← Back to blog
BLOG · AUGUST 2026

Why Most AI Clinical Safety Cases Miss the Point

DCB0129 compliance is not the same as clinical safety. Here is why the gap matters - and what good looks like.
SARAH AMANI15 AUGUST 20262 MIN READ

The number of AI products entering NHS clinical workflows is accelerating. The number with rigorous clinical safety cases is not keeping pace.

The Compliance Trap

Too many vendors treat DCB0129 as a documentation exercise - a hurdle to clear before go-live. The Clinical Safety Case Report becomes a retrospective justification rather than an active risk management tool.

This matters because AI systems behave differently from the static digital health tools DCB0129 was designed for. They drift. They hallucinate. They perform differently across populations. A safety case written at a point in time decays the moment the model is retrained or the user population shifts.

What Good Looks Like

A meaningful clinical safety case for AI should:

  • Map hazards to real clinical workflows, not hypothetical ones. If clinicians are using an ambient scribe in a mental health assessment, the hazard log needs to account for what happens when affect, tone, and non-verbal cues are lost in transcription.
  • Define ongoing monitoring obligations, not just pre-deployment checks. How will you know when the system's performance has degraded below your acceptable threshold?
  • Be honest about residual risk. Every system carries residual risk. The question is whether that risk is understood, quantified where possible, and accepted by an accountable clinical safety officer - not buried in an appendix.

The Role of the Clinical Safety Officer

An accredited CSO is not a rubber stamp. They are the named individual accountable for confirming that clinical risks have been managed to an acceptable level. That accountability means being willing to say no - to delay a deployment, to require additional mitigations, to escalate when commercial pressure conflicts with patient safety.

This is uncomfortable. It is also the point.


TrustPoint Digital Health provides independent clinical safety governance for AI and digital health products under DCB0129 and DCB0160. Get in touch to discuss your safety case.

CLINICAL SAFETY & DIGITAL HEALTH & CLINICAL AI GOVERNANCE

TrustPoint provides independent clinical safety governance for digital health and clinical AI - DCB0129/DCB0160 safety cases, readiness reviews, and post-market surveillance.

Book a consultation ↗Email us ↗