Table 1—

Regression models to assess socioeconomic and medical correlates of depression

Logistic regression model predicting depression status
PredictorsBSEWaldOR (95% CI)
Younger age−0.040.025.39*0.96 (0.93–0.99)
Unemployment1.070.455.75*2.92 (1.22–6.99)
Lack of home ownership1.340.497.31*3.81 (1.45–10.03)
No. of medications0.120.064.56*1.13 (1.01–1.26)
Duration of diabetes−0.030.021.150.98 (0.93–1.02)
BMI0.050.034.43*1.10 (1.00–1.11)
Hosmer-Lemeshow goodness-of-fit testχ2(8, n = 152) = 4.22; P = 0.84ROC = 0.80R2 = 0.23
Multiple regression model to assess correlates of depression severity
PredictorsBβSE
Younger age−0.27−0.270.09
Unemployment4.300.18*2.01
Income−0.64−0.090.67
Lack of home ownership2.070.072.34
Ability to make ends meet−1.79−0.180.99
Exercise−0.83−0.130.49
BMI0.110.080.11
No. of complications−1.07−0.130.73
Duration of diabetes−0.05−0.030.13
No. of medications0.450.16*0.23
Overall model testF = 8.33R2 = 0.43Adjusted R2 = 0.43
  • *

    * P < 0.05,

  • P < 0.0001.