Table 4.
Vitamin D mediates group differences in pressure pain threshold at the index knee (Model 3) and ipsilateral forearm (Model 4)
| Effect | Coefficient | SE | t | Sig. | 95% CI Bias Corrected |
|---|---|---|---|---|---|
| Model 3 | |||||
| c | 95.88 | 30.12 | 3.18 | 0.0020 | |
| a | 7.39 | 1.85 | 4.00 | 0.0001 | |
| b | 4.00 | 1.70 | 2.36 | 0.0205 | |
| c' | 66.31 | 31.93 | 2.08 | 0.0408 | |
| a × b | 29.10 | 14.84 | d | (3.43 – 60.86) | |
| Model 4 | |||||
| c | 60.21 | 30.26 | 1.99 | 0.0497 | |
| a | 7.39 | 1.85 | 4.00 | 0.0001 | |
| b | 4.95 | 1.68 | 2.96 | 0.0040 | |
| c' | 23.61 | 31.54 | 0.75 | 0.4561 | |
| a × b | 36.34 | 14.27 | d | (12.44 – 67.55) | |
Table 4 shows unstandardized coefficients for the mediated effect of group differences in pressure pain threshold at the index knee and ipsilateral forearm through vitamin D level, controlling for age, sex, BMI, and CES-D. a race to vitamin D; b, direct effect of vitamin D on pressure pain threshold; c, total effect of race on pressure pain threshold; c', direct effect of race on pressure pain threshold, controlling for vitamin D; a × b indirect effect of race on pressure pain threshold through vitamin D.
A p-value for the indirect effect is not provided because such a p-value is contingent upon a normal distribution of the indirect effect. Given that the product of the a and b path coefficients is always positively skewed, interpretation of this p-value can be misleading (40).