e conducted a randomized controlled trial of EpxDiabetes, a novel digital health intervention as an adjunct therapy to reduce HbA1c and fasting blood glucose (FBG) among patients with type 2 diabetes mellitus (T2DM). In addition, we examined the effect of social determinants of health on our system. Methods: Sixty-five (n = 65) patients were randomized at a primary care clinic. Self-reported FBG data were collected by EpxDiabetes automated phone calls or text messages. Only intervention group responses were shared with providers, facilitating follow-up and bidirectional communication. ΔHbA1c and ΔFBG were analyzed after 6 months. Results: There was an absolute HbA1c reduction of 0.69% in the intervention group (95% confidence interval [CI], -1.41 to 0.02) and an absolute reduction of 0.03% in the control group (95% CI, -0.88 to 0.82). For those with baseline HbA1c >8%, HbA1c decreased significantly by 1.17% in the intervention group (95% CI, -1.90 to -0.44), and decreased by 0.02% in the control group (95% CI, -0.99 to 0.94). FBG decreased in the intervention group by 21.6 mg/dL (95% CI, -37.56 to -5.639), and increased 13.0 mg/dL in the control group (95% CI, -47.67 to 73.69). Engagement (proportion responding to ≥25% of texts or calls over 4 weeks) was 58% for the intervention group (95% CI, 0.373-0.627) and 48% for the control group (95% CI, 0.296-0.621). Smoking, number of comorbidities, and response rate were significant predictors of ΔHbA1c. Conclusions: EpxDiabetes helps to reduce HbA1c in patients with uncontrolled T2DM and fosters patient-provider communication; it has definite merit as an adjunct therapy in diabetes management. Future work will focus on improving the acceptability of the system and implementation on a larger scale trial.