Table 2. Evaluation of consistency for primary outcome (complete response).
| Side | Direct | Indirect | Difference | P | |||
|---|---|---|---|---|---|---|---|
| Coef | Std. Err. | Coef | Std. Err. | Coef. | Std. Err. | ||
| A E | .7639229 | .1615678 | .6257291 | 130.9475 | .1381938 | 130.9476 | 0.999 |
| B G | .935809 | .1176838 | .1210314 | 57.23982 | .8147776 | 57.23995 | 0.989 |
| B D | .0035546 | .0949245 | 1.873384 | 145.1808 | 1.876939 | 145.1808 | 0.990 |
| B F | .0248976 | .4128513 | 1.194039 | 187.4053 | 1.218937 | 187.4058 | 0.992 |
| C H | .8186631 | .3626814 | 2.037544 | 187.729 | 1.218881 | 187.7295 | 0.995 |
| C E | .6675231 | .2220261 | .64144 | 148.881 | .0260831 | 148.8811 | 1.000 |
| E I | .5017233 | .1302082 | .7175594 | 153.7626 | .2158361 | 153.7627 | 0.999 |
This result is based on a node-splitting model. Direct estimates are always compared with indirect ones. If the P value is > 0.05, the comparison for this node is not inconsistent.
Abbreviations: Coef, regression coefficient; Std. Err., Standard error.