Table 3.
Node-splitting tests of direct–indirect consistency across the three outcome networks.
| Outcome | Treatment comparison | Direct estimate (SE) |
Indirect estimate (SE) |
Difference, direct-indirect (SE) |
P value for inconsistency |
|---|---|---|---|---|---|
| Outcome 1 | |||||
| AE vs RSE | 0.242 (0.489) | 0.252 (0.557) | −0.010 (0.742) | 0.989 | |
| AE vs CON | 0.759 (0.254) | 1.093 (1.301) | −0.334 (1.326) | 0.801 | |
| YME vs MBE | −0.011 (0.972) | −0.993 (0.885) | 0.983 (1.315) | 0.455 | |
| YME vs RSE | −1.027 (0.961) | −0.353 (0.871) | −0.674 (1.297) | 0.603 | |
| YME vs CON | −0.209 (0.691) | 0.147 (1.319) | −0.356 (1.490) | 0.811 | |
| MBE vs RSE | −0.257 (0.956) | −0.061 (0.544) | −0.196 (1.100) | 0.859 | |
| MBE vs ME | −0.070 (0.952) | −0.745 (0.552) | 0.675 (1.101) | 0.540 | |
| MBE vs CON | 0.346 (0.388) | 1.148 (1.264) | −0.802 (1.324) | 0.545 | |
| RSE vs ME | 0.091 (0.665) | −0.848 (0.555) | 0.940 (0.866) | 0.278 | |
| RSE vs CON | 0.277 (0.350) | 1.559 (0.722) | −1.281 (0.804) | 0.111 | |
| ME vs CON | 0.983 (0.346) | 1.108 (1.210) | −0.125 (1.258) | 0.921 | |
| Outcome 2 | |||||
| AE vs YME | −0.336 (1.066) | 0.633 (0.669) | −0.969 (1.259) | 0.442 | |
| AE vs MBE | 0.043 (1.078) | 1.012 (0.557) | −0.970 (1.214) | 0.424 | |
| AE vs RSE | 1.833 (0.770) | 0.524 (0.708) | 1.310 (1.049) | 0.212 | |
| AE vs ME | 0.480 (0.778) | 0.201 (0.752) | 0.279 (1.083) | 0.796 | |
| AE vs CON | 1.191 (0.394) | 1.506 (1.015) | −0.314 (1.088) | 0.773 | |
| YME vs CON | 0.731 (0.476) | 3.265 (1.967) | −2.534 (2.023) | 0.211 | |
| MBE vs RSE | 0.387 (1.094) | 0.299 (0.647) | 0.087 (1.271) | 0.945 | |
| MBE vs ME | 0.320 (1.073) | −0.770 (0.657) | 1.090 (1.259) | 0.386 | |
| MBE vs CON | 0.254 (0.365) | 2.465 (1.274) | −2.211 (1.325) | 0.095 | |
| RSE vs CON | 0.329 (0.473) | −1.463 (1.250) | 1.792 (1.338) | 0.181 | |
| ME vs CON | 0.931 (0.494) | 0.637 (1.370) | 0.294 (1.461) | 0.840 | |
| Outcome 3 | |||||
| AE vs YME | 1.400 (3.015) | 1.680 (4.184) | −0.279 (5.158) | 0.957 | |
| AE vs RSE | 0.245 (2.330) | 1.642 (3.017) | −1.397 (3.812) | 0.714 | |
| AE vs ME | 1.081 (3.680) | 2.967 (2.889) | −1.886 (4.678) | 0.687 | |
| AE vs CON | 2.540 (1.365) | 2.197 (5.299) | 0.343 (5.472) | 0.950 | |
| YME vs ME | 3.705 (5.163) | −0.650 (3.560) | 4.355 (6.271) | 0.487 | |
| YME vs CON | 0.493 (2.598) | 3.396 (5.502) | −2.903 (6.083) | 0.633 | |
| MBE vs RSE | 0.327 (3.688) | −0.873 (4.724) | 1.199 (5.993) | 0.841 | |
| MBE vs ME | 1.843 (3.690) | 0.334 (5.329) | 1.509 (6.482) | 0.816 | |
| MBE vs CON | 1.025 (3.001) | 4.169 (6.191) | −3.144 (6.880) | 0.648 | |
| RSE vs ME | 1.516 (3.693) | 1.456 (3.245) | 0.060 (4.916) | 0.990 | |
| RSE vs CON | 1.262 (1.834) | 4.500 (4.357) | −3.239 (4.727) | 0.493 | |
| ME vs CON | 0.011 (2.335) | 1.346 (4.771) | −1.335 (5.312) | 0.802 | |