Medical Device Compliance & Regulatory Newsletter August 2026

Importance

Generative AI promises transformative benefits for healthcare personalized outputs, adaptive decision support, and enhanced patient-clinician interactions. But it also introduces risks such as:

• Hallucinations/confabulations that appear authentic but are incorrect.
• Unclear boundaries of intended use, especially with open-ended inputs.
• Opaque foundation models developed by third parties.
• Performance drift across different environments and over time.

FDA’s Center for Devices and Radiological Health (CDRH) is seeking stakeholder input to shape a nimble, risk-based regulatory approach.

Risk Framework Proposal

The paper introduces a two-axis risk framework:

• Activity axis: How independently the device directs or takes action (from non-directive information to fully autonomous action-taking).
• Consequences axis: The severity of harm if outputs are incorrect.

Risk increases as devices move toward autonomous action-taking with severe consequences.

FDA is asking for feedback on how to apply this framework to:

• Action-directing vs. non-directive outputs (e.g., general info vs. specific treatment recommendations).
• Patient-facing vs. clinician-facing outputs, recognizing differences in domain knowledge.
• Generalist vs. specialist use cases, where expertise affects risk.
• Multi-turn conversations, where outputs may shift from informational to directive.
• Care escalation functions, balancing risks of under-escalation (missed care) vs. over-escalation (unnecessary interventions).

Competency-Based Evaluation

Traditional device evaluation methods may not suit GenAI’s variability. FDA is considering a competency-based approach, inspired by how clinicians are trained and licensed:

• Defined scope of practice.
• Time-limited certification.
• Ongoing monitoring and reevaluation.
• Accountability for developers and deploying institutions.

This model could ensure GenAI-enabled devices remain safe and effective throughout their lifecycle.

What’s Next

FDA is requesting feedback from industry, clinicians, patients, and regulators. Key areas for input include:

• Risk assessment dimensions beyond activity and consequences.
• Safeguards for patient-facing outputs.
• Evaluation methods for multi-turn conversational devices.
• Balancing innovation with accountability.

Conclusion

Generative AI-enabled devices represent a new frontier in MedTech regulation. FDA’s discussion paper signals openness to innovative oversight models while reinforcing its commitment to patient safety. For manufacturers, consultants, and healthcare leaders, this is the moment to engage with FDA, shape the framework, and prepare for a future where GenAI becomes integral to medical device functionality.

https://www.fda.gov/medical-devices/digital-health-center-excellence/considerations-regulation-generative-ai-enabled-medical-devices-discussion-paper-and-request

Therapeutic Goods Administration (TGA) has set out clear expectations for manufacturers of in-vitro diagnostic (IVD) self-tests for chlamydia, gonorrhoea, and syphilis. The update provides a clear picture of what manufacturers and sponsors need to consider when preparing these self-tests for the Australian market.

Importance

STIs a growing concern in Australia, with rising report for chlamydia, gonorrhoea, and syphilis over the past decade. While these are treatable with antibiotics, early detection is critical to reducing transmission and complications. Self-tests can bridge gaps in access, especially in regional and remote communities.

Clinical performance expectations

The TGA requires self-tests to meet high standards of sensitivity and specificity:

• Chlamydia & Gonorrhoea: ≥95% sensitivity and ≥99% specificity
• Syphilis: ≥95% sensitivity and ≥98% specificity

These thresholds ensure that tests are accurate enough to be trusted in the hands of lay users.

Usability and comprehension

Since these tests are designed for non-professionals, usability studies are important and essential. The TGA expects manufacturers to demonstrate that instructions are clear, results are interpretable, and invalid test rates remain below 5%.

Risks and mitigation

Self-testing carries inherent risks:

• Incorrect specimen collection
• Misinterpretation of weak or invalid results
• Failure to seek medical follow-up

For syphilis, antibody-based tests add complexity, as they cannot distinguish between active and past infections. The TGA emphasizes that risk mitigation must be built into device design, labelling, and user education.

Post-Market Monitoring and Standard Conditions of Inclusion

Once self-tests are included in the Australian Register of Therapeutic Goods (ARTG), manufacturers and sponsors must comply with ongoing post-market monitoring requirements.

The TGA expects sponsors to:

• Conduct active post-market surveillance to identify emerging risks, including false negatives, false positives, or usability issues.
• Report adverse events and performance concerns promptly, ensuring transparency and accountability.

Final thought

The TGA’s updated guidance provides a roadmap for responsible innovation. By setting clear performance thresholds and emphasizing usability, risk mitigation, and confirmatory testing, the regulator ensures that self-tests are safe in the hands of everyday users.

For manufacturers and sponsors, the challenge is to balance accessibility with accuracy.