Skip to main content
Journal Club4 min read

The Best App in the World, and No One on the Ward to Use It

Twenty clinicians explain why good mental-health apps never reach patients. The obstacle is almost never the technology. It is whose job it is to introduce the tool, watch the alerts, and answer when something looks wrong — questions no software answers.

Dr. Sven Jungmann

Dr. Sven Jungmann

CEO

Editorial collage of a smartphone with a blank teal screen lying on an empty hospital bedside table, with a single amber accent at the screen's edge.

Between an app that exists and an app a patient actually opens sits a gap that no amount of engineering closes. A new paper puts twenty clinicians in front of that gap and asks them, plainly, what keeps a perfectly good mental-health tool from reaching the young cancer patients who need it. Their answers have almost nothing to do with the software.

The study sets its scene with two sobering numbers it borrows rather than produces: roughly seventeen years for research evidence to reach the bedside, and only about one in five evidence-based interventions ever entering routine care. Against that backdrop, Hong and colleagues, writing in the Journal of Medical Internet Research in February 2026, interviewed twenty providers at a single academic centre — Seattle Children's Cancer and Blood Disorders Clinic — who care for adolescents and young adults with cancer. The group spanned oncologists, nurses, psychologists, social workers, therapists and an occupational therapist. Each sat for one semistructured interview of roughly forty minutes, and the transcripts were read through the Consolidated Framework for Implementation Research (CFIR), a standard map of why health innovations take hold or fail. The authors did not use the whole framework; they pre-selected sixteen of its sixty-seven constructs as relevant and analysed against those.

What the design can and cannot tell you

It is worth being precise about the evidence tier, because qualitative work is routinely either oversold or waved away. This is twenty people at one site, talking. It measures perceptions, not behaviour: no app was deployed, no patient outcome was followed, nothing was counted that a sceptic could plot. What that buys you is the thing a trial cannot — the texture of why adoption stalls, in the words of the people who would have to do the adopting. Read at that tier it is genuinely useful; read as proof that any particular intervention works it is nothing of the kind, and the authors make no such claim.

Where it actually breaks down

Start with the barriers, because that is where the paper earns its keep. Cost surfaced again and again: nearly half the clinicians were uneasy about apps patients must pay for, in a group already carrying the financial weight of cancer. That is an equity problem, not a usability one. Heavy caseloads and a wariness of yet another stream of reminders came up too. But the deepest obstacle was structural. Providers described working in "silos", with no settled answer to a chain of plain questions. Who introduces the app? Who watches whether it is used? Who responds when it flags a worrying symptom? Where those answers are missing, no one owns the tool — and an ownerless tool is not used.

You could have built the best thing in the world. What you need is a champion on the ground, on the ward.

That line, from one provider, is the closest thing the study has to a thesis. And the clinicians were specific about who the champion should be: not whoever holds the most formal authority, but the bedside nurses and social workers who see the patient most often. The mechanism that carries a tool from installed to used is a person embedded in the daily routine of the ward — not a procurement decision, not a launch email.

What clinicians said would help

The facilitators were strikingly consistent across roles, and they mirror the barriers. Clinicians wanted real evidence before recommending anything — one wanted "a product that was vetted by psychologists and other clinicians" with demonstrated efficacy. They wanted patients involved in the design rather than consulted as an afterthought, an interface young patients would actually tolerate, and the tool wired into the record system and devices they already use rather than bolted on beside them. None of this is surprising. All of it is about trust and fit, not features.

Why it matters here

Silos, integration into the record system, unclear roles and the absence of a local champion are properties of health systems, not of Seattle. For any European hospital trying to move a digital psychosocial intervention from pilot to routine, the paper hands over a short, awkward checklist worth answering before the contract is signed. Who is the champion on this ward, by name? Are the roles between nursing, social work and clinical psychology actually settled, or only assumed? Is the tool inside the clinical information system or beside it? And can the patients who need it reach it without paying? Digital health interventions rarely fail because the software is bad. They fail because no one decided whose job it was to make them work.

Source: Hong SJH, Patton M, Barton KS, Palermo TM, Mulholland K, Chow EJ, Lau N. Facilitators and Barriers to Implementing Mobile Mental Health Interventions: Qualitative Study of the Consolidated Framework for Implementation Research in Pediatric Oncology Providers. J Med Internet Res 2026;28:e87533. Funded by the US National Cancer Institute (grant 1K08CA263474). This is a single-centre qualitative interview study of twenty providers' perceptions — strong on the why of adoption, silent by design on whether any given app improves outcomes.

#Journal Club#Implementation Science#Digital Health#mHealth#Qualitative Research

Keep reading

Editorial collage of an older person's wrist with a plain band rendered as a teal arc, faint activity waveforms below, and one amber dot marking a single external validation link.
Journal Club

Wearables and Dementia: A Strong Signal on Thin Validation

Forty-nine studies suggest disturbed sleep and activity shadow cognitive decline by years. Only three tested their model outside the lab that built it. The signal is real; the case that it works as a screening tool is not yet made.

Dr. Sven JungmannCEO
Editorial collage of a recovery-room patient's hand on a bedrail framed by a teal circle, with twenty-eight uneven navy bars behind it and one amber stripe standing apart.
Journal Club

An AUROC of 0.805, Sitting on 97 Percent Heterogeneity

Twenty-eight machine-learning models claim to predict delirium after heart surgery. Pooled, they look clinically useful. Read the validation methods and the heterogeneity, and the single number stops meaning what it appears to.

Dr. Sven JungmannCEO

This analysis comes from the people behind Visite.

Our weekly newsletter on AI in medicine. Every Friday, rigorously checked.

By signing up you agree to receive Grand Rounds by email. Unsubscribe anytime. More in our privacy policy.

Want to see this in your hospital?

30 minutes. Your questions. Our physician-founder shows you the platform personally.

Book a demo

No commitment. No sales pitch. Physician to physician.