Table 3.
Multivariable logistic regression analysis of decannulation factors
| Factors | Odds ratio (95% CI)a | P valuea |
| Clinician factors | ||
| Principal work facility | ||
| Acute careb | 1.00 | |
| Weaning | 0.79 (0.45, 1.40) | 0.424 |
| Rehabilitation | 0.52 (0.26, 1.04) | 0.063 |
| Chronic care | 0.28 (0.10, 0.77) | 0.014 |
| Patient scenario characteristics | ||
| Level of consciousness, alert versus drowsyb | 4.76 (3.27, 6.90) | <0.001 |
| Cough effectiveness, strong versus weakb | 3.84 (2.66, 5.54) | <0.001 |
| Secretions, scant thin versus moderate thickb | 2.23 (1.56, 3.19) | <0.001 |
| Oxygenation, 95% saturation, FiO2 of 0.30 versus FiO2 of 0.50b | 2.04 (1.45, 2.86) | <0.001 |
aOdds ratios with corresponding 95% confidence intervals (CIs) and P values were calculated by multivariable logistic regression analysis using robust estimates of variance and are for the comparison of clinical scenarios in which decannulation was recommended with those in which it was not. Odds ratios of greater than 1.0 favor clinician recommendation for tracheostomy decannulation. Odds ratios of less than 1.0 favor clinician recommendation against tracheostomy decannulation. bClinical scenarios with this factor served as the reference group. FiO2, fraction of inspired oxygen.