A novel tool to predict youth who will show recommended usage of diabetes technologies

Controversy exists regarding which individuals will benefit most from commencement of diabetes technologies such as continuous subcutaneous insulin infusion (CSII) or continuous glucose monitoring systems (CGMS), such as 'real-time' sensor-augmented pumping (SAP). Because higher usage correlates with haemoglobin A1c (HbA1c) achieved, we aimed to predict future usage of technologies using a questionnaire-based tool.

The tool was distributed to two groups of youth with type 1 diabetes; group A (n = 50; mean age 12 ± 2.5 yr) which subsequently commenced 'real-time' CGMS and group B (n = 47; mean age 13 ± 3 yr) which commenced CSII utilisation.

For the CGMS group, recommended usage was ≥5 days (70%) per week [≥70% = high usage (HU); <70% = low usage (LU)], assessed at 3 months. In the CSII group, HU was quantified as entering ≥5 blood sugars per day to the pump and LU as <5 blood sugars per day, at 6 months from initiation. Binary logistic regression with forward stepwise conditional was used to utilise tool scales and calculate an applied formula.

Of the CGMS group, using gender, baseline HbA1c, and two subscales of the tool generated a formula which predicted both high and low usage with 92% accuracy. Twelve (24%) showed HU vs. 38 who exhibited LU at 3 months. Of the CSII group, 32 (68%) exhibited HU vs. 15 who exhibited LU at 6 months. Four tool items plus gender predicted HU/LU with 95% accuracy.

This pilot study resulted in successful prediction of individuals who will and those who will not go on to show recommended usage of CSII and CGMS.