Table 4.
Logistic regression results for predicting length of stay
| Predictor | Sample | Odds Ratio | 95% CI | P Value |
|---|---|---|---|---|
| Initial AM-PAC mobility | All (n=1453) | 0.957 | 0.933-0.982 | <.001 |
| COVID-19 as primary/admitting | ||||
| Yes (n=690) | 0.946 | 0.906-0.988 | .012 | |
| No (n=761) | 0.953 | 0.922-0.985 | .004 | |
| ICU | ||||
| Yes (n=519) | 0.983 | 0.944-1.023 | .389 | |
| No (n=929) | 0.930 | 0.899-0.962 | <.001 | |
| Cohort | ||||
| Cohort (n=847) | 0.958 | 0.926-0.991 | .014 | |
| Cohort (n=606) | 0.952 | 0.914-0.991 | .018 | |
| Initial AM-PAC activity | All (n=1200) | 0.980 | 0.951-1.011 | .209 |
| COVID-19 as primary/admitting (n=566) | 0.978 | 0.930-1.029 | .392 |
NOTE. Results of the main (sample=all) and subanalyses on different subsets of the cohort (eg, only those with COVID-19 as a primary or admitting diagnosis, etc). Additional subanalyses using AM-PAC activity score were not conducted because of the lack of significance of the main analysis. Model P<.001 for all analyses. Nagelkerke R2 ranged from 0.481-0.525.