For most of its history, NHS mental health commissioning has measured what is easiest to count: referrals made, appointments attended, numbers seen. Whether those appointments helped the people attending them was rarely part of the picture. As the Public Accounts Committee noted in 2023, data on NHS mental health services still lagged well behind that for physical health, with particular concern about the lack of data on patient outcomes and experiences. Commissioners could tell you how many people had passed through a servicebut not reliably tell you how many had improved.
The NHS 10-year plan commits to expanding outcome measurement across all mental health services, building on what NHS Talking Therapies demonstrated is possible when outcome data is collected as routine. The shift has direct commercial implications for pharmaceutical and medtech companies working in mental health: the evidence commissioners now want to see is different from what was sufficient five years ago, and the gap between clinical trial data and real-world NHS performance is under more scrutiny.
Before IAPT (Improving Access to Psychological Therapies) launched in 2008, outcome data collection in NHS psychological therapies was patchy at best. Attempts to measure outcomes at the start and end of treatment were undermined by high rates of missing data: patients who dropped out, who tended to have worse outcomes, simply disappeared from the record. Services were commissioned on the basis of activity, and activity grew, but there was no reliable way to know whether the growth in appointments translated into growth in recovery.
IAPT solved this with a methodological change that was straightforward in design and significant in effect. Rather than asking patients to complete outcome questionnaires only at the start and end of treatment, the programme collected standardised symptom scores at every session. The result was outcome data for 98% of patients completing a course of treatment, a completion rate that transformed what commissioners could see and act on. For the first time, it was possible to compare recovery rates across services, identify what was working and what was not, and tie continued investment to demonstrated outcomes rather than headcount.
NHS Talking Therapies, which succeeded IAPT in 2023, has built on that foundation. The programme treated 670,419 people completing a course of treatment in 2024-25, up from 144,210 in 2012-13. The government's target is that 50% of eligible referrals should move to recovery, and the 2026-27 Medium Term Planning Framework sets a further ambition of 51% reliable recovery and 69% reliable improvement by the end of 2026-27, with commissioning targets tied to outcome measures, not throughput.
The programme also illustrates the limits of activity data as a proxy for impact. Referrals have grown from around 880,000 in 2012-13 to 1.8 million in 2023-24. But the proportion of referrals completing treatment has remained around 37% throughout that period, with attrition a persistent problem. Growing the number of people entering a service tells you something about access but it tells you much less about whether the service is working for the people who need it most.
NHS Talking Therapies operates primarily in primary care, covering depression and common anxiety disorders. The harder measurement challenge has always been community mental health: services for people with severe and complex needs, where outcomes are harder to define, recovery takes longer, and the relationship between clinical contact and improvement is less direct than it is in a structured course of therapy.
Research in community mental health services has consistently found that higher contact frequency between clinicians and patients does not reliably correlate with better clinical outcomes. More appointments do not necessarily mean more improvement. Without standardised outcome data collected as routine, there is no way to identify which interventions are working, for which patients, or under what conditions. Services optimised for throughput can look productive while delivering little measurable benefit.
NHS England has addressed this by recommending three Patient-Reported Outcome Measures (PROMs) for use across community mental health services, with full implementation across all services. Each is completed by the service user rather than the clinician, capturing a dimension of experience and recovery that clinical assessment alone cannot reliably access:
The Royal College of Psychiatrists’ implementation guidance makes clear that these measures are intended to be embedded in electronic patient record systems and collected as routine, not as a research add-on. Early implementations are beginning to show what routine DIALOG data can reveal across large patient cohorts: which aspects of care and life satisfaction tend to improve over time, which remain static regardless of clinical contact, and where the system is consistently failing to address what matters most to the people using it. This is population-level insight that NHS mental health services have not previously had.
For pharmaceutical companies with medicines in mental health, the shift to outcome-based commissioning changes what the NHS expects an evidence package to demonstrate. Clinical trial data remains the foundation, but the questions that committees and commissioners are increasingly asking go beyond primary endpoints.
NICE appraisals for mental health medicines now routinely consider patient-reported quality of life data alongside clinical efficacy measures. As PROMs become standard in community mental health services, commissioners will also look at whether real-world NHS data, collected through the same instruments as those now embedded in service delivery, supports the outcomes predicted by trial evidence. A medicine that performs well in a trial population but whose real-world NHS performance in complex patients is unknown will face harder questions at appraisal and at local commissioning level.
The practical implication is that post-launch evidence generation in NHS mental health populations is likely to become a standard commercial expectation. Companies that build PROMs collection into real-world evidence programmes from the point of launch, designing studies that use the same instruments NHS services are now collecting routinely, are better placed to connect to the data infrastructure now being built.
For medtech and digital health companies, the evidence generation challenge is more immediate and more specific. NICE’s first published early value assessment under the new HealthTech Programme was for digital CBT technologies for children and young people with anxiety and low mood: a conditional recommendation allowing NHS use while further real-world evidence is generated. The evidence generation plan published alongside that assessment sets out specifically what data is needed to convert the conditional recommendation into full guidance:
For companies in early value assessment, the path from a conditional recommendation to full guidance, and potentially to the National HealthTech Access Programme with its mandatory funding obligation, runs through the evidence generation period. Companies that actively engage with their evidence-generation plans, rather than treating them as a post-recommendation administrative step, are considerably better placed to complete that journey.
The alignment between NHS PROMs and the outcome measures NICE asks for in evidence generation plans is also an opportunity. Technologies that are designed to collect and report on standardised measures, in formats that connect with NHS electronic patient record systems, are easier for NHS services to adopt and easier for NICE committees to evaluate. Building that alignment in early, rather than retrofitting it during the evidence generation period, reduces friction at every stage of the market access process.
The shift from activity to outcomes in NHS mental health commissioning is being operationalised through the NHS Talking Therapies programme, through the PROMs framework now being embedded in community mental health services, and through the 10-year plan’s explicit commitment to expand outcome measurement further.
For pharmaceutical and medtech companies, the adjustment is commercial: understanding that the questions commissioners ask have changed, and that evidence packages need to be designed to answer those questions from the outset. A medicine or technology with a strong evidence base in NHS mental health populations, measured on the instruments commissioners are now using to assess value, has a genuinely stronger market access position than one that requires commissioners to extrapolate from trial data alone.
In a system that is finally asking the right questions about whether care is working, that is a meaningful advantage.
This post is part of a series on mental health as a commercial and clinical priority. Read the first post: Mental health as a national priority: what the new landscape means for pharma and medtech.
CHASE works with pharmaceutical, medtech and digital health companies on NHS commercial strategy and NHS–Industry Partnerships, including in mental health. Our NHS–Industry Partnerships team supports companies navigating NICE evaluation, NHS procurement, and evidence-generation planning across mental health and beyond. Get in touch to find out more about how we can work with you.
Explore our latest thinking, event updates and industry insights to stay informed.
Precision medicine makes pharma and medtech convergence a commercial necessity. CHASE examines what that means for companies and the NHS in practice.
NHS AI tools that work technically are common. Tools that change care at scale are not. CHASE examines what genuine pathway improvement from AI requires.
Record NHS mental health spending, a new cross-government strategy and the neighbourhood health shift are reshaping commercial landscape for pharma and medtech
.svg)
