Clinical evaluation is where many China registration projects gain or lose months. The framework under Decree No. 739 (Articles 24–29) and Order No. 47 (Articles 33–46) offers more flexibility than its reputation suggests — if the route is chosen deliberately and the evidence is structured to match reviewer expectations. This article maps the three routes, the exemption catalogue, the same-variety (equivalence) logic, when China requires local clinical trials, and how foreign clinical data can be used.
1. When is clinical evaluation required?
Medical device registration and filing generally require clinical evaluation. Two statutory exemptions exist (Decree No. 739, Article 24):
- devices with a well-defined working mechanism, settled design, mature production process, whose equivalents have been in long clinical use without serious adverse event records, and whose conventional use has not changed;
- devices whose safety and effectiveness can be demonstrated through non-clinical evaluation.
The NMPA maintains the catalogue of devices exempt from clinical evaluation (Order No. 47, Article 34) — it is the first document to check for any mature product category. Filing under the exemption still requires the full non-clinical evidence set: exemption from clinical evaluation is not exemption from evidence.
2. Route 1 — The exemption catalogue
If your product appears in the catalogue with a matching scope, the submission uses non-clinical evidence alone: product technical requirements, test reports against them, risk management, and the supporting research documentation. The practical work is confirming the catalogue entry matches your intended use — scope mismatches, not the catalogue itself, are the usual source of deficiency letters. Check the catalogue before planning any clinical work; a large share of mature devices never need a Chinese trial.
3. Route 2 — Same-variety (equivalence) clinical evaluation
Where no exemption applies, the default evidence path compares your device with an equivalent device already on the market and analyzes the equivalent’s clinical literature and data (Order No. 47, Articles 35–36). The submission must show:
- a structured comparison with the same-variety device: intended use, technical features, key performance, safety-relevant design choices;
- analysis and evaluation of the equivalent’s clinical data (published literature, and where available, long-term usage data);
- for every difference between your device and the equivalent: scientific evidence showing the difference does not adversely affect safety and effectiveness — bench data, additional testing or, where the gap cannot be closed, clinical data.
The equivalence route is documentation-heavy but trial-free when the differences are defensible. Choosing a defensible predicate device and closing the difference table early is the single best predictor of a smooth review.
4. Route 3 — Clinical trials in China
Trials are required where existing data cannot confirm safety and effectiveness (Decree No. 739, Article 25; Order No. 47, Article 35). The rules that matter most:
- GCP and institutions — trials follow China’s Good Clinical Practice for medical devices and run at qualified, filed trial institutions (Decree No. 739, Article 26);
- Filing before start — the sponsor files the trial with the provincial drug administration of the sponsor’s location before commencement (Order No. 47, Article 37);
- Approval for high-risk Class III — trials of high-risk Class III devices on the published catalogue need prior NMPA approval; the agency decides within 60 days of acceptance, and silence counts as consent (Decree No. 739, Article 27; Order No. 47, Articles 38 and 40). Approved trials must actually start within 3 years (Order No. 47, Article 45);
- Ethics and consent — ethical review and informed consent are mandatory; trial subjects may not be charged for participation (Decree No. 739, Article 28).
5. Using foreign clinical data
China accepts overseas clinical trial data under defined conditions — the NMPA’s published guidance on accepting foreign (境外) clinical trial data sets expectations on ethics compliance, study quality, population applicability and the need for Chinese-specific evidence where race-sensitive or population-dependent performance is in play. In practice, foreign data usually shortens rather than replaces the clinical chapter: expect a gap analysis against Chinese expectations, plus possible supplementary evidence for use conditions or populations not covered abroad.
6. Planning the route: what reviewers look for
Whatever the route, the review asks the same question: does the evidence chain prove your device, as designed and manufactured, is safe and effective for the claimed intended use in the intended population? Deficiency patterns concentrate in four places: predicate mismatches (route chosen before scope confirmed), difference tables left with open items, test reports not aligned with the product technical requirements, and foreign data submitted without an applicability analysis. Early route selection — ideally before bench testing is finalized — keeps the evidence chain coherent.
