The study examined the SALIENT consortium, a UK research programme funded by the Economic and Social Research Council, part of UK Research and Innovation, with additional support from the government’s Evaluation Accelerator Fund, the Department for Environment, Food and Rural Affairs and the Food Standards Agency. SALIENT’s ambition was explicitly translational: rather than conducting research in isolation and hoping policymakers would eventually take notice, the consortium set out to co-design, implement and evaluate public-facing food system interventions with government partners and food system stakeholders, targeting improvements in both health and sustainability while maintaining affordability and commercial viability. In total, forty organisations engaged with the consortium, and eight of those progressed far enough to deliver and evaluate a combined ten interventions.

To understand how and why this engagement process worked for some organisations and faltered for others, the research team, led by Sarah Shaw of the University of Cambridge’s MRC Epidemiology Unit and the University of Southampton, together with colleagues from the London School of Hygiene and Tropical Medicine, the University of Hertfordshire, the University of Liverpool and the University of Oxford, conducted a qualitative evaluation grounded in theory-based evaluation principles. The methodological approach is worth unpacking. Rather than simply asking participants whether they enjoyed the collaboration, the evaluators first constructed a theory of change, a formal description of the causal pathway through which the SALIENT process was expected to produce its intended outcomes. This artefact then served as an analytical scaffold against which the interview data could be interpreted, allowing the team to identify precisely where the assumed mechanisms of change held up and where they broke down.

The evidence base for the evaluation consisted of thirty-eight semi-structured interviews with three categories of participants: government partners, academic researchers and food system representatives. Interview transcripts were subjected to thematic analysis, a systematic qualitative method in which researchers code the data, group codes into candidate themes, and refine those themes iteratively until they capture the patterns running through the corpus. By interpreting the resulting themes in relation to the theory of change, the evaluators could trace not just what participants experienced but why the engagement process unfolded as it did, a distinction that matters enormously for anyone planning to replicate this model of collaborative research.

The first major finding concerns fragmentation. Participants described a food system landscape in which existing health and sustainability actions were scattered, disconnected and often duplicative, shaped by the differing priorities, agendas and constraints of the stakeholders involved. A local authority might be focused on food poverty, a retailer on supply chain resilience, a regulator on safety standards, and a researcher on generating generalisable evidence. None of these agendas is wrong, but their misalignment means that even well-intentioned initiatives can pull in different directions, dissipating resources and momentum. The evaluation suggests that this fragmentation is not merely an administrative nuisance but a structural barrier that any cross-sectoral collaboration must confront from the outset, ideally through explicit governance arrangements that acknowledge and mediate competing interests rather than pretending they do not exist.

Power dynamics emerged as a second, closely related theme. The researchers found that stakeholder participation was strongest when the research aligned with the stakeholder’s core goals and offered tangible benefits, and weakest when it did not. This may sound obvious, but it carries a profound implication for how collaborative research is designed. Organisations do not donate their time, staff and data to academic projects out of altruism; they engage because the collaboration serves their institutional interests, whether that is access to evidence, reputational advantage, regulatory preparedness or practical problem-solving. Where the perceived benefit was diffuse or distant, engagement thinned out. The lesson for future consortia is that the value proposition for each partner must be articulated early and honestly, and that researchers who assume stakeholders will simply rally around the scientific mission are likely to be disappointed.

The third finding is perhaps the most methodologically provocative. Pragmatic co-design processes, in which interventions were shaped jointly with stakeholders to build trust and ensure feasibility, did indeed strengthen relationships and make implementation more realistic. But the evaluation found a hidden cost: interventions designed primarily for practical acceptability sometimes had limited potential to address the critical uncertainties in the evidence base. In other words, the very compromises that made an intervention deliverable in the real world could also blunt its scientific value, leaving fundamental questions about what works and why unanswered. This tension between rigour and relevance is a perennial problem in implementation science, but the SALIENT evaluation documents it with unusual clarity, showing that co-design is not a cost-free good but a balancing act in which the research questions themselves can be diluted.

Compounding this tension, the study found that academic processes were often perceived as misaligned with stakeholders’ practical needs. Timelines built around grant cycles, ethics approvals and publication schedules rarely map neatly onto the operational rhythms of businesses and local authorities, which must respond to market pressures and political timetables. Reporting requirements, evaluation protocols and the language of research governance can feel alien to partners whose expertise lies in logistics, catering or retail. The evaluation suggests that these frictions are not trivial irritations; they shape whether stakeholders remain engaged at all, and they can determine whether the evidence produced is actually usable by the people it is meant to serve.

So what should be done differently? The authors conclude that long-term government direction, stronger accountability and appropriate incentives are needed to foster meaningful collaboration between researchers and food system stakeholders. Short, project-based funding windows, they imply, are poorly suited to the slow work of building trust across sectors. Future partnerships, the study argues, should balance flexibility with rigour, invest in sustained relationships rather than one-off consultations, and establish strong governance, communication and evaluation processes early, before disagreements about priorities and power have a chance to fester. The goal is to generate robust, scalable evidence for meaningful food system change, and the evaluation makes clear that the machinery of collaboration is not a preliminary to that evidence but a determinant of it.

The significance of this work extends well beyond the UK food sector. Governments worldwide are grappling with diet-related disease, food insecurity and the environmental footprint of agriculture, and the prevailing policy fashion favours multi-stakeholder partnerships and co-produced solutions. SALIENT’s evaluation is a reminder that these fashionable formats are neither self-executing nor universally benign. They can build trust and deliver interventions, but they can also reproduce existing inequalities of power, fragment effort and soften scientific ambition. For researchers, the message is to design engagement around stakeholders’ real incentives and to protect the evidential core of their studies. For funders and governments, the message is that meaningful food system transformation requires patient investment in the unglamorous infrastructure of collaboration: governance, accountability and time. The study, which received ethics approval from the University of Cambridge Humanities and Social Sciences Research Ethics Committees and was published open access, offers a rare, honest blueprint for the difficult work ahead.

Subject of Research: Cross-sectoral collaborative research for food system transformation

Article Title: Conducting cross-sectoral collaborative research for food system transformation: learnings from a qualitative evaluation study

Article References: Shaw, S., Huse, O., Savory, B., Cooke, N., Thompson, C., Loopstra, R., Scarborough, P., Ogilvie, D., Cummins, S., White, M., & Adams, J. (2026). Conducting cross-sectoral collaborative research for food system transformation: learnings from a qualitative evaluation study. Health Research Policy and Systems. https://doi.org/10.1186/s12961-026-01540-y

Image Credits: AI Generated

DOI: 10.1186/s12961-026-01540-y

Keywords: food systems, cross-sectoral collaboration, co-design, qualitative evaluation, theory of change, stakeholder engagement, public health, sustainability, research policy, SALIENT consortium, power dynamics, evidence-based policy

The study, conducted by Mie Pyrmer, Ditte Maria Sivertsen and Signe Wegmann Düring, researchers affiliated with the CAG Dual Diagnosis unit of Mental Health Services in the Capital Region of Denmark, the University of Copenhagen and the Department of Research at Central and Western Zealand Hospital, set out to answer a deceptively simple question: what is it actually like to be a patient inside a coordinated treatment model? Coordination schemes look impressive on organizational charts, but the researchers wanted to know whether the promised coherence reaches the people it is meant to serve.

Dual diagnosis, often abbreviated DD in the clinical literature, refers to the co-occurrence of a psychiatric disorder and a substance use disorder in the same person. The combination is clinically demanding for reasons that go beyond simple addition of two conditions. Substance use can mask or mimic psychiatric symptoms, psychiatric illness can drive relapse into drug or alcohol use, and each condition is typically treated by a different arm of the public system. In Denmark, substance use treatment is organized at the municipal level, while outpatient psychiatry belongs to the regions. A single patient may therefore have two treatment teams, two treatment plans, two sets of appointments and, in the worst case, two sets of goals that pull in opposite directions.

SPOR, whose Danish name Sammenhængende Psykiatri Og Rusmiddelbehandling translates as Concurrent Psychiatric and Substance use Treatment, was created in the Capital Region of Denmark precisely to close that gap. The model establishes structured collaboration between municipal substance use centres and regional outpatient psychiatry, including elements familiar from other integrated care frameworks such as network meetings across sectors and designated contact persons who follow the patient through the system. The researchers interviewed eight participants with dual diagnosis who were currently receiving cross-sectoral SPOR interventions, using semi-structured interviews to capture their experiences in their own words.

The central finding is sobering. Participants reported inadequate cross-sector communication, and that shortfall had a corrosive effect on their sense of care. Rather than experiencing the system as coordinated, patients felt they had to take responsibility for coherence themselves, relaying information between clinicians, reminding one service what another had decided, and effectively acting as their own case managers. For people already struggling to balance substance use, psychiatric symptoms and the demands of everyday life, this added burden is more than an administrative nuisance. It places the heaviest organizational work on the shoulders of the least stable participants in the network.

The pinball metaphor captures the mechanics of the problem. When information does not travel between sectors, the patient does. Each referral, each handover and each change of plan sends the person bouncing from one service to another, with momentum supplied by the system rather than by any deliberate clinical strategy. The researchers found that this dynamic undermined trust and made patients feel unsafe and uncared for at precisely the moments when continuity mattered most. In effect, the coordination model existed on paper, but the lived experience of coordination depended on whether communication happened in practice.

Network meetings, a core component of the SPOR model, received a more nuanced verdict. Participants considered these meetings, in which professionals from both the municipal and regional sides convene around a shared patient, to be important in principle. When they functioned well, they offered a visible demonstration that the two halves of the treatment system were aligned. In practice, however, the study found that network meetings were not conducted stringently. They were often postponed, inconsistently held or otherwise not delivered as designed, and as a result they frequently failed to feel helpful to the patients they were intended to serve. The lesson is a familiar one in implementation science: an intervention is only as good as its fidelity. A coordination mechanism that exists intermittently may be worse than none at all, because it teaches patients not to expect follow-through.

Against these systemic failures, one element of the model emerged as decisively positive: the contact person. Participants reported that the individual clinician assigned to follow them had a considerable influence on whether they felt safe and cared for during treatment. It was considered particularly helpful when the contact person was mindful of the participants’ ongoing struggle to balance substance use, psychiatric symptoms and the rest of their lives. In other words, what patients valued most was not primarily a structural feature of SPOR but a relational one, a trusting therapeutic alliance with a professional who understood the whole person rather than a single diagnosis. This finding echoes a large body of research on the therapeutic relationship, which consistently identifies alliance quality as one of the strongest predictors of engagement and outcomes across mental health and addiction treatment.

The authors conclude that good collaborative practices, including effective cross-sector communication, properly conducted network meetings and continuous feedback, alongside a trusting therapeutic alliance with the contact person, are considered meaningful by patients with dual diagnosis. They note that this aligns with existing research emphasizing how difficult it is to achieve continuous collaboration between hospital-based and municipality-based services. The structural divide their study documents is not unique to Denmark; health systems internationally struggle with the seam between addiction services, which are often locally administered and socially oriented, and psychiatric services, which are typically regionally or centrally administered and medically oriented. Patients with concurrent disorders fall exactly into that seam, and the new study suggests that the seam is felt most sharply by the patients themselves.

The methodological approach deserves attention as well. Semi-structured interviews with eight participants cannot quantify how common these experiences are, but qualitative designs of this kind are well suited to revealing how a policy model is actually lived. The study was carried out according to the ethical principles of the Declaration of Helsinki, approved by the Capital Region Data Protection Agency, and conducted with verbal and written informed consent from all participants. The researchers recruited through clinicians and a peer employee at Psychiatric Centre Amager, and they acknowledge the participants who, in their words, had the courage to be vulnerable and share personal experiences. The study received no external funding and the authors declare no competing interests. Published open access in BMC Psychiatry on 23 September 2026, the paper offers health planners a clear, patient-centred checklist for integrated care: communicate across sectors as a default, hold network meetings reliably rather than occasionally, close the feedback loop, and above all invest in contact persons who can see the whole person. When those elements hold together, patients stop feeling like pinballs and start feeling like people in treatment.

Subject of Research: Patient experiences with coordinated cross-sectoral treatment for dual diagnosis in Denmark

Article Title: ‘I feel like a pinball’: patient experiences with a coordinated treatment model in dual diagnosis

Article References: Pyrmer, M., Sivertsen, D. M., & Düring, S. W. (2026). ‘I feel like a pinball’: patient experiences with a coordinated treatment model in dual diagnosis. BMC Psychiatry. https://doi.org/10.1186/s12888-026-08658-x

Image Credits: AI Generated

DOI: 10.1186/s12888-026-08658-x

Keywords: dual diagnosis, substance use disorder, mental health, cross-sectoral collaboration, coordinated treatment, patient experiences, qualitative research, therapeutic alliance, Denmark, psychiatry, network meetings, contact person