IPM Take
Digital health has spent years asking whether patients will use the technology. This study suggests the better question may be whether health systems are prepared to support them when they do.
Both groups had access to the same diabetes application. What changed was the pathway around it. Patients in the nurse-led model received digital-readiness assessment, structured onboarding, adherence monitoring, personalised feedback and escalation when engagement dropped. After 24 weeks, adherence was substantially higher and HbA1c reductions were greater than with usual care.
That does not prove the pathway caused the improvement. But it points toward a broader implementation lesson: the value of digital therapeutics may depend as much on workforce, workflow and continuity of care as on the software itself.
Executive Summary
A retrospective cohort study published in Frontiers in Public Health compared 120 adults with type 2 diabetes managed through either a structured nurse-led digital therapeutics pathway or usual care at a tertiary hospital in China. Complete 24-week outcome data were available for 113 patients.
Both pathways used the same hospital-adopted diabetes self-management application, which supported glucose logging, medication reminders, secure messaging and remote monitoring. The nurse-led pathway added readiness assessment, education, dashboard-based adherence monitoring, tiered follow-up, feedback and coordination with physicians and technical staff. Medication decisions remained under physician responsibility.
At 24 weeks, digital adherence averaged 78.3% in the nurse-led group versus 54.7% with usual care, a difference of 23.6 percentage points. HbA1c fell by 0.8 percentage points versus 0.3 percentage points, respectively. Patient satisfaction, self-management scores and glycaemic variability also favoured the nurse-supported pathway.
The authors emphasised that the observational design cannot establish causality and called for prospective multicentre evaluation before broader implementation claims are made.
Why it matters
- HTA bodies: Digital-health assessment may need to look beyond the software itself. Outcomes could depend heavily on the service model, professional support and escalation pathway wrapped around the technology.
- Payers: Funding a digital tool without funding the workforce required to sustain engagement may underestimate the real cost of effective implementation. Future reimbursement models may need to distinguish between software access and digitally enabled care.
- Industry / innovation partners: The findings reinforce an uncomfortable truth for digital-health developers: engagement cannot always be engineered through notifications and interface design alone. Integration into clinical workflow may be part of the therapeutic proposition.
Digital therapeutics are often presented as a way to extend diabetes management beyond the clinic. But putting an application on a patient’s phone is only the first step.
A new real-world study suggests that what happens around the technology may matter just as much as the technology itself.
Researchers examined routine-care records from adults with type 2 diabetes followed for 24 weeks. Sixty patients had entered a structured nurse-led digital therapeutics pathway, while 60 received usual care. Complete follow-up data were available for 57 and 56 patients, respectively.
The distinction between the two groups is important.
Patients in usual care were given brief instructions for downloading the same mobile application and a printed manual before continuing routine outpatient follow-up. The nurse-led group received something far more intensive: digital-readiness assessment, individual education, adherence monitoring through a clinical dashboard, regular feedback and escalating support when use began to fall.
Patients with strong engagement received routine follow-up. Those showing declining adherence received more frequent calls and review of glucose records. Patients with low adherence could be escalated to an in-person visit, re-education and physician review. In other words, the digital platform was not operating as a standalone product. It was embedded inside a human care pathway.
The outcomes diverged.
Adherence reached 78.3% in the nurse-led pathway compared with 54.7% in usual care. Mean HbA1c decreased by 0.8 percentage points in the nurse-led group and 0.3 percentage points under usual care. Nearly 79% of patients in the structured pathway achieved an HbA1c reduction of at least 0.5 percentage points, compared with just over 30% in the comparison group.
There was, however, a workforce cost.
The researchers estimated that nurses spent an average of 3.5 hours per week delivering the pathway across their assigned caseload, covering education, assessment, adherence monitoring, feedback and coordination.
That number may prove as important for policymakers as the improvement in HbA1c.
Digital health is often associated with efficiency and scalability. Yet this model worked by adding structured professional attention around the technology, not by removing clinicians from the pathway. Scaling such an approach therefore raises practical questions about nursing capacity, reimbursement, workflow design and which patients need intensive support.
The study cannot answer those questions yet. It was retrospective, conducted at one centre, involved a relatively small population and lacked complete information on factors including medication changes, comorbidities and digital-access barriers. The authors explicitly caution against causal or broad implementation claims.
But the signal is worth watching.
The next generation of diabetes digital care may not be about finding a smarter app.
It may be about building a smarter system around it.
