Abstract: Youth with Type 1 Diabetes (T1D) Improve Time in Range by Use of a Relational Agent Embedded in a Smart Phone App

Sophie Maccoll, Emily L. DeWit, Priscilla Connell, Britaney Spartz, David D. Williams, Amey Waghmode, Brent Lockee, Mark Clements
 

Background and Aims: Youth with T1D struggle to achieve blood glucose (BG) TIR targets. However, their inclination for technology makes a glucose self-management app an ideal intervention. We aimed to improve TIR using a relational agent smart phone app. The Happy Bob (HB) app uses Dexcom Continuous Glucose Monitor (CGM) data with personified nudges to gamify engagement for users in BG monitoring. 

Methods: We collaborated with HB developers to customize an app for youth ≥12 years with T1D who use Dexcom and a smartphone. We analyzed participants’ CGM data 28-days pre- and post-enrollment and required ≥ 14-days of wear time in both 28-day periods.  

Results: We divided the cohort into two groups: youth with a change in TIR <+5.0% (n=38) vs those with a change ≥+5.0% (n=14). Notably, the ≥5.0% group had lower %TIR values pre-intervention compared to the <5.0% group (Mean±SD: 37.71±11.67 vs 59.47±18.62; p=0.0002). The ≥5.0% group showed an increase in TIR pre- to post-intervention (37.71±11.67 vs 48.25±12.59; p<0.0001). We also found that for 23 individuals with pre-intervention %TIR <50%, there was a significant increase in their %TIR pre- vs post-intervention (35.35±9.41 vs 40.35±12.05; p=0.0114), while for the total cohort of N=52, average %TIR was 53.61±19.53 vs 54.29±17.72 pre- to post-intervention (p=0.5419).  

Conclusions: In this cohort, TIR values did not improve pre- to post-intervention overall but did improve in youth with a low starting TIR. Future work will examine how HB’s impact is amplified with additional app functionality and when bundled into a remote patient monitoring program.