Digital Support for Chronic Disease

Diabetes apps

HbA1c is a blood test that reflects average blood sugar over about two to three months. A 2016 systematic review of 14 randomized trials (1,360 people) found that people with type 2 diabetes using an app had an HbA1c about 0.49 percentage points lower than control groups (95% CI 0.30 to 0.68). The authors rated the evidence as moderate quality, noted the effect was larger with health care professional feedback, and said there was not enough data for type 1 diabetes. [1]

A later review of 21 trials (1,550 people) found a 0.57-point reduction for type 2 diabetes and 0.49 for type 1, but with substantial differences between trials. The reduction was 0.24 points with no professional feedback, 0.33 with low-frequency feedback and 1.12 with high-frequency feedback. [2] Type 1 evidence was rated low quality. [2]

What this means: apps tested as one part of care, often with clinician contact, were linked to lower HbA1c. Apps used alone were linked to smaller changes. These trials tested a mix of different products, so they do not show that a specific app on your phone works. HbA1c is also a lab marker; neither review showed fewer complications. Evidence level: promising but limited.

Symptom monitoring with a care team

Another model asks people to report symptoms electronically so the clinic can respond. In the PRO-TECT cluster-randomized trial, 52 oncology practices were assigned to weekly electronic symptom surveys with alerts to the care team, or usual care. Among 1,191 patients with metastatic cancer there was no survival difference (hazard ratio 0.99, 95% CI 0.83 to 1.17). Emergency visits were less frequent (hazard ratio 0.84), and time to worsening of physical function and symptoms was longer. [3]

This was cancer care in a specific health system model, not a general app. The benefit came with a team acting on alerts. [3] See symptom tracking apps and remote patient monitoring.

Many people stop using the tools

A review of 17 studies of app-based interventions for chronic disease found a pooled dropout rate of 43% (95% CI 29 to 57), with very large differences between studies. [4] Reasons included social, demographic and behavioral factors. [4] Trial results therefore describe people who kept using the app. See meaningful outcomes vs app engagement.

Who should be careful

• People who use insulin or other medicines that can cause low blood sugar should not adjust doses from an app's suggestion unless their clinician has set that up.
• Tools that give treatment suggestions may be regulated as medical devices; the FDA says many low-risk self-management tools get enforcement discretion. [5] See regulatory status of digital health products.
• Tracking can add worry for some people. A tool is only as helpful as the follow-up it triggers.

Questions to ask about a chronic-disease digital product

• Which condition and which outcome did the trial measure (symptoms, hospital visits, a lab marker)?
• Was a clinician part of the program?
• How many people dropped out, and were they counted?
• Who paid for the study?