The Effects and Effect Sizes of Real-Time Continuous Glucose Monitoring on Patient-Reported Outcomes: A Secondary Analysis of the HypoDE Study

Background: While real-time continuous glucose monitoring (rtCGM) has proven its efficacy for glycemic control and avoidance of hypoglycemia, evidence on its effects on patient-reported outcomes is still inconclusive. This secondary analysis of the HypoDE study analyzed effect sizes of rtCGM on patient-reported outcomes and compared them with the effect sizes for glycemic outcomes.

Materials and methods: The intervention group using rtCGM (n = 75) and the control group using self-monitored blood glucose measurements (n = 66) in the HypoDE study were compared. Effect sizes for the patient-reported outcome measures were calculated as the standardized between-group difference in baseline-adjusted follow-up scores (Cohen's d).

Results: rtCGM had significant effects with medium effect sizes on satisfaction with the glucose monitoring device (d = 0.50; 95% confidence interval [CI] 0.84-0.17), fear of hypoglycemia (d = 0.32; 95% CI 0.66 to -0.01), and hypoglycemia-related distress (d = 0.41; 95% CI 0.74-0.08). However, nonsignificant effects and rather small effect sizes were found for general diabetes distress (d = 0.21; 95% CI 0.54 to -0.12), hypoglycemia unawareness (d = 0.03; 95% CI 0.37 to -0.30), and self-reported health status (d = 0.27; 95% CI 0.60 to -0.08), while the effect sizes for the hypoglycemia-related glycemic endpoints were large (d > 1.0). Based on these effect sizes for patient-reported outcomes, a much larger sample size would result in detecting significant differences in patient-reported outcomes.

Conclusions: Effect sizes of rtCGM on patient-reported outcomes are substantially smaller than those on glycemic outcomes. Studies on the efficacy of rtCGM previously had a glycemic endpoint with sample size calculations based on the larger effect sizes for glycemic outcomes. Thus, previous studies were neither powered to detect effects on patient-reported outcomes nor was the participant recruitment tailored to specific patient-reported outcomes. To demonstrate whether rtCGM affects patient-reported outcomes, future studies should consider this comparison of effect sizes.