Diabetes apps lower blood sugar but do not get people moving

Type 2 diabetes is managed day to day by the person who has it, far from any clinic, which is exactly the gap digital health tools promise to fill. Apps, text messages, online platforms and remote monitoring can nudge, remind and track between appointments, potentially at lower cost than face-to-face care. Their record on blood glucose has looked promising, but whether they actually change behaviour, above all physical activity, has been inconsistent across trials. This review set out to test both questions across the key clinical endpoints at once.

Following PRISMA guidelines with a registered protocol, the authors searched five databases through February 2025 for randomised controlled trials comparing digital health interventions against usual care in people with type 2 diabetes. Eligible tools included mobile applications, phone calls or SMS, online platforms and remote monitoring. From 9,499 records they analysed 118 trials covering 21,662 participants. The outcomes pooled were HbA1c, fasting blood glucose, post-meal glucose, a standard index of insulin resistance called HOMA-IR, and physical activity, with risk of bias assessed using Cochrane's tool.

The digital tools delivered on glucose: HbA1c fell by 0.32 to 0.54 percentage points versus usual care, fasting glucose dropped by 0.30 to 0.85, and post-meal glucose improved as well. Online platforms produced the largest HbA1c reduction, at 0.54. The behavioural story was different: physical activity did not significantly improve, and neither did insulin resistance measured by HOMA-IR. Costs also favoured the digital groups, averaging $269.31 against $465.37 for usual care. The trials varied considerably in their results, with intervention duration and sample size explaining part of that spread.

The glucose findings confirm and extend a body of work suggesting digital support genuinely helps people with type 2 diabetes manage blood sugar, and with 118 trials pooled this is one of the larger syntheses available. The activity null complicates the popular assumption that apps improve health by making people move more; whatever drove the glucose gains here, it was not a measurable increase in physical activity. The authors call for adaptive designs aimed at sustained behavioural change and for proper long-term cost-effectiveness work, which remain the open questions.

If you or someone close to you manages type 2 diabetes with an app, a text-message service or remote monitoring, this pooled evidence is a reason to expect real movement in glucose numbers, with online platforms looking strongest. It is equally a reason not to expect the technology to make you more active, because across these trials it did not. The lower average costs in the digital groups are a further point in their favour. On this evidence the tools work as management aids rather than as motivation machines.