Digital Mental Health Check-In Tool Struggles to Take Hold in NHS Rehabilitation Wards

Routine outcome monitoring has long been promoted as one of the most promising low-cost innovations in mental health care. The idea is deceptively simple: instead of waiting for annual reviews or crisis events, patients regularly rate their quality of life and treatment experiences, and clinicians use those ratings to steer care in real time. One of the best-studied versions of this approach is DIALOG, a structured scale in which patients score their satisfaction across key life domains, and its companion intervention DIALOG+, which turns those scores into a structured conversation between patient and clinician aimed at solving concrete problems. A new mixed-methods feasibility trial, published in BMC Psychiatry by researchers at the University of Sheffield, Sheffield Health and Social Care Trust and King’s College London, now offers a sobering look at what happens when this approach is transplanted into one of the most challenging environments in mental health services: psychiatric rehabilitation inpatient wards in the English National Health Service.

The trial, led by Sinead McLernon and colleagues, set out to answer a question that sounds procedural but carries real clinical weight: is it actually possible to embed DIALOG and DIALOG+ in inpatient rehabilitation settings, where patients often have complex and enduring mental health difficulties and where care is delivered by multidisciplinary teams under constant operational pressure? The study was designed not as a test of effectiveness but as a feasibility and acceptability trial, the crucial first step in the pipeline of implementation science. Before a large-scale evaluation can be justified, researchers must establish whether the intervention can be delivered as intended, whether patients and staff will engage with it, and what practical barriers stand in the way. To that end, the team registered their protocol prospectively on ClinicalTrials.gov in May 2024 and received ethical approval from the Health Research Authority in July 2024, following the UK Policy Framework for Health and Social Care research with informed consent from all participants.

Methodologically, the study used a cluster allocation design, meaning that whole wards rather than individual patients were assigned to the implementation conditions. Feasibility data were collected over a five-month intervention period using a deliberately adapted, paper-based format of DIALOG and DIALOG+, a pragmatic choice reflecting the realities of inpatient wards where access to tablets or shared digital devices cannot be assumed. The researchers set pre-established progression criteria, thresholds that would determine whether the study could reasonably advance to a full trial. Staff recruitment and training proved achievable: seventeen staff members were recruited and trained, demonstrating that the service was willing to engage with the research at the organizational level. The patient side of the equation, however, told a different story.

Patient recruitment fell short of the pre-defined progression criteria, and engagement once enrolled was strikingly low. Across the five-month period, the mean number of DIALOG+ sessions completed per patient was just 1.9 out of a planned 10 sessions. Of the eight patients who participated, only four completed more than a single outcome measure. In implementation science terms, these adherence figures signal that the intervention could not be delivered with anything approaching fidelity in this context. The authors are careful to frame these numbers cautiously given the very small sample size, but the pattern is difficult to ignore: a structured monitoring and solution-focused intervention designed to be delivered repeatedly over months was, in practice, delivered once or twice.

To understand why, the researchers turned to qualitative methods. Patient participants completed qualitative surveys about their experiences, while staff participants took part in focus groups and interviews. Thematic analysis of this material identified four overarching themes that shaped whether DIALOG+ could take root. The first concerned staff and stakeholders’ affective attitude towards the feasibility objectives themselves, that is, their emotional responses to what the research was asking of them. The second theme captured affective attitudes towards the intervention, including how staff and patients felt about the DIALOG+ approach on its own merits. The third theme, alignment with current practice, examined whether the tool fitted the existing rhythms and routines of ward life. The fourth and perhaps most revealing theme was the inpatient setting itself, recognizing that the environment imposes constraints that outpatient or community services do not face.

These themes converge on a central insight: the failure was not necessarily one of the intervention’s design but of its integration. The authors identify insufficient adherence and limited integration within multidisciplinary team meetings as key implementation barriers. DIALOG+ is built around the premise that patient-rated outcomes feed directly into care planning conversations, ideally surfacing in the regular reviews where decisions about support are made. When the ratings remain isolated, completed sporadically and discussed outside the team’s main decision-making forums, the mechanism of action is broken. The tool becomes paperwork rather than practice. This finding echoes a broader lesson from implementation research across healthcare: innovations that require changes to team routines need active, sustained support to alter those routines, and simply training individual staff members is rarely enough.

The implications for the wider field of routine outcome monitoring are significant. DIALOG+ has accumulated encouraging evidence in community mental health settings, where structured quality-of-life assessments paired with solution-focused discussions have been associated with improved outcomes. Rehabilitation inpatient wards, however, serve patients with some of the most complex needs in the system, often involving long stays, fluctuating engagement, cognitive and motivational difficulties, and care delivered by shifting combinations of nurses, occupational therapists, psychologists and psychiatrists. The trial suggests that the assumptions about patient throughput and staff continuity that work in community settings may simply not transfer. A patient who is acutely unwell or ambivalent about their care may not reliably complete a ten-session monitoring schedule, however well intentioned the design.

Importantly, the authors do not conclude that DIALOG+ is unsuitable for this population. Instead, they offer concrete directions for future implementation trials: providing dedicated implementation support, additional training, and ongoing supervision to facilitate intervention fidelity and integration into routine multidisciplinary team practice. This recommendation aligns with contemporary implementation frameworks that distinguish between the effort of introducing an intervention and the effort of sustaining it. Dedicated implementation support might include embedding the tool into ward meeting agendas, assigning named staff champions, building completion into care pathways, and providing regular supervision to troubleshoot barriers as they emerge, rather than discovering them at the end of a five-month window.

The study also demonstrates the value of rigorous feasibility methodology itself. By setting progression criteria in advance, registering the trial prospectively, and combining quantitative adherence data with qualitative thematic analysis, the researchers produced a result that is honest and actionable rather than simply disappointing. Many interventions that fail quietly in routine services are never reported, leaving the field to repeat the same mistakes. Publishing a negative feasibility result, with a clear account of the four barrier themes, gives future teams a map of the specific obstacles they must address before investing in a full-scale trial of outcome monitoring in inpatient rehabilitation.

For patients on psychiatric rehabilitation wards, the stakes of getting this right are considerable. Structured monitoring of quality of life has the potential to give voice to people whose preferences are often overlooked in institutional settings, and to convert vague dissatisfaction into specific, solvable problems. The Sheffield-led trial shows that the technology and the clinical logic are portable only up to a point; the surrounding system must carry the intervention, and in this setting it did not. Whether a better-supported, perhaps digitally enabled version of DIALOG+ can succeed where the paper-based approach faltered remains an open question, one that the authors’ recommendations now make possible to test properly. The study, published open access under a Creative Commons licence with support from the University of Sheffield Institutional Open Access Fund, stands as a candid data point in the ongoing effort to make measurement-based mental health care a reality for the patients who need it most.

Subject of Research: Feasibility of implementing the DIALOG+ routine outcome monitoring system in psychiatric rehabilitation inpatient care

Article Title: Is it possible to implement the DIALOG+ routine outcome monitoring system in a psychiatric rehabilitation inpatient setting? A mixed methods feasibility trial in the UK National Health Service

Article References: McLernon, S., Bone, C., Fergusson, A., & Delgadillo, J. (2026). Is it possible to implement the DIALOG+ routine outcome monitoring system in a psychiatric rehabilitation inpatient setting? A mixed methods feasibility trial in the UK National Health Service. BMC Psychiatry. https://doi.org/10.1186/s12888-026-08669-8

Image Credits: AI Generated

DOI: 10.1186/s12888-026-08669-8

Keywords: DIALOG+, routine outcome monitoring, psychiatric rehabilitation, inpatient care, feasibility trial, mixed methods, NHS, implementation science, mental health, quality of life, multidisciplinary team, BMC Psychiatry