Personal Innovativeness Emerges as the Strongest Driver of mHealth Uptake in Diabetes
Among Chinese adults living with diabetes mellitus, willingness to use a mobile health app is shaped less by abstract technology promises than by how people feel about the tool and how readily they embrace new digital habits. In a cross-sectional survey of 445 participants conducted from March to April 2025, investigators tested an extended UTAUT2 framework that added two psychological constructs: personal innovativeness and attitude. The resulting model did more than fit the data; it mapped the route from curiosity and comfort with new technology to actual intention to engage with mHealth for self-management.
Personal innovativeness stood out immediately. It was the primary predictor of attitude, with a standardized path coefficient of β = 0.335. In practical terms, adults who are naturally more open to new tools were more likely to view mHealth favorably, and that favorable view then helped convert interest into intention. This matters in diabetes care because app-based self-management only works when patients return to the platform day after day, not when they download it once and abandon it after a brief trial.
The study also found two pathways with medium effects on behavioral intention: effort expectancy and attitude. Effort expectancy carried a coefficient of β = 0.388 and an f² of 0.260, while attitude had a coefficient of β = 0.324 and an f² of 0.150. Those numbers point to a familiar clinical truth with digital tools: if the interface feels cumbersome, the value of the app shrinks quickly. A diabetes app can offer glucose logs, reminders, or education, but it has to feel simple enough to use during daily routines that already include medication timing, dietary choices, and symptom monitoring.
Why Ease of Use and Belief in the App Still Shape Behavior
Performance expectancy and facilitating conditions remained part of the model, but their direct effects on behavioral intention were small, with f² values below 0.02. In plain clinical terms, usefulness and environmental support still matter, yet they did not carry the same weight as ease of use and attitude in this sample. The pattern suggests that adults with diabetes may first ask whether an app feels manageable and worth their attention before deciding whether it could improve care.
Attitude also acted as a mediator between the predictor variables and behavioral intention. This is the kind of finding that has direct relevance for app design and patient counseling. A person may understand that mHealth can track blood glucose, support adherence, or deliver reminders, but that understanding alone is not enough. The internal response to the technology — confidence, receptivity, and positive expectation — seems to be the bridge between awareness and willingness to use it.
The model explained 29.8% of the variance in attitude and 51.0% of the variance in behavioral intention. For a behavioral adoption study, those values suggest meaningful explanatory power without pretending to settle the question completely. Digital engagement is never determined by one layer of influence. Age, clinical burden, prior technology exposure, trust, literacy, and social context all sit behind the choice to use an app, even when they are not fully captured in a single framework.
What the Extended UTAUT2 Model Adds to Diabetes Self-Management
UTAUT2 has long been used to study technology acceptance, but this analysis extends the model in a way that fits diabetes care particularly well. Self-management requires repeated behavior change, and repeated behavior change depends on more than technical capability. It depends on how a tool is perceived at the bedside and at home, where patients decide whether to open the app, enter data, respond to prompts, or ignore the notification banner.
By incorporating personal innovativeness and attitude, the model captured a more human pathway to adoption. Personal innovativeness influences how quickly someone approaches a new digital tool, while attitude reflects the emotional and cognitive appraisal of that tool. When both are favorable, intention rises. That sequence has obvious implications for clinicians recommending mHealth platforms: a brief recommendation is unlikely to be enough if the patient already sees digital self-management as confusing, intrusive, or low value.
The data also reinforce why usability cannot be an afterthought. Effort expectancy had the largest direct effect on intention, which means the most persuasive app is still the one that asks the least of the user. For diabetes care, where many people are already managing multiple medications, dietary restrictions, and self-monitoring tasks, every extra step in navigation can erode engagement. Simpler workflows, clearer language, and fewer barriers to entry are likely to matter as much as feature lists.
Clinical Reach Beyond the Survey Sample
The study was conducted in Chinese adults with diabetes, and that context is central. Digital behavior is shaped by local norms, platform familiarity, and health system structure. A model that explains use intention in one population may not transfer unchanged to another, especially where smartphone habits, health literacy, and access to clinician support differ. The sample size of 445 offers a useful base for structural equation modeling, but it does not erase the need for validation in other settings.
Still, the findings speak directly to frontline diabetes care. Patients do not adopt mHealth because developers describe its functions well; they adopt it when the tool feels easy, credible, and personally relevant. Confidence in the app’s value has to be matched by confidence in one’s own ability to use it. That alignment between perceived utility and perceived usability may be the most practical target for interventions aimed at improving digital adherence.
For providers, the message is straightforward. Recommending an app is not the same as securing engagement. Counseling may need to address attitude, anticipated effort, and personal comfort with technology, not just the app’s features. Developers face a parallel challenge: build systems that reduce friction, support early success, and make first use feel natural rather than technical.
The remaining gap is not whether mHealth can support diabetes self-management in theory, but how to make adoption durable across different patient groups and care environments. This study establishes a workable behavioral model, yet it remains limited to cross-sectional intention rather than actual long-term use, leaving the durability of those intentions untested.