Table 2.
Prediction models for response to teduglutide and placebo
| Predictors selected by regression models | Teduglutide
|
Placebo
|
||
|---|---|---|---|---|
| Coefficient | Odds ratio | Coefficient | Odds ratio | |
| Intercept | −1.263 | −0.988 | ||
| Percentage of colon remaining | −0.020 | 0.980 | 0.003 | 1.003 |
| Cause of major intestinal resection (volvulus versus other) | 1.326 | 3.765 | 0.214 | 1.238 |
| Presence of ileocecal valve (yes versus no) | −0.598 | 0.550 | ||
| Age (years) | 0.023 | 1.023 | ||
| Baseline PS volume at randomization (>6 L per week versus ≤6 L per week) | 1.558 | 4.750 | ||
| Time since start of PS dependency (years) | 0.010 | 1.010 | ||
| Model performance | ||||
| C statisticsa | 0.88 | 0.74 | ||
| P-value of Hosmer–Lemeshow testb | 0.15 | 0.26 | ||
Notes:
a
A C statistic of 1 corresponds to perfect prediction; a C statistic of 0.5 corresponds to a prediction rule that does not perform any better than a random prediction.
b
The Hosmer–Lemeshow test is a goodness-of-fit test. P<0.05 indicates a lack of fit.
Abbreviation: PS, parenteral support.