Model | UT estimate | −95% CI | +95% CI | P value* | R2M (R2C) | ICC | AIC | BIC |
Logistic models (ORs) | |||||||
Tests: any ordered? | |||||||
Model I | 0.75 | 0.40 | 1.41 | 0.181t | 14% (18%) | 5% | 498 | 525 |
Model II | 0.72 | 0.38 | 1.36 | 0.161t | 14% (18%) | 5% | 496 | 528 |
Model III | 0.72 | 0.38 | 1.35 | 0.161t | 17% (21%) | 5% | 491 | 530 |
Model IV | 0.74 | 0.40 | 1.36 | 0.171t | 18% (21%) | 4% | 493 | 541 |
Treatments: any ordered? | |||||||
Model I | 0.88 | 0.52 | 1.49 | 0.321t | 6% (7%) | 1% | 515 | 543 |
Model II | 0.87 | 0.52 | 1.45 | 0.301t | 9% | <1% | 508 | 539 |
Model III | 0.90 | 0.54 | 1.52 | 0.351t | 11% | <1% | 506 | 545 |
Model IV | 0.93 | 0.55 | 1.59 | 0.401t | 14% | <1% | 500 | 548 |
Treatments or tests: any ordered? | |||||||
Model I | 0.65 | 0.29 | 1.47 | 0.151t | 11% (22%) | 13% | 451 | 478 |
Model II | 0.64 | 0.29 | 1.43 | 0.141t | 11% (22%) | 12% | 451 | 483 |
Model III | 0.66 | 0.30 | 1.46 | 0.151t | 13% (23%) | 11% | 448 | 488 |
Model IV | 0.71 | 0.35 | 1.45 | 0.161t | 18% (23%) | 7% | 446 | 494 |
Return attendance 30 days after presentation | |||||||
Model I | 0.95 | 0.47 | 1.92 | 0.892t | 4% | <1% | 320 | 348 |
Model II | 0.97 | 0.47 | 2.00 | 0.942t | 9% | <1% | 312 | 344 |
Model III | 0.95 | 0.46 | 1.95 | 0.882t | 13% | <1% | 311 | 351 |
Model IV | 0.97 | 0.46 | 2.04 | 0.942t | 13% | <1% | 314 | 362 |
Admission 30 days postpresentation | |||||||
Model I | 0.77 | 0.28 | 2.11 | 0.612t | 18% | <1% | 191 | 219 |
Model II | 0.79 | 0.28 | 2.26 | 0.662t | 24% | <1% | 185 | 217 |
Model III | 0.80 | 0.28 | 2.30 | 0.672t | 24% | <1% | 189 | 229 |
Model IV | 0.82 | 0.28 | 2.35 | 0.712t | 24% | <1% | 192 | 240 |
24-hour or shorter length of stay | |||||||
Model I | 0.91 | 0.35 | 2.34 | 0.431t | 98% | <1% | 196 | 224 |
Model II | 0.91 | 0.35 | 2.34 | 0.421t | 98% | <1% | 198 | 230 |
Model III | 0.97 | 0.36 | 2.59 | 0.481t | 98% | <1% | 198 | 238 |
Model IV | 0.97 | 0.36 | 2.59 | 0.731t | 98% | <1% | 198 | 245 |
Linear models (unstandardised β)† | |||||||
Test and treatment costs | |||||||
Model I | −0.01 | −0.06 | 0.05 | 0.371t | 8% (12%) | 4% | −162 | −130 |
Model II | −0.01 | −0.06 | 0.04 | 0.361t | 9% (13%) | 3% | −156 | −120 |
Model III | −0.01 | −0.06 | 0.04 | 0.311t | 13% (16%) | 3% | −150 | −106 |
Model IV | −0.01 | −0.06 | 0.04 | 0.321t | 13% (16%) | 4% | −130 | −79 |
+ admission costs | |||||||
Model I | 0.02 | −0.24 | 0.28 | 0.561t | 16% (17%) | 1% | 1122 | 1153 |
Model II | −0.002 | −0.24 | 0.24 | 0.491t | 23% | <1% | 1088 | 1123 |
Model III | −0.005 | −0.24 | 0.23 | 0.481t | 27% | <1% | 1082 | 1126 |
Model IV | −0.002 | −0.23 | 0.23 | 0.481t | 28% | <1% | 1094 | 1145 |
+ reattendance and return admission costs | |||||||
Model I | −0.05 | −0.34 | 0.24 | 0.371t | 17% (19%) | 2% | 1182 | 1213 |
Model II | −0.07 | −0.33 | 0.19 | 0.291t | 25% (26%) | <1% | 1141 | 1177 |
Model III | −0.08 | −0.33 | 0.17 | 0.281t | 27% | <1% | 1146 | 1190 |
Model IV | −0.07 | −0.32 | 0.18 | 0.281t | 28% | <1% | 1156 | 1208 |
R2M=variance explained by fixed effects.
R2C=variance explained by fixed+random effects (where random effects variance minimal R2C=R2M).
Bold AIC/BIC 2=model with the lowest AIC+BIC.
*1t/2t=one/two-tailed tests.
†Note the natural logarithm of all outcomes was used to mitigate skew after a constant of 182 was added to cost outcomes.
AIC, Akaike Information Criteria; BIC, Bayesian Information Criteria; ICC, intraclass correlation coefficients.