Background. A growing number of studies have documented the impact of telemedicine on diabetes management. Limited systematic reviews have failed to yield consistent and convincing results due to modest sample sizes, lack of standardized methodology for meta-analysis, and/or inadequate attention to risks of bias.

Objectives.

1) To apply PRISMA guidelines in conducting a meta-analysis study on the impact of telemedicine in the management of diabetes;

2) To examine and interpret the results, both overall and stratified by subgroups;

3) To assess the robustness of findings by looking into the risks of bias within and across studies.

Methods.  A literature search identified randomized controlled trials examining the effect of telemedicine interventions on HbA1c levels in patients with Type 1 or Type 2 diabetes published until August 8, 2014. The search terms used were “telehealth,” “telemedicine,” and “diabetes.”  Thirty-eight studies were retained for analysis. The meta-analysis was performed using the Comprehensive Meta-Analysis v2 software program to compute Hedges’s g as the summary measure of mean change in HbA1c across studies using a random effects model. We examined overall results as well as stratified results by type of diabetes (Type 1 or 2), age group (younger than 40, 40 to younger than 60, 60 and older), duration of study (6 months or less, more than 6 months), and intervention level (remote monitoring or teleconsultation).

Results. Overall, the results of this meta-analysis favor telemedicine intervention compared to usual care (Hedges’s g= -0.44, SE= 0.09, p<0.001). Subgroup analysis reveals that the effect of telemedicine differs by all subgroups. In particular, teleconsultation appears to be a more effective approach than remote monitoring.

Discussion/Conclusion. Telemedicine is generally effective in improving treatment outcomes for diabetes. Despite its technological advantages relative to teleconsultation, the lack of corresponding effectiveness of remote monitoring in improving diabetes treatment outcomes warrants further research.