Personal and Clinical Predictors of Poor Metabolic Control in Children with Type 1 Diabetes in Jordan
Table 3
Linear regression analysis for possible predictors of HbA1c levels.
Variables
Univariate linear regression
Multivariate linear regression¥
◊
95% CI
value
◊
95% CI
value
Age
0.097
0.049-0.145
<0.001
0.065
0.014-0.116
0.013
Duration of diabetes
0.095
0.020-0.170
0.013
BMI
0.073
0.022-0.124
0.005
Gender
MaleØ
Female
-0.043∆
-0.0405-0.320
0.817
DKA as presentation
YesØ
No
-0.170∆
-0.0540-0.199
0.36
Counts of carbohydrates
NoØ
Yes
-0.549
-0.925 to -0.173
0.004
-0.503
-0.870 to -0.135
0.008
Number of clinic visits during last year of follow-up
≤4Ø
>4
0.328
-0.034–0.690
0.075
0.381
0.039–0.723
0.029
School grades
0.002
<80Ø
80-90
-0.439∆
-1.116–0.238
0.203
90-100
-1.034∆
-1.697 to -0.371
0.002
-0.353∆
-0.698 to -0.009
0.044
None
-1.002∆
-1.782 to -0.222
0.012
Receive insulin at school
0.007
YesØ
No
-0.855∆
-1.453 to -0.240
0.005
-0.743∆
-1.310 to -0.176
0.010
Children who do not attend school
-0.505∆
-1.059-0.049
0.074
Who decided insulin dose
0.006
MotherØ
Patient himself/herself
0.675
-0.360-1.709
0.200
Others
0.760
0.271-1.250
0.002
0.580
0.075-1.085
0.024
Current insulin regimen
0.432
TIDx,Ø
MDIΣ
-0.057
-0.444-0.330
0.772
Pump
-1.369
-3.458-0.720
0.198
¥The model had a value < 0.001, , and .xTriple dose insulin regimen. ΣMultiple dose insulin regimen. ◊Unstandardized coefficient. ØReference group. ∆Negative values of coefficients represent values of HbA1c lower than those of the reference group.