The diagnostics advisory committee considered evidence on digital front door technologies to gather service user information for NHS Talking Therapies for anxiety and depression assessments (from now, NHS Talking Therapies) from several sources. This included evidence submitted by Limbic and Wysa, a review of clinical and cost evidence by the external assessment group (EAG), and responses from stakeholders. Full details are available in the project documents for this guidance.
Digital front door technologies are indicated for people over 16 years with suspected common mental health conditions, as specified in the NHS Talking Therapies for anxiety and depression manual (2024). Many people may have more than 1 of these conditions.
The NHS Talking Therapies services pathway is divided into 5 steps in the NHS Talking Therapies manual: presentation, referral, pathway starts, assessment and next steps. Referrals to NHS Talking Therapies can come from primary, secondary or community care, or through self-referral. Methods include paper forms, letters, phone calls, emails, online forms or digital front door technology. Both people using the service and healthcare professionals can initiate referrals. Sometimes, healthcare professionals may guide people to self-refer and provide information on available methods, or help them in the process. The clinical experts highlighted that the referral process for assessments for NHS Talking Therapies in current practice varies between NHS Talking Therapies providers. While all referrals include routine information collection, the amount and type of information collected can vary.
The referral process includes checking whether a person is at risk from NHS Talking Therapies, or eligibility for them. It also includes potentially considering likely suitability for NHS Talking Therapies from other information included in the referral or possibly other healthcare records to which the service has access. This is also called triage or screening. If there are concerns about risk, a healthcare professional will contact the person to ensure safety, gather more information and direct them to appropriate services. Eligibility is based on age and GP location, with some services also available for young people aged 16 to 17 years. People for whom NHS Talking Therapies services are unsuitable are not offered an initial clinical assessment and are sent a letter explaining the referral rejection. The method of triaging people for assessments for NHS Talking Therapies varies, for example, it might include:
doing eligibility checks at initial clinical assessments
eligibility checks led by an administrator or healthcare professional before initial clinical assessments.
The comparator for this early value assessment was the process of referring people to NHS Talking Therapies without using digital front door technologies.
In England, 1 in 6 people have a common mental health condition (like anxiety and depression) in any given week (McManus et al. 2016). The Five Year Forward View for Mental Health from 2016 set out that NHS England should increase access to evidence-based psychological therapies to reach 25% of people in need. This was so that at least 600,000 more adults with common mental health conditions could access NHS Talking Therapies services each year by 2020 to 2021 (1.5 million in total). The NHS Long Term Plan then increased this to an additional 380,000 adults having access by 2023 to 2024 (1.9 million in total). The clinical experts said that the key access number is no longer the number of people seen for an assessment. Rather, it is the number of people who have an assessment and then have a course of treatment. Currently, this is about 670,000 people annually (around 13% of the community prevalence of anxiety and depression). But this number is expected to rise substantially over the next 5 years.
There are several challenges when accessing NHS Talking Therapies. One issue is information sharing. The quality of collected information before assessments for NHS Talking Therapies is often poor, which means that healthcare practitioners have to spend extra time gathering the necessary information during the actual assessments. Some services have to manually copy and paste information to populate fields. Also, manually entering referral details for people for clinical assessments for NHS Talking Therapies can be a burden for administrative staff. Another significant challenge is improving access to NHS Talking Therapies services. The committee noted that it is important to make it easier for people to refer themselves for assessments whenever they need to, rather than relying on referrals from healthcare providers. This could help more people get the support they need in a timely manner. Also, some people find face-to-face interactions uncomfortable, which can be a barrier to seeking help.
Digital front door technologies enable people to refer themselves for assessment at any time to access NHS Talking Therapies services. They capture information at the point when people are seeking help by offering 24‑hour access, 7 days a week, through web portals and mobile apps. Digital front door technologies also remove the need for face-to-face interactions, which could promote access for those who find this a barrier. Many incorporate artificial intelligence (AI)‑driven symptom checkers and have decision-support tools that automatically signpost people who are not eligible for other services. A key innovation is that they can interface with electronic health records and improve clinical workflows by streamlining processes.
The committee considered evidence from 10 sources for Limbic Access and 3 studies for Wysa Digital Referral Assistant (DRA). Data for Limbic Access was available from multiple sources including:
2 large UK-based peer-reviewed studies
1 validation study
2 company evaluation studies
1 online survey of psychological wellbeing practitioners
1 evaluation of feedback responses from users of Limbic Access
1 research study about model training
2 NICE request-for-information responses.
Data for Wysa DRA was real-world experience from users reported in 3 Wysa NICE request-for-information responses. None of the data provided was comparative or sourced from published research studies. Most of the evidence came from routine data collected by NHS Talking Therapies services or company-reported data on the performance of their digital front door technologies. The evidence for patient-reported outcomes was sourced from user responses to questions asked after completing referral information through a digital front door technology. Most studies were non-comparative, primarily evaluating the strengths and weaknesses of these technologies.
The extent to which the available clinical evidence is generalisable to all NHS Talking Therapies service providers is unclear. This is because the referral process and system for people to access initial assessment varies across NHS Talking Therapies services. The committee noted that it was not possible to synthesise the limited available clinical evidence because of the heterogeneity of the non-comparative data. The committee acknowledged that there was more evidence for Limbic Access than for Wysa DRA. But it noted that the overall quality of evidence assessed was broadly comparable for Limbic Access and Wysa DRA. This was not a reflection of equal strength in peer-reviewed evidence but a comment on the source of most of the information available for both technologies at the time of assessment. Most of the information came either directly from the companies themselves, or from published data for which most of the authors were company affiliated. This did not invalidate the findings, but it did highlight the importance of independent, externally reviewed research. But the committee acknowledged that the published data (only available for Limbic Access) was in peer-reviewed studies.
The committee noted a lack of evidence on the impact for people unable to access the service or for people referred elsewhere. The clinical experts highlighted that triaging people out of the service without a formal clinical assessment does not align with the NHS Talking Therapies manual. They emphasised that the technology should prioritise collecting risk-related information to inform clinical decision making rather than excluding people before the assessments. The companies clarified that their technologies do not automatically exclude users. One clinical expert said that it is the companies' responsibility to manage clinical risk by ensuring compliance with the Digital Clinical Safety Assurance process. This verifies that health information technology used by healthcare professionals is safe and meets national standards. Also, the EAG addressed that digital front door technologies are an optional referral method because other pathways to NHS Talking Therapies assessments remain available. In addition, the evidence showed high levels of self-reported satisfaction among people using digital front door technologies to access NHS Talking Therapies. Also, the committee acknowledged that the technology is used to gather information to inform and help the healthcare professional when doing an assessment for NHS Talking Therapies. It does not make a clinical judgement independently from the healthcare professional. Based on these considerations, the committee concluded that digital front door technologies are low risk to use with further evidence generation.
The committee considered the potential benefits of using digital front door technologies to improve the quality of clinical assessments for NHS Talking Therapies. The clinical experts explained that routine information collection by less experienced staff in current practice is often insufficient, leading to incorrect diagnosis. Digital front door technologies could enhance routine information gathering before assessment, helping healthcare professionals make more informed decisions during the initial clinical assessment. This could result in more accurate diagnoses and improved treatment pathway selection. But the clinical and patient experts also raised concerns that the AI‑driven algorithms in some of these technologies may selectively present information to healthcare professionals. This could, in some circumstances, influence clinical decision making in ways that increase the risk of inaccurate diagnosis. The committee discussed that further evidence comparing clinical assessment outcomes with and without digital front door technologies is needed. The clinical experts were asked how the quality of information collected by the technologies could be estimated. They explained it could be done by comparing the problem descriptors like ICD‑10 diagnostic codes identified in initial clinical assessment with those from an internationally recognised 'gold standard' assessment.
The committee discussed the potential time saved on collecting routine information from using digital front door technologies during initial clinical assessments. Both Limbic and Wysa reported that their technology can reduce assessment time. Limbic Access reportedly saves 12.7 minutes (data from a peer-reviewed study) and Wysa DRA reportedly saves between 16.0 and 21.0 minutes (data from the company's response) for each assessment. The EAG highlighted that potential time saved on collecting routine information could have various effects, including:
allowing the clinical assessor to:
review the distilled information when preparing for the assessment
highlight particular areas for discussion
free up appointment time for more personalised and tailored conversations
make a more accurate and higher-quality clinical assessment
reducing the administrative burden and decreasing the need to recollect information
reallocating saved administrative time by non-clinical staff to other tasks, potentially reducing waiting times.
The clinical experts explained that the time saved on collecting routine information in advance could enable healthcare professionals to spend more time having quality conversations with people. It would not be used to reduce face-to-face time spent with the person having the assessment. This could mean that they would be able to better identify problems, generate hope and start engaging people in treatment. The committee recognised that evidence on time saved on collecting routine information was limited, and the net time saved remains uncertain. But it noted substantial potential for improving system efficiency.
The patient experts shared their experience of accessing NHS Talking Therapies and the impact of common mental health conditions. They highlighted that current services struggle to meet demand, leading to long waiting times for assessment and treatment. Delayed access can negatively affect relationships, employment, and mental wellbeing (for example, increasing suicidal thoughts) and reduce daily functioning. People using the service thought that digital front door technologies could offer benefits such as faster access, improved identification of appropriate treatment pathways and reduced duplication of information sharing. But they were concerned about digital exclusion, literacy barriers and the affordability of technology, which may widen inequalities, particularly for vulnerable groups. Additional concerns included:
risk to data privacy
the risk of a technology making implicit judgements about a person's mental health condition based on their responses
referrals or treatment pathways potentially being influenced without a formal clinical assessment
potential misdirection to inappropriate services
a lack of accuracy of AI‑driven conclusions
whether feedback from people using the service would be adequately considered
whether these technologies might unintentionally exacerbate inequalities.
The committee emphasised the need for safeguards to ensure these technologies provide appropriate support to people using the service.
NICE developed a questionnaire for people who have used the NHS Talking Therapies service. This was to get public responses on using a digital technology before clinical assessments for this service. NICE received a total of 433 responses to the questionnaire. The responses highlighted potential benefits and concerns. Most of the respondents were 25 to 59 years (73%). Women comprised 74% of the sample. Willingness to use digital technologies occurred in 82% of respondents. Reasons given included efficiency, convenience, flexibility and a sense of control over the information provided. But concerns were raised about data privacy, risk of digital exclusion because of literacy and accessibility issues, and the impersonal nature of digital interactions. Also, 64% of respondents waited less than 3 months for an assessment. When asked about routine information collection, 82% stated they were asked to provide details in advance, primarily through phone calls or digital platforms. Some respondents valued the structured approach of the digital technologies. Others expressed difficulty in talking about symptoms digitally, and said they preferred in-person interactions with healthcare professionals. The clinical experts noted that using digital front door technologies could reduce the need for phone calls to collect routine information, potentially saving time in the referral process. But the committee suggested that alternative referral methods to access NHS Talking Therapies services should still be available.
The EAG did not find any published economic evidence that met its systematic literature review inclusion criteria for this early value assessment. So, there was insufficient evidence to compare standard referral practice to NHS Talking Therapies with and without digital front door technologies to:
build a conceptual economic model
generate any reliable economic results.
To compare the benefits and costs of standard NHS pre-assessment referral practice to NHS Talking Therapies with and without digital front door technologies, the EAG did:
an exploratory economic analysis
a threshold analysis
a scenario analysis.
Evidence supporting time savings during the initial clinical assessment with these technologies was limited. But the required time savings to offset the licence costs of Limbic Access or Wysa DRA are minimal. If there are no time savings, a very small quality-adjusted life year (QALY) gain of 0.0003 QALYs for each referral would be needed for either technology to be considered cost effective. This would be at a willingness-to-pay threshold of £20,000 for each QALY gained, assuming only licence costs are considered. Also, there was no quantitative data to support claims that digital front door technologies reduce administrative burden. The committee acknowledged the uncertainty in the economic findings because of the limited evidence. But it recognised the potential for these technologies to improve system efficiency.
The committee noted that Limbic Access and Wysa DRA are designed to be accessible. This includes older people, people from minority backgrounds and disabled people. One peer-reviewed study (Habicht et al. 2024) showed that Limbic Access is effective at increasing referrals from people typically underrepresented in mental healthcare. This included people with diverse gender identities and people from ethnic minority backgrounds. The clinical experts said that, since the introduction of digital front door technologies, referrals from ethnic minority groups have increased. They also suggested that digital front door technologies may help reduce access barriers for some historically harder-to-reach populations.
The EAG highlighted that some people may not have access to a computer, smartphone or laptop. Also, people with low motivation or cognitive challenges may disengage from digital platforms before their referral is complete. The committee noted the considerations raised by people using the service and replying to the NICE questionnaire around the risk of digital exclusion because of literacy and accessibility issues. Older adults or people with low digital literacy may face barriers to using digital front door technologies to access NHS Talking Therapies. There may also be issues for people with English as an additional language because translations may not be appropriate or the content may not be culturally relevant. AI‑based chatbots may be unable to interpret information provided by people with English as an additional language, leading to miscommunication. Marginalised populations, including people experiencing domestic violence or housing insecurity, may avoid using digital services because of concerns over confidentiality.
The clinical experts also noted that people whose first language is not English may face barriers when using digital front door technologies. They added that some people with anxiety, depression, or severe mental illness may struggle with engaging in digital interactions, particularly if they prefer face-to-face support. Also, people with severe physical disabilities may need adaptive technologies that are not always integrated into digital front door systems.
The committee agreed that, if NHS Talking Therapies services continue to provide multiple referral methods to access them, people attempting to access them are unlikely to be disadvantaged if digital front door technologies are introduced.
The committee agreed that there were evidence gaps for the technologies assessed in this early value assessment, including:
Quality of the information collected: more evidence is needed to validate the quality of information collected by digital front door technologies and their potential to enhance the quality of clinical assessments in NHS Talking Therapies. Improving the quality of assessments could lead to better and more accurate treatment decisions and more effective care pathways. The impact of these technologies on clinical decision making is unclear. Some outcomes will be collected through NHS Talking Therapies services. But the clinical experts emphasised the need for comparative studies assessing people referred using digital front door technologies compared with people referred through traditional routes. Also, evaluating the sensitivity and specificity of problem descriptors identified through clinical assessments would provide valuable insights into the reliability of these technologies in supporting clinical decision making.
Acceptability: more data is needed on the completion rate of referrals through digital front door technologies. It is also needed on the acceptability of these technologies among healthcare professionals. This is important in evaluating their effectiveness and integration into NHS Talking Therapies services. Understanding both aspects is helpful for assessing the feasibility, usability and overall impact of digital front door technologies in improving access to NHS Talking Therapies.
Resource and system impact: more information is needed on the impact on administrative burden and time saved by using digital front door technology to assess the net time saved. This will be helpful to understand wider system efficiencies.
Reported outcomes from people using the service: more evidence is needed on feedback from people using the service about ease of access, usability, information clarity and relevance, comfort and privacy. This will help the committee understand how these technologies benefit people using the service.
Costs: there was limited information on training, promotion and digital safety assurance costs of digital front door technologies. The clinical experts highlighted that digital safety assurance costs can be significant. Additional cost data would help the committee better assess the overall value of these technologies.