Table 3.
Effect of the intervention on potential mediators and the associations between changes in mediators and changes in weight at 3- and 9-months (using imputation for missing data)
| Hypothesized mediators | Month | Direct effect of intervention on weight | Intervention effect on potential mediators | Association between potential mediators and weight change | Mediated effect | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| C’ | (SE) | P | A | (SE) | P | B | (SE) | P | AB | (SE) | 95 % CI | AB/(C’ + AB) | ||
| Fruit | 3 | −1.12 | 0.34 | <0.01 | 0.21 | 0.09 | 0.02 | −0.30 | 0.24 | 0.24 | −0.06 | 0.07 | −0.27, 0.04 | 3.26 % |
| 9 | −1.38 | 0.55 | 0.01 | −0.10 | 0.39 | 0.82 | 0.00 | 0.10 | −0.23, 0.19 | 0.17 % | ||||
| Vegetables | 3 | −1.12 | 0.34 | <0.01 | 0.48 | 0.12 | <0.01 | −0.70 | 0.18 | <0.01 | −0.34 | 0.14 | −0.70, −0.11 | 18.48 % |
| 9 | −1.38 | 0.55 | 0.01 | −0.79 | 0.28 | <0.01 | −0.39 | 0.20 | −0.92, −0.08 | 17.00 % | ||||
| SSB | 3 | −1.12 | 0.34 | <0.01 | 0.27 | 0.08 | <0.01 | −0.50 | 0.27 | 0.07 | −0.14 | 0.09 | −0.34, −0.04 | 7.61 % |
| 9 | −1.38 | 0.55 | 0.01 | −0.61 | 0.26 | 0.02 | −0.29 | 0.14 | −0.64, −0.07 | 17.37 % | ||||
| Take-out meals | 3 | −1.12 | 0.34 | <0.01 | 0.29 | 0.07 | <0.01 | −0.38 | 0.30 | 0.20 | −0.11 | 0.10 | −0.35, 0.06 | 5.41 % |
| 9 | −1.38 | 0.55 | 0.01 | −0.34 | 0.44 | 0.44 | −0.11 | 0.13 | −0.40, 0.11 | 4.80 % | ||||
| PA MET mins | 3 | −1.12 | 0.34 | <0.01 | 221.45 | 201.8 | 0.27 | 0.00 | 0.00 | 0.20 | −0.03 | 0.06 | −0.22, 0.03 | 1.63 % |
| 9 | −1.38 | 0.55 | 0.01 | 0.00 | 0.00 | 0.43 | 0.04 | 0.05 | −0.02, 0.21 | −1.74 % | ||||
| PA total days | 3 | −1.12 | 0.34 | <0.01 | 0.95 | 0.39 | 0.02 | −0.04 | 0.06 | 0.49 | −0.04 | 0.08 | −0.24, 0.10 | 2.17 % |
| 9 | −1.38 | 0.55 | 0.01 | −0.16 | 0.1 | 0.12 | −0.16 | 0.14 | −0.53, 0.02 | 6.97 % | ||||
Table design adapted from Hollis et al. (2013) [13]; C’ = unstandardised regression coefficient of the intervention predicting change in weight with mediator in the model. (SE standard error); A = unstandardised regression coefficient of the intervention condition predicting hypothesised mediators; B = unstandardised regression coefficient of the hypothesised mediator predicting weight with intervention condition included in the model; AB = product-of-coefficients estimate. (95 % CI = 95 % confidence interval, Bootstrap bias corrected 95 % confidence intervals of the mediated effect); AB/(C’ + AB) = Proportion of intervention weight effect that was mediated