Table 3.
Multiple regression analyses of association between interhospital transfer and ED length of stay or in-hospital outcomes in unmatched and propensity-matched patients.
| Unadjusted β or OR (95% CI) | Adjusted β or OR (95% CI) | |
|---|---|---|
| Unmatched cohort (IHT, n=1856; non-IHT, n=16295) | ||
| ED LOS (minutes)a | −315.17c (−423.04, −207.30) | −266.22c (−374.19, −158.26) |
| Hospital LOS (days)a | 2.55c (2.02, 3.10) | 1.01c (0.50, 1.54) |
| In-hospital deathb | 1.63c (1.37, 1.93) | 1.23c (1.02, 1.48) |
| Propensity-matched cohort (IHT, n=1856; non-IHT, n=7424) | ||
| ED LOS (minutes)a | −257.90c (−370.98, −144.83) | −174.08c (−286.42, −61.74) |
| Hospital LOS (days)a | 2.19c (1.60, 2.79) | 1.17c (0.61, 1.74) |
| In-hospital deathb | 1.35c (1.12, 1.64) | 1.32c (1.08, 1.59) |
IHT, interhospital transfer; LOS, length of stay; β, regression coefficient (ED LOS and hospital LOS); OR, odds ratio (in-hospital death); CI, confidence interval.
Non-IHT was used as the reference group for analyses.
In the propensity-matched analyses, the IHT and non-IHT groups were matched for sex, age, and other covariates that had significant differences between two groups to achieve a 1:4 balanced cohort. These covariates included hour of ED visit, year of ED visit, TTAS level, and coexisting diseases.
Linear regression with β
: Logistic regression with OR. Adjusted β or OR was obtained after adjustment for potential confounders.
P < 0.05.