Effect of Multimodal App-Based Interventions on Glycemic Control in Patients With Type 2 Diabetes: Systematic Review and Meta-Analysis

Background: Digital technologies for type 2 diabetes mellitus (T2DM) care hold great potential to improve patients' health in the long term. Only a subset of telemedicine offerings are digital interventions that meet the criteria for prescribable digitale Gesundheitsanwendung (digital health apps; DiGAs) in Germany. Digital treatments further provide vast amounts of patient data that are important to generate evidence. Objective: This systematic review aims to analyze the efficacy of multimodal digital therapies that mainly meet the DiGA criteria for patients with T2DM and to elicit the potential of such therapies. This includes evidence from randomized controlled trials (RCTs) as well as from real-world data. The outcome of interest was a reduction in glycated hemoglobin (hemoglobin A(1c) [HbA(1c)]; long-term blood glucose measurements). Methods: A systematic literature search was conducted in the literature bases PubMed, LIVIVO, and Cochrane, based on the predefined PICO (Population; Intervention; Control; Outcome) scheme. Identified studies were assessed for risk of bias, pragmatism, and overall quality of evidence. Meta-analyses were conducted for between group differences using RCTs only, and for within-group differences using RCTs and non-RCTs, to examine the effect of the interventions on HbA(1c). Results: In total, 795 records were identified, of which 24 were eligible for this systematic review and 23 studies were eligible for the meta-analysis. The results of the meta-analyses showed significant and clinically relevant reductions in HbA(1c) in patients with T2DM. Regarding the between-group difference for HbA(1c) reduction, the pooled effect of the RCTs showed a reduction of -0.36% (95% CI -0.59% to -0.14%; P<.001), favoring app-based interventions. The average mean within-group reduction in HbA(1c) was -0.79 (95% CI -1.02 to -0.55), with no significant difference between RCTs (-0.69, 95% CI -1.13 to -0.24) and non-RCTs (-0.87, 95% CI -1.16 to -0.57; P<.01, differences between RCTs and RCTs P=.44). A pragmatism rating showed that both study types were on average (very) pragmatic, that is, close to usual care. However, the overall quality of evidence was low to very low. Conclusions: This systematic review shows that digital therapies that mainly meet the DiGA criteria can effectively improve HbA(1c) in patients with T2DM. The integration of digital health care into usual care holds great potential and should be considered as a complementary option to usual care in the future.