One in five youth with T1D experience worsening A1c values between quarterly visits. We evaluated the effectiveness of Remote Patient Monitoring (RPM), a direct-to-consumer telehealth intervention offering problem-solving and education to assess glucose patterns for youth predicted by a machine learning model to experience a significant rise in A1c 70-180 days following routine clinical visits.
Patients received care at a tertiary diabetes clinic in the U.S. Midwest. Supervised machine learning was used to develop a random forest-based model to predict 90-day change in A1c. Clinic staff reviewed weekly lists of patients with a predicted 90-day rise in A1c of ≥3mmol/mol. From these lists, 69 patients under 20yrs old with baseline A1c ≥55mmol/mol were enrolled in RPM. Youth received 1-6 brief telehealth sessions with trained interventionists over 90-days before their next routine clinic visit. Families reviewed device data with interventionists during each session and received personalized insulin regimen adjustments and problem-solving support.
Study cohort was 77% white, 8% Hispanic, 54% female, 36% on insulin pump & CGM, median age 14.32yrs (IQR=11.28,16.41), baseline A1c 64mmol/mol (58,73), and follow-up A1c 67mmol/mol (61,79). Sixty-two percent of the 69 RPM patients did not have A1c rise ≥3mmol/mol compared to 53% of the 524 non-RPM patients (p=0.155).
RPM might lead to improved glycemic levels by preventing clinically significant 90-day rise in A1c. Future research should evaluate the efficiency and effectiveness of this intervention, and the ideal dose, in a randomized controlled setting to identify factors associated with intervention efficacy and optimize care.
Data-Driven Science, Telehealth, Type 1 Diabetes