To conduct a systematic review focusing on the effectiveness of behavioral programs for type 1 diabetes (T1DM) and identifying factors contributing to program effectiveness for type 2 diabetes (T2DM).
MEDLINE®, Cochrane Central Register of Controlled Trials, Embase®, CINAHL, PsycINFO® (January 1, 1993, to January 2015), and PubMed® (2015); ClinicalTrials.gov, World Health Organization International Clinical Trials Registry Platform, conference proceedings (2011–14); reference lists of relevant studies.
Two reviewers independently assessed studies for fit with predetermined selection criteria and assessed risk of bias. We included prospective controlled studies and randomized controlled trials (RCTs) for T1DM and RCTs for T2DM, evaluating behavioral programs compared with usual care, active controls (e.g., didactic education), or other behavioral programs. One reviewer extracted data, with verification by a second reviewer. For T1DM, we conducted pairwise meta-analysis to assess program effectiveness; subgroup analyses to examine patient variables (e.g., age, race/ethnicity, glycemic control); and metaregressions to assess potential moderators of effectiveness, such as program components (i.e., diabetes self-management education [DSME], DSME plus support, lifestyle), intensity, delivery format, and personnel. For T2DM, we conducted network meta-analysis (incorporating direct and indirect comparisons) to assess potential moderation of program effectiveness, and subgroup analyses to assess the impact of patient variables. Strength of evidence (SOE) for key outcomes in T1DM was assessed to determine our confidence in the results.
The searches identified 47,149 citations, of which we included 34 studies for T1DM and 132 RCTs for T2DM. All trials had a medium or high overall risk of bias.
For T1DM, there was moderate SOE showing greater reductions in percent hemoglobin A1c (HbA1c) levels at 6-month postintervention followup for individuals receiving a behavioral program compared with usual care (0.31) or an active control (0.44); both were statistically significant, and the latter was considered clinically important based on our prespecified threshold of ≥0.4 unit change in percent HbA1c. There was low SOE showing no difference in HbA1c at end of intervention and at 12-month or longer followup. Generic health-related quality of life was no different at end of intervention in comparisons with usual care (moderate SOE). There was either low SOE or insufficient SOE for all other outcomes, including self-management and lifestyle behaviors, body composition, diabetes-specific quality of life, diabetes distress, and complications. From the subgroup analysis for percent HbA1c by age in comparison with usual care, the effect for the adult subgroup appeared to be greater (0.28) than the effect for the youth subgroup (0.00) at end of intervention, although neither result reached statistical significance. In comparisons with active controls, the SOE of the findings for youths and adults was insufficient. Program intensity (duration, contact hours, frequency of contacts) appeared not to influence program effectiveness for T1DM; individual delivery (vs. group) may be beneficial.
For T2DM, relative to usual care, the effect sizes for all minimally intensive (≤10 contact hours) DSME programs were not considered clinically important based on our prespecified threshold of ≥0.4 unit change in percent HbA1c for glycemic control. Programs having greater benefit for HbA1c reduction were more often delivered in person. For body mass index, lifestyle programs (usually combining structured diet and exercise) provided the most benefit. In subgroup analyses, results for reduced HbA1c favored participants with suboptimal baseline glycemic control (≥7% HbA1c), adults <65 years, and minority participants (sample ≥75% nonwhite and/or Hispanic); the findings by race/ethnicity were confounded by poorer baseline glycemic control among minorities.
Behavioral programs for T1DM offer some benefit for glycemic control when followup extends beyond end of intervention up to 6 months. There was no statistically significant difference at end of intervention or followup timepoints longer than 6 months, although our confidence in these findings is low and benefit cannot be ruled out. More evidence is required to determine the effects of behavioral programs for other outcomes, including lifestyle behaviors, body composition, diabetes-specific quality of life, diabetes distress, and complications. For T2DM, our analyses showed limited benefit in glycemic control from DSME programs offering ≤10 hours of contact with delivery personnel and suggested that in-person delivery of behavioral programs is more beneficial than communicating the information with incorporation of technology. Behavioral programs seem to benefit individuals having suboptimal or poor glycemic control more than those with good control. Tailoring programs to ethnic minorities appears to be beneficial.
Pillay J, Chordiya P, Dhakal S, Vandermeer B, Hartling L, Armstrong MJ, Butalia S, Donovan LE, Sigal RJ, Featherstone R, Nuspl M, Dryden DM. Behavioral Programs for Diabetes Mellitus. Evidence Report/Technology Assessment No. 221. (Prepared by the University of Alberta Evidence-based Practice Center under Contract No. 290-2012-00013-I.) AHRQ Publication No. 15-E003-EF. Rockville, MD: Agency for Healthcare Research and Quality; September 2015. www.effectivehealthcare.ahrq.gov/reports/final/cfm. DOI: https://doi.org/10.23970/AHRQEPCERTA221.
Pillay J, Armstrong MJ, Butalia S, Donovan LE, Sigal RJ, Chordiya P, et al. Behavioral Programs for Type 1 Diabetes Mellitus: A Systematic Review and Meta-analysis. Ann Intern Med. [Epub ahead of print 29 September 2015] doi:10.7326/M15-1399
Pillay J, Armstrong MJ, Butalia S, Donovan LE, Sigal RJ, Vandermeer B, et al. Behavioral Programs for Type 2 Diabetes Mellitus: A Systematic Review and Network Meta-analysis for Effect Moderation. Ann Intern Med. [Epub ahead of print 29 September 2015] doi:10.7326/M15-1400