Table 3.
Multivariable analysis on the training dataset
| Variable | Odds ratio | 95% confidence interval | p value |
|---|---|---|---|
| Age | 1.00 | 0.99–1.01 | 0.95 |
| Female sex | 0.87 | 0.76–0.99 | 0.039* |
| Decreasing frailty score (JHACG) | 0.86 | 0.71–1.04 | 0.12** |
| Private insurance status | 1.06 | 0.99–1.14 | 0.11** |
| Increasing median income quartile by zip code | 0.99 | 0.93–1.06 | 0.86 |
| Hospital type | 0.91 | 0.70–1.20 | 0.51 |
| Increasing number of diagnoses | 1.05 | 1.04–1.07 | <0.0001* |
| Malnutrition | 1.04 | 0.74–1.45 | 0.81 |
| Elective admission | 0.62 | 0.53–0.72 | <0.0001* |
| Nonroutine discharge | 0.84 | 0.68–1.04 | 0.12** |
| Increasing hospital bed size | 1.01 | 0.92–1.11 | 0.85 |
| Neurological MDCs | 3.83 | 3.33–4.42 | <0.0001* |
MDC, major diagnostic category.
*Statistically significant. All statistically significant variables were included in the final regression model.
**Not statistically significant but found to be a predictive variable upon tuning parameters/hyperparameters using the validation dataset.