Table 2.
Ethnic differences in blood sugar control (HbA1c) with ‘energy to spare’ as the mediator, before and after adjusting for selected biopsychosocial covariates
| Variables | Path a (ethnicity → energy) | Path b (energy → HbA1c) | Path c (ethnicity → HbA1c) | Path a*b or indirect effect (ethnicity → energy → HbA1c) | Total effect |
|---|---|---|---|---|---|
| Unadjusted | −0.32 (−0.49, −0.160)*** | −0.92 (−1.29, −0.55)*** | −2.15 (−3.64, −0.67)** | 0.30 (0.13, 0.58)** | −1.83 (−3.32, −0.35)* |
| Adjusted for age (0 to 90), and gender (male = 1, female = 2), employment earnings (receiving = 1, not = 0) | −0.26 (−0.44, −0.09)** | −0.44 (−0.80, −0.07)* | −4.05 (−5.61, −2.49)*** | 0.11 (0.01, 0.32)**** | −3.91 (−5.47, −2.36)*** |
| Adjusted for age (0 to 90), and gender (male = 1, female = 2), employment earnings (receiving = 1, not = 0), BMI score | −0.19 (−0.37, −0.00)* | −0.29 (−0.66, 0.07)*** | −4.14 (−5.72, −2.57)*** | 0.05 (−0.00, 0.21) | −4.05 (−5.61, −2.48)*** |
| Adjusted for age (0 to 90), and gender (male = 1, female = 2), employment earnings (receiving = 1, not = 0), BMI score, high blood pressure (yes = 1, no = 0) | −0.19 (−0.38, −0.01)* | −0.20 (−0.57, 0.15) | −4.05 (−5.61, −2.48)*** | 0.04 (−0.02, 0.17) | −3.97 (−5.53, −2.41)*** |
| Adjusted for age (0 to 90), and gender (male = 1, female = 2), employment earnings (receiving = 1, not = 0), BMI score, high blood pressure (yes = 1, no = 0), diabetes status (had = 1, never had = 0) | −0.21 (−0.39, −0.02)* | −0.04 (−0.35, 0.26) | −2.74 (−4.06, −1.41)*** | 0.00 (−0.05, 0.11) | −2.70 (−4.02, −1.38)*** |
| Adjusted for age (0 to 90), and gender (male = 1, female = 2), employment earnings (receiving = 1, not = 0), BMI score, high blood pressure (yes = 1, no = 0), diabetes status (had = 1, never had = 0), cigarette smoking (number smoked), physical activity (IPAQ ‘active’ = 1, ‘not active’ = 0) | 0.24 (−0.44, −0.04)* | −0.04 (−0.36, 0.28) | −2.27 (−3.72, −0.83)** | 0.01 (−0.07, 0.12) | −2.24 (−3.68, −0.81)** |
The table does not include the direct effect of variable X (ethnicity) on variable Y (HbA1c) unadjusted for variance attributable to the mediator (perceived ‘energy to spare’)
*(p < 0.05), **(p < 0.01), ***(p < 0.001), ****p > 0.05 (indirect effect is significant based on CI’s, albeit conservative Sobel test not significant)