Table 5.
Model fit statistics.
| C-Statistic | Ability to distinguish |
Pseudo-R2 | AIC | LRT | ||
|---|---|---|---|---|---|---|
| Wound complications (n=2870) | Age | 0.537 | 7.4% | 0.0012 | 1419.243 | 3.4369 |
| Functional status | 0.511 | 2.2% | 0.0008 | 1425.249 | 5.4310 | |
| Preoperative albumin | 0.561 | 12.2% | 0.0034 | 1413.036 | 9.6434 | |
| Procedure score | 0.593 | 18.6% | 0.0109 | 1391.343 | 31.3372 | |
| Combination of age, functional status, albumin | 0.588 | 17.6% | 0.0067 | 1413.329 | 19.3506 | |
| All four measures (age, functional status, albumin, procedure score) | 0.642 | 28.4% | 0.0169 | 1389.839 | 48.8411 | |
| Full model | 0.678 | 35.6% | 0.0252 | 1397.325 | 73.3545 | |
| Severe complications (n=2870) | Age | 0.567 | 13.4% | 0.0038 | 1427.440 | 10.9687 |
| Functional status | 0.533 | 6.6% | 0.0022 | 1417.753 | 20.6553 | |
| Preoperative albumin | 0.651 | 30.2% | 0.0199 | 1380.832 | 57.5768 | |
| Procedure score | 0.648 | 29.6% | 0.0218 | 1375.213 | 63.1950 | |
| Combination of age, functional status, albumin | 0.665 | 33% | 0.0266 | 1373.103 | 77.3050 | |
| All four measures (age, functional status, albumin, procedure score) | 0.720 | 44% | 0.0441 | 1326.896 | 129.5123 | |
| Full model | 0.743 | 48.6% | 0.0532 | 1331.600 | 156.8090 | |
| Any complication (n=2870) | Age | 0.555 | 11% | 0.0046 | 2091.445 | 13.1221 |
| Functional status | 0.520 | 4% | 0.0055 | 2088.809 | 15.7581 | |
| Preoperative albumin | 0.610 | 22% | 0.0173 | 2054.415 | 50.1513 | |
| Procedure score | 0.637 | 27.4% | 0.0320 | 2011.340 | 93.2270 | |
| Combination of age, functional status, albumin | 0.626 | 25.2% | 0.0238 | 2047.416 | 69.1509 | |
| All four measures (age, functional status, albumin, procedure score) | 0.688 | 37.6% | 0.0504 | 1974.189 | 148.3781 | |
| Full model | 0.702 | 40.4% | 0.0580 | 1983.202 | 171.3646 | |
| Prolonged hospitalization (n=2870) | Age | 0.582 | 16.4% | 0.0164 | 3156.133 | 47.4187 |
| Functional status | 0.526 | 5.2% | 0.0161 | 3157.011 | 46.5404 | |
| Preoperative albumin | 0.664 | 32.8% | 0.0718 | 2989.583 | 213.9682 | |
| Procedure score | 0.662 | 32.4% | 0.0808 | 2961.806 | 241.7456 | |
| Combination of age, functional status, albumin | 0.690 | 38% | 0.0911 | 2941.442 | 274.1096 | |
| All four measures (age, functional status, albumin, procedure score) | 0.753 | 50.6% | 0.1540 | 2741.417 | 480.1349 | |
| Full model | 0.779 | 55.8% | 0.1810 | 2680.337 | 573.2142 | |
| Non-routine discharge (n=1959) | Age | 0.712 | 42.4% | 0.0424 | 952.002 | 84.9319 |
| Functional status | 0.549 | 9.8% | 0.0177 | 1001.914 | 35.0197 | |
| Preoperative albumin | 0.666 | 33.2% | 0.0250 | 987.309 | 49.6255 | |
| Procedure score | 0.578 | 15.6% | 0.0060 | 1025.146 | 11.7883 | |
| Combination of age, functional status, albumin | 0.777 | 55.4% | 0.0716 | 903.415 | 145.5188 | |
| All four measures (age, functional status, albumin, procedure score) | 0.785 | 57% | 0.0752 | 901.838 | 153.0959 | |
| Full model | 0.802 | 60.4% | 0.0877 | 893.031 | 179.5974 | |
| Transfusion (n=2319) | Age | 0.556 | 11.2% | 0.0093 | 2943.943 | 21.7242 |
| Functional status | 0.505 | 1% | 0.0021 | 2960.882 | 4.7846 | |
| Preoperative albumin | 0.626 | 25.2% | 0.0513 | 2843.558 | 122.1086 | |
| Procedure score | 0.657 | 31.4% | 0.0838 | 2762.649 | 203.0175 | |
| Combination of age, functional status, albumin | 0.642 | 28.4% | 0.0584 | 2838.119 | 139.5482 | |
| All four measures (age, functional status, albumin, procedure score) | 0.714 | 42.8% | 0.1234 | 2678.186 | 305.4804 | |
| Full model | 0.724 | 44.8% | 0.1372 | 2665.431 | 342.2354 | |
| Non-routine discharge (n=2870) | Age | 0.711 | 42.2% | 0.0295 | 1064.521 | 85.8993 |
| Functional status | 0.545 | 9% | 0.0087 | 1125.304 | 25.1159 | |
| Preoperative albumin | 0.655 | 31% | 0.0156 | 1105.310 | 45.1098 | |
| Procedure score | 0.592 | 18.4% | 0.0059 | 1133.306 | 17.1145 | |
| Combination of age, functional status, albumin | 0.767 | 53.4% | 0.0458 | 1027.978 | 134.4422 | |
| All four measures (age, functional status, albumin, procedure score) | 0.779 | 55.8% | 0.0497 | 1021.978 | 146.4418 | |
| Full model | 0.869 | 73.8% | 0.0980 | 904.377 | 296.0429 | |
| Readmission (n=1959) | Age | 0.519 | 3.8% | 0.0005 | 1307.245 | 1.0269 |
| Functional status | 0.503 | 0.6% | 0.0002 | 1305.856 | 0.4165 | |
| Preoperative albumin | 0.541 | 8.2% | 0.0027 | 1302.939 | 5.3328 | |
| Procedure score | 0.549 | 9.8% | 0.0033 | 1301.791 | 6.4815 | |
| Combination of age, functional status, albumin | 0.550 | 10% | 0.0033 | 1311.807 | 6.4652 | |
| All four measures (age, functional status, albumin, procedure score) | 0.572 | 14.4% | 0.0063 | 1311.936 | 12.3358 | |
| Full model | 0.599 | 19.8% | 0.0110 | 1324.515 | 21.7572 | |
| Reoperation (n=2870) | Age | 0.538 | 7.6% | 0.0007 | 906.815 | 1.9986 |
| Functional status | 0.506 | 1.2% | 0.0003 | 907.995 | 0.8179 | |
| Preoperative albumin | 0.577 | 15.4 | 0.0028 | 900.725 | 8.0886 | |
| Procedure score | 0.574 | 14.8 | 0.0056 | 892.674 | 16.1394 | |
| Combination of age, functional status, albumin | 0.591 | 18.2% | 0.0038 | 909.918 | 10.8956 | |
| All four measures (age, functional status, albumin, procedure score) | 0.641 | 28.2% | 0.0090 | 900.887 | 25.9263 | |
| Full model | 0.691 | 38.2% | 0.0176 | 907.731 | 51.0820 | |
Full model includes: age, race, ASA, BMI, albumin, preoperative functional status, year of diagnosis, and number of cytoreductive procedures. C-statistics measures discriminative power of the model. A value of 0.5 (null model) indicates that the model’s predictions are no better than chance, whereas a value closed to 1 indicates that the model has a good prediction. Ability to distinguish is calculated from c-statistics (=((C-statistics − 0.5)/0.5)*100%). Higher Pseudo-R2 indicates that the model explains more observed variation in the outcome. When only one variable is included in the model, the lower the AIC, the higher the importance of the variable in the model. The LRT compares the null model (no variables) to a model including one variable or one group of variables. Higher LRT indicates greater improvement in fit when the specified group of variables is included.