Comparisons of School-Day Glycemia in Different Settings for Children with Type 1 Diabetes Using Continuous Glucose Monitoring

Table 6

Regression modeling comparing CGM metric by setting.

	Fully-adjusted model β(95% CI)	-value
TAR (%)
In-school	(Ref)
Weekend	−32.00 (−38.92, −25.07)	<0.001
Virtual school	−47.50 (−54.66, −40.34)	<0.001
TIR (%)
In-school	(Ref)
Weekend	31.84 (25.09, 38.60)	<0.001
Virtual school	47.70 (40.77, 54.63)	<0.001
TBR (%)
In-school	(Ref)
Weekend	−0.01 (−1.35, 1.32)	0.985
Virtual school	−0.20 (−1.17, 0.77)	0.691
Mean glucose^† (mg/dL)
In-school	(Ref)
Weekend	−52.60 (−63.79, −41.41)	<0.001
Virtual school	−74.05 (−85.73, −62.36)	<0.001
Standard deviation^† (mg/dL)
In-school	(Ref)
Weekend	−17.98 (−23.55, −12.42)	<0.001
Virtual school	−14.47 (−19.03, −9.92)	<0.001
CV^† (%)
In-school	(Ref)
Weekend	−2.83 (−5.79, 0.13)	0.061
Virtual school	2.45 (−0.04, 4.95)	0.054

TAR, time above range; TIR, time in range; TBR, time below range; SD, standard deviation; CV, coefficient of variation. Fully adjusted model includes all the potential sociodemographic and clinical predictors: hour, hour², interaction term between setting and hour, race, sex, age, duration of diabetes, regimen, body mass index, insulin dose (units/kilogram), hemoglobin A1c, primary insurance, and area deprivation index national percentile. ^†Values are weighted for the number of readings/hour.