Summary: In a 2-centre factorial randomised controlled trial of 273 Chinese adults with type 2 diabetes mellitus (HbA1c at least 7.0%), self-management with the Dnurse smartphone app did not significantly improve the HbA1c attainment rate over six months versus usual care or telephone follow-up (no significant difference across the four groups, reported p approximately 0.79). Systolic blood pressure was significantly lower in the two app-containing arms (p<0.05), self-monitoring compliance did not differ, and app satisfaction scored above 4.5 on each item.
PICO Summary
| Element | Detail |
|---|---|
| Population | 273 smartphone-literate adults (aged at least 18 years) with type 2 diabetes and HbA1c at least 7.0%, recruited from outpatient endocrinology clinics at two tertiary hospitals in Chengdu, China; 2-centre factorial-design superiority RCT; 220 (80.6%) followed to 6 months. |
| Intervention | Dnurse self-management app, alone (G3) or combined with telephone follow-up (G4), over 6 months. Four equal arms by factorial allocation; the two app-containing arms together comprised roughly half of the 273 participants. |
| Comparison | Usual care (G1) and telephone follow-up alone (G2), without app use. |
| Outcome | Primary outcome (HbA1c attainment rate): no significant difference among the four groups (p>0.05; reported p approximately 0.787), although attainment was numerically higher in the two app arms. Secondary outcomes: systolic blood pressure significantly lower in the two app arms (p0.05); each app-satisfaction item scored above 4.5. No effect estimates with 95% confidence intervals, ARR, or NNT were reported in the abstract. |
Dnurse app for T2DM self-management
Factorial RCT · type 2 diabetes · 6 months
The Dnurse app did not significantly improve the HbA1c <7% attainment rate over 6 months (44.8% vs 34.0%, p=0.787). Systolic blood pressure was lower in the app arms, but the primary glycaemic endpoint was negative.
Expert Commentary
This factorial randomised trial is best read as a negative study on its primary endpoint: the Dnurse app did not significantly raise the HbA1c attainment rate over six months, with a reported p value near 0.79 that argues against a true between-group difference rather than mere underpowering. The numerically higher attainment in the app arms is hypothesis-generating only and should not be presented as glycaemic benefit. The significant fall in systolic blood pressure is a secondary signal, vulnerable to multiplicity across several secondary outcomes and to chance, and is not corroborated by any improvement in self-monitoring compliance. High satisfaction scores reflect acceptability, not effectiveness. A central limitation is that the trial was unblinded and open-label, so attention and reporting effects could plausibly drive softer endpoints while leaving the objective primary outcome unmoved; attrition of roughly one in five patients by month six and single-region recruitment further temper generalisability, and no confidence intervals or absolute risk figures were available to gauge precision. Can I use this with my patients? Not yet as a glucose-lowering tool: there is no convincing evidence the app improves HbA1c, so it is reasonable only as an optional engagement or monitoring aid for motivated, smartphone-literate adults who want it. Adequately powered, blinded-outcome trials reporting absolute effects are needed before this app is recommended for metabolic control.
References
Tang H, Tan H, Zhang J, Zhang M, Chen M, Kong L, et al. Factorial design study of self-management using Dnurse App in T2DM patients. Front Endocrinol (Lausanne). 2025;16:1420578. doi:10.3389/fendo.2025.1420578