Clinical Evaluation Requirements for Thailand Medical Device Registration

1. Key takeaways

Clinical data requirements in Thailand follow the risk class. Class 4 devices must submit the complete CSDT together with clinical evidence, and the Thai FDA review of clinical evaluation for high-risk products has become increasingly strict, demanding a thorough literature evaluation or a comparison report against equivalent competitor devices. Class 2 and Class 3 devices may need clinical or performance data depending on the nature of the device and the approvals already obtained in reference countries. The practical consequence is that the clinical evaluation report is not an optional annex but a core module of the dossier, and for Class 4 it is a hard gate: no pathway choice removes the evidence requirement, and pathway selection can only compress review time, not cancel the evidence obligation. Teams should therefore treat the CER as a project-critical deliverable with its own schedule, its own owners, and its own dependencies, rather than as a document assembled at the end from whatever the technical files happen to contain.

• Class 4: complete CSDT plus clinical evidence, with increasingly strict review of the clinical evaluation and demands for detailed literature evaluation or equivalent-device comparison reports.
• Class 2/3: clinical or performance data may be needed depending on the device and on approvals already obtained in reference countries.
• The CER is organised in three layers: clinical background and state of the art, the evaluation evidence itself, and conclusions with the risk-benefit determination.
• The evidence chain must be complete; submitting conclusions without the search strategy, the literature list, and the comparison logic triggers a deficiency request.
• Foreign clinical data can generally be relied on, but whether local bridging evidence is needed should be confirmed for the specific device; Indonesian rules on mandatory local clinical verification must not be applied to Thailand.
• For Class 3 and Class 4 SaMD, clinical evidence expectations approach the level of the EU MDR and the US FDA, and AI-based continually learning algorithms require a locked version at submission plus a post-market update strategy.

2. Legal basis

The document requirements come from the Thai FDA website Class 2-4 document list matrix, which shows the exemptions available on the Abridged route; those exemptions cover the verification and validation summary, the risk analysis, and disposal documentation, and they do not include the clinical evaluation. The Thai FDA review interpretation for Class 4 clinical evaluation calls for a literature evaluation or a comparison report against equivalent devices. The MoPH classification notification B.E. 2562 (2019) provides the Class 4 examples used to determine which products fall into the highest-risk tier. For software as a medical device and AI algorithms, the requirements for a locked version at submission and a post-market update strategy apply to Class 3 and Class 4 products. On accelerated routes, the Abridged route exempts the verification and validation summary, the risk analysis, and disposal documentation but not the clinical evaluation, while the Reliance route based on HSA approval exempts the expert review, also not the clinical evidence itself; the Reliance exemption was established by the announcement of 2024-06-04, issued in B.E. 2567.

3. Expert insights

The most common sequencing error in Thai projects is starting the clinical evaluation report after the rest of the CSDT is assembled. The correct point to begin the CER is before CSDT compilation starts, for two reasons. First, the literature search and the equivalent-device comparison take time, and they may require external database resources that need to be arranged in advance. Second, if the search reveals that the evidence is insufficient, there is still time to run supplementary verification testing or to adjust the submission strategy; a team that leaves the CER to the end discovers an evidence gap when the whole project must stop and wait for data. A second insight concerns borders: Thailand follows a practical estimate under which foreign clinical or technical data can usually be relied upon, but whether local bridging evidence is needed must be confirmed device by device, and the Indonesian approach of mandatory local clinical verification must not be borrowed. Finally, for AI-based continually learning algorithms, the locked version requirement deserves careful drafting: the submitted version fixes a definite algorithm version, and the dossier must explain how subsequent version updates will be handled, whether through a variation or a new submission. Ambiguity here resurfaces at every future update, so resolving it once, clearly, pays for itself.

4. Practical recommendations

• Start the CER before CSDT compilation begins, so that literature searching, equivalent-device comparison, and any evidence gap testing happen while the project can still absorb them.
• Structure the CER in three layers: clinical background and state of the art, evaluation evidence covering literature, equivalent-device comparison and clinical trial data, and conclusions with the risk-benefit determination.
• Include the search strategy, the literature list, and the comparison logic in the submission; conclusions alone will trigger a deficiency request.
• For Class 2/3, assess early whether clinical or performance data are needed given the device nature and approvals already held in reference countries.
• For Class 3/4 SaMD, prepare the locked version statement and the post-market update strategy for AI-based continually learning algorithms before submission.
• Do not assume the Abridged or Reliance routes remove clinical evidence: the Abridged route exempts the verification and validation summary, risk analysis and disposal documentation, and the Reliance route exempts expert review, while the Class 4 clinical evidence gate stands in all cases.

5. Verification checklist

• □ Verification basis date: 26 October 2026.
• □ Thai FDA website Class 2-4 document list matrix, confirming the Abridged exemptions do not include the clinical evaluation.
• □ Thai FDA review interpretation for Class 4 clinical evaluation, requiring a literature evaluation or an equivalent-device comparison report.
• □ MoPH classification notification B.E. 2562 (2019), Class 4 examples.
• □ Requirements for the locked version and post-market update strategy for SaMD and AI algorithms.
• □ HSA Reliance route exemption of expert review, announcement of 2024-06-04, B.E. 2567.

How GRZAN can help. GRZAN provides medical device and IVD registration, quality system, authorised representative and local testing support across Southeast Asia, including Thailand. Our team prepares dossiers to ASEAN CSDT structure, manages submissions and deficiency responses with the Thai FDA, and supports post-market vigilance obligations. Business enquiries: info@grzanglobal.com