Dr Rajiv Chandegra
12.12.2024·Healthcare, Technology, Policy

Part of series: UK Health Tech Compliance·Part 2 of 3

Navigate:
Pathway through a structured environment

UK Health Tech Compliance: Design the Pathways

Building on Part 1: See the System, where we mapped the five regulatory subsystems and identified leverage points, this guide focuses on designing the specific pathways through each requirement.


MHRA registration and marking

The UKCA marking requirement eventually replaces CE marking, though CE marked devices remain acceptable until:

  • June 30, 2028 - Legacy devices certified under older directives
  • June 30, 2030 - Devices certified under EU MDR

This extended transition provides strategic flexibility. Most companies maintain CE marking for dual UK/EU market access rather than pursuing UKCA-only routes.

Non-UK manufacturers must appoint a UK Responsible Person who:

  • Registers devices with MHRA (currently £261 per application, transitioning to ~£300 annually per GMDN Level 2 category from April 2026)
  • Maintains technical documentation accessible to MHRA
  • Serves as regulatory liaison for all UK compliance matters

Clinical evaluation requirements

Evidence requirements scale with classification:

ClassEvidence RequirementTypical Cost
Class ILiterature review, equivalence£5,000-£15,000
Class IIaRobust clinical data£15,000-£30,000
Class IIbComprehensive clinical data£25,000-£50,000
Class IIIClinical investigations£100,000-£5,000,000+

Software and AI pathways

The MHRA Software and AI as Medical Device Change Programme establishes framework for classifying software based on functional impact. The ten Good Machine Learning Practice principles (developed jointly by MHRA, FDA, and Health Canada) emphasise:

  • Multi-disciplinary expertise throughout product lifecycle
  • Representative training data independent from test datasets
  • Focus on human-AI team performance, not standalone algorithm metrics
  • Testing in clinically relevant settings
  • Deployed model monitoring in real-world environments

For adaptive AI systems: Predetermined Change Control Plans (PCCPs) must define bounded modifications, update methodologies, impact assessment procedures, and implementation controls - approved before deployment.


NICE evidence generation

"Sleepio (MTG70) received first-ever digital therapeutic guidance in May 2022 after eight years of evidence generation including twelve RCTs. NICE confirmed Sleepio as clinically effective and cost-saving at £4.52 per person over standard treatment after one year."

From April 2026, NICE technology appraisal programme extends to medical devices, diagnostics, and digital health technologies - placing these on legal parity with medicines. NHS organisations will be legally required to provide funding within three months of positive NICE recommendations.

Early Value Assessment (EVA) offers faster pathway for promising technologies:

  • 3-6 months for initial review
  • 12-24 months for evidence generation with NHS deployment
  • Full NICE guidance upon evidence maturity

NHS Digital Technology Assessment Criteria

DTAC divides into five assessed sections:


Timeline reality by classification

ClassificationTimeline to MarketPrimary Driver
Class I (self-certified)6-12 monthsDocumentation completeness
Class I (UK Approved Body)12-18 monthsSterile/measuring function assessment
Class IIa18-24 monthsUK Approved Body review (3-6 months)
Class IIb24-48 monthsClinical investigation (6-18 months)
Class III36-60+ monthsClinical investigation (12-36 months)

What's next

With the pathways designed, Part 3: Implementation & Case Studies covers quality management foundation, acceleration strategies, post-market surveillance, NHS integration, and real-world case studies from Sleepio, Skin Analytics, and Doccla.


Related reading: