Table 3

Multiple linear regression analysis to assess independent baseline factors of insulin-associated weight gain after 12 months of insulin therapy by including interaction terms and change in insulin dose over time

VariableUnstandardized coefficientsSE95% CIP value
Constant−8.6355.023−18.745 to 1.4750.092
PAID score*0.9640.3840.192 to 1.7360.016
Initial (baseline) insulin dose (units/day)0.1650.0600.044 to 0.2850.008
∆ insulin dose**0.0480.0190.009 to 0.0870.017
Age (years)0.1060.082−0.059 to 0.2710.202
(Age × PAID score)−0.0150.006−0.028 to −0.0020.025
  • Linear regression analysis with weight gain after 12 months as dependent variable. Variables were selected using a manual backward selection method. The baseline variables considered for the model were age (years), body fat (%), HbA1c (% or mmol/mol), insulin dose (units/day), PAID score, ∆ insulin dose (change in insulin dose at 12 months compared with baseline), and the interaction terms: (HbA1c × PAID), (HbA1c × body fat), (HbA1c × initial insulin dose), (HbA1c × age), (HbA1c × ∆ insulin dose), (PAID × body fat), (PAID × initial insulin dose), (PAID × age), (PAID × ∆ insulin dose), (initial insulin dose × body fat), (age × body fat), (∆ insulin dose × body fat), (age × initial insulin dose), (initial insulin dose × ∆ insulin dose), and (∆ insulin dose × age). Only the variables that had a P value <0.05 remained in the final model (R2 = 0.43, adjusted R2 = 0.36). 95% CI, for unstandardized coefficients. In the formula, body weight gain = −8.635 + 0.106 × (age) + 0.964 × (PAID score) + −0.015 × (age × PAID score) + 0.165 × (initial insulin dose) + 0.048 × (∆ insulin dose).

  • *Diabetes-related distress was measured by PAID scale.

  • **∆ insulin dose is the change in insulin dose at 12 months compared with baseline.