Intervention Enhancement Strategies Among Adults With Type 2 Diabetes in a Very Low–Carbohydrate Web-Based Program: Evaluating the Impact With a Randomized Trial

Abstract

Background Adults with type 2 diabetes may experience health benefits, including glycemic control and weight loss, from following a very low–carbohydrate, ketogenic (VLC) diet. However, it is unclear which ancillary strategies may enhance these effects. Objective This pilot study aims to estimate the effect sizes of 3 intervention enhancement strategies (text messages, gifts, and breath vs urine ketone self-monitoring) that may improve outcomes of a 12-month web-based ad libitum VLC diet and lifestyle intervention for adults with type 2 diabetes. The primary intervention also included other components to improve adherence and well-being, including positive affect and mindfulness as well as coaching. Methods Overweight or obese adults (n=44; BMI 25-45 kg/m2) with type 2 diabetes (glycated hemoglobin [HbA1c] ≥6.5%), who had been prescribed either no glucose-lowering medications or metformin alone, participated in a 12-month web-based intervention. Using a 2×2×2 randomized factorial design, we compared 3 enhancement strategies: (1) near-daily text messages about the intervention’s recommended behaviors (texts n=22 vs no texts n=22), (2) mailed gifts of diet-relevant foods and cookbooks (6 rounds of mailed gifts n=21 vs no gifts n=23), and (3) urine- or breath-based ketone self-monitoring (urine n=21 vs breath n=23). We assessed HbA1c and weight at baseline and at 4, 8, and 12 months. We evaluated whether each strategy exerted a differential impact on HbA1c and weight at 12 months against an a priori threshold of Cohen d of 0.5 or greater. Results We retained 73% (32/44) of the participants at 12 months. The intervention, across all conditions, led to improvements in glucose control and reductions in body weight at the 12-month follow-up. In intent-to-treat (ITT) analyses, the mean HbA1c reduction was 1.0% (SD 1.6) and the mean weight reduction was 5.3% (SD 6.0), whereas among study completers, these reductions were 1.2% (SD 1.7) and 6.3% (SD 6.4), respectively, all with a P value of less than .001. In ITT analyses, no enhancement strategy met the effect size threshold. Considering only study completers, 2 strategies showed a differential effect size of at least a d value of 0.5 or greater Conclusions Text messages, gifts of food and cookbooks, and urine-based ketone self-monitoring may potentially enhance the glycemic or weight loss benefits of a web-based VLC diet and lifestyle intervention for individuals with type 2 diabetes. Future research could investigate other enhancement strategies to help create even more effective solutions for the treatment of type 2 diabetes. Trial Registration ClinicalTrials.gov NCT02676648; http://clinicaltrials.gov/ct2/show/NCT02676648