The MHRA weighs in on Digital Mental Health Technologies (DMHTs)

Across the world, demand for mental-health support is exploding. Anxiety, depression, and sleep disorders are no longer fringe concerns but are headline issues, with health services struggling to keep up with the demand. Long waiting lists, staff shortages, and patchy access mean millions of people are left to cope alone.

Into this gap has rushed a wave of Digital Mental Health Technologies (DMHTs) such as apps, platforms, and AI-powered tools promising fast, private, and tailored support. The global market has ballooned, and with it, expectations. These tools are no longer just digital diaries; many are edging into territory traditionally occupied by clinicians, triggering closer regulatory scrutiny.

The newest generation of DMHTs is bold. Some analyse language to detect emotional shifts, others triage symptoms, and many claim to “guide” users along therapeutic paths. But when does helpful become clinical? And when does clinical become risky? As developers increasingly turn to AI, adaptive algorithms, and conversational agents, regulators are asking one crucial question: is this tool simply supporting well-being, or is it behaving like a medical device? The answer is critical to preserving patient safety and determines whether a product can be placed on the market. 

To help clarify this point, the UK’s Medicines and Healthcare products Regulatory Agency (MHRA) has weighed in on the issue providing dedicated guidance to help manufacturers determine when a Digital Mental Health Technology qualifies as a medical device. This guidance is part of a wider initiative supported by public investment to build safer, clearer regulatory expectations for mental-health-focused digital innovation. The framework offers clarity in two key areas:

1.      Intended purpose: what the tool claims to do.

2.      Functional impact: what the tool is actually capable of doing in practice.

A DMHT may fall under medical-device regulation if it is intended to diagnose, treat, mitigate, or monitor a mental-health condition. However, the MHRA has emphasised that functionality matters just as much as stated intent. A tool stating that it “measures depressive severity” or “helps identify anxiety disorders” is likely to be considered medical in purpose, while one stating that it “supports general well-being”, “encourages reflection”, or “promotes mindfulness” may fall outside medical regulation.

One of the most significant shifts in the MHRA’s guidance is the emphasis on functional impact which is equivalent to how much influence the tool has on users’ mental-health decisions or outcomes. Specifically, low-functionality DMHTs typically have limited influence on clinical decisions and do not provide personalised interpretation with therapeutic consequences and could be simple symptom checklists, tools that total up questionnaire scores (e.g., summing mood-survey responses) without giving diagnostic statements, apps or portals that deliver fixed information about stress, sleep hygiene, or resilience. Low-functionality tools are often excluded from medical-device classification because they don’t directly influence diagnosis or treatment decisions.

High-Functionality DMHTs on the other hand are tools that actively interpret user information, generate tailored insights, or influence clinical actions. They typically use AI, generative models, or adaptive questioning systems. These may be chatbots that analyse user input and provide tailored emotional or behavioural guidance, rather than pre-written generic responses, dynamic risk-assessment tools or systems that ask variable, algorithm-driven follow-up questions and output a probability or severity label for mental-health conditions or even automated clinical summaries that  record consultations and produce structured summaries to be incorporated into medical records without final clinician review.

The MHRA’s guidance marks a significant step toward creating a structured and predictable regulatory environment for digital mental-health innovation but it does not provide the definitive outline for the developing world of DMHTs. Developers of AI-powered DMHTs should be prepared to monitor changes to guidelines closely and align with broader software and AI regulatory programmes such as the MHRA’s Software and AI as a Medical Device Change Programme Roadmap. Understanding these requirements is essential not only for UK market access but also for operating in jurisdictions that are developing similar frameworks for AI-enabled medical technologies. Manufacturers that invest early in regulatory compliance, transparency, and robust safety measures will be better positioned to react in a timely manner and maintain their competitive advantage.