Table 3.
Unadjusted and adjusted associations of depressive symptoms and diabetes distress at baseline with 2 year biomedical outcomes
| Baseline predictor | 2 year HbA1c, % | Any incident macrovascular complication (%) | Any incident microvascular complication (%) | |||||
|---|---|---|---|---|---|---|---|---|
| Unadjusted | Adjusteda | Unadjusted | Partially adjustedb | Fully adjustedc | Unadjusted | Partially adjustedb | Fully adjustedc | |
| Depressive symptoms: PHQ-9 ≥10 | ||||||||
| β (95% CI) or OR (95% CI)d,e | β 0.23 (0.02, 0.45) | β 0.068 (−0.13, 0.26) | OR 2.18 (1.05, 4.55) | OR 2.78 (1.19, 6.49) | OR 2.29 (0.88, 5.95) | OR 1.10 (0.78, 1.54) | OR 1.33 (0.90, 1.96) | OR 1.32 (0.85, 2.05) |
| p value | 0.031 | 0.49 | 0.037 | 0.018 | 0.088 | 0.60 | 0.15 | 0.22 |
| Diabetes distress: PAID ≥40 | ||||||||
| β (95% CI) or OR (95% CI)d,e | β 0.23 (0.06, 0.53) | β 0.032 (−0.24, 0.30) | OR 1.19 (0.36, 3.95) | OR 2.66 (0.73, 9.69) | OR 1.09 (0.14, 8.82) | OR 1.13 (0.72, 1.79) | OR 1.28 (0.76, 2.16) | OR 1.45 (0.79, 2.67) |
| p value | 0.12 | 0.82 | 0.78 | 0.14 | 0.94 | 0.59 | 0.36 | 0.23 |
| Total PHQ-9 score | ||||||||
| β (95% CI) or OR (95% CI)d,e | β 0.012 (0.05, 0.20) | β 0.02 (−0.05, 0.09) | OR 1.39 (1.00, 1.94) | OR 1.62 (1.12, 2.35) | OR 1.49 (0.99, 2.21) | OR 1.03 (0.91, 1.16) | OR 1.08 (0.93, 1.24) | OR 1.08 (0.92, 1.27) |
| p value | 0.001 | 0.63 | 0.05 | 0.010 | 0.051 | 0.64 | 0.31 | 0.34 |
| Total PAID score | ||||||||
| β (95% CI) or OR (95% CI) | β 0.11 (0.06, 0.17) | β 0.02 (−0.04, 0.08) | OR 0.98 (0.75, 1.26) | OR 1.12 (0.83, 1.51) | OR 0.97 (0.70, 1.34) | OR 1.04 (0.94, 1.14) | OR 1.09 (0.97, 1.22) | OR 1.08 (0.95, 1.23) |
| p value | <0.001 | 0.51 | 0.85 | 0.47 | 0.87 | 0.45 | 0.14 | 0.24 |
| Total number in model | 1234 | 1147 | 1247 | 1049 | 848 | 1334 | 1118 | 884 |
aMultivariable linear regression model (outcome: 2 year HbA1c) adjusted for baseline HbA1c, age, sex, non-white ethnicity, prescription of oral hypoglycaemic medication at 2 years and prescription of insulin at 2 years
bMultivariable logistic regression model (outcome: any incident complication during the 2 year follow-up) adjusted for age, sex, non-white ethnicity, baseline BMI, baseline systolic BP, baseline smoking status, baseline serum cholesterol, baseline HbA1c, prescription of oral hypoglycaemic medication at 2 years and prescription of insulin at 2 years
cMultivariable logistic regression model (outcome: any incident complication during the 2 year follow-up) adjusted for age, sex, non-white ethnicity, baseline BMI, baseline systolic BP, baseline smoking status, baseline serum cholesterol, baseline HbA1c, prescription of oral hypoglycaemic medication at 2 years, prescription of insulin at 2 years and baseline IL-1RA concentration (natural log-transformed)