Behavioral Programs for Type 2 Diabetes Mellitus A Systematic Review and Network Meta-analysis

Background: Behavioral programs may improve outcomes for individuals with type 2 diabetes mellitus, but there is a large diversity of behavioral interventions and uncertainty about how to optimize the effectiveness of these programs. Purpose: To identify factors moderating the effectiveness of behavioral programs for adults with type 2 diabetes. Data Sources: 6 databases (1993 to January 2015), conference proceedings (2011 to 2014), and reference lists. Study Selection: Duplicate screening and selection of 132 randomized, controlled trials evaluating behavioral programs compared with usual care, active controls, or other behavioral programs. Data Extraction: One reviewer extracted and another verified data. Two reviewers independently assessed risk of bias. Data Synthesis: Behavioral programs were grouped on the basis of program content and delivery methods. A Bayesian network meta-analysis showed that most lifestyle and diabetes self-management education and support programs (usually offering >= 11 contact hours) led to clinically important improvements in glycemic control (>= 0.4% reduction in hemoglobin A1c [HbA1c]), whereas most diabetes self-management education programs without added support-especially those offering 10 or fewer contact hours-provided little benefit. Programs with higher effect sizes were more often delivered in person than via technology. Lifestyle programs led to the greatest reductions in body mass index. Reductions in HbA1c seemed to be greater for participants with a baseline HbA1c level of 7.0% or greater, adults younger than 65 years, and minority persons (subgroups with >= 75% nonwhite participants). Limitations: All trials had medium or high risk of bias. Subgroup analyses were indirect, and therefore exploratory. Most outcomes were reported immediately after the interventions. Conclusion: Diabetes self-management education offering 10 or fewer hours of contact with delivery personnel provided little benefit. Behavioral programs seem to benefit persons with suboptimal or poor glycemic control more than those with good control.