Digital Mental Health Technology: Device Characterisation, Regulatory Qualification and Classification

Key Findings

• — Qualification turns on two characteristics: the manufacturer's intended purpose (wellbeing support versus aiding treatment of a mental health condition) and functionality (how the product works).
• — 'AI chatbot' is listed alongside e-coach, questionnaires, triage and personalised content as a DMHT function; a function that processes data using AI is classed as high functionality, which bears on whether the product qualifies as a medical device.
• — Worked example: user sign-in through a generative-AI chatbot that responds to the user's input 'with the aim of creating a welcoming and compassionate environment' is rated high functionality even though its stated purpose is data collection.
• — The guidance enumerates 29 DMHT functions, from fixed-content psychoeducation to just-in-time adaptive interventions, digital games and VR/AR, each with qualification reasoning.
• — Version history: v1.1 (2025-02-13) reworded two examples; v1.2 (2025-07-03) added stakeholder-engagement detail, guidance on when a DMHT does not qualify as SaMD, and EU MDCG 2019-11 Rev 1 alignment; v1.3 (2026-07-29) changed the wording of function 11.

Methodology Notes

Regulatory guidance, not an empirical study. 54-page PDF (version 1.3, 886,831 bytes) published under the Open Government Licence at gov.uk; first published 2025-02-03, current version 2026-07-29. Developed with stakeholder engagement described in section 2. The document addresses qualification and classification only; clinical evaluation and evidence standards are covered by sibling MHRA/NICE outputs of the same programme. Companion gov.uk documents from the programme, 'Digital mental health technology: user and public perspectives on identifying and reporting harms' (2025-11-20) and 'Report on adverse event reporting in DMHTs' (2025-08-01), are not covered by this entry.

Sources