sources

CMA guidelines (中华医学会) — Chinese Medical Association clinical practice guidelines, hundreds of specialties
NHC clinical pathways (国家卫生健康委员会) — National Health Commission; mandated for hospital reimbursement
Chinese specialty society guidelines — sub-societies under CMA

Dimension	License & Display	Train on Content
AI Approved?	Legally unclear. CAC AI regulations (2023) require clear IP rights on source content. Most CMA guidelines copyrighted by journal publishers.	High legal risk — training rights not contemplated in existing agreements. Very limited precedent.
Contract Type	License with CMA Publishing House + government data agreement (NHC). Local Chinese entity likely required to hold contracts.	Same + explicit training rights + local Chinese entity required + IP indemnification.
Steps	1. Engage local legal counsel (PIPL, DSL) 2. Establish or partner with China-registered entity 3. Identify & prioritize guideline sources 4. Negotiate with CMA Publishing House 5. NHC pathway data agreement 6. CAC AI regulation compliance review 7. Build content integration	Steps 1–6 same as display, then: 7. Negotiate training rights (very hard, limited precedent) 8. NMPA medical AI software registration assessment 9. Data localization compliance (DSL/PIPL) 10. Ongoing regulatory monitoring
Cost	Opaque — no public pricing. Local entity setup adds significant overhead.	Substantially higher than display + regulatory compliance costs.
Access Method	No API. Web scraping + structured export via negotiated deal, OR partner with local health IT company (e.g., Yidu Cloud, LinkDoc).	Same — bulk content via local partner likely the only realistic path.
Key Risks	CAC AI rules, data localization (PIPL/DSL), IP ambiguity across sources.	NMPA device registration may be triggered depending on product claims. Regulatory exposure is significant.
Staff Needed	Local Chinese clinical editor (essential), China-registered legal counsel, local partner.	Same + ML engineer + regulatory affairs specialist (NMPA).
Timeline	9–18 months	18–24+ months

Note on China: Most Western CDS companies either partner with a Chinese health IT company or skip China entirely. The practical need for local presence, NMPA registration risk, and data localization requirements make this a substantial investment. Decide first: is China a near-term target or a 3-year horizon?

Prepared March 2026 · For internal use only