Background: Hospital admissions at a tertiary care hospital occurs mostly by two routes: 1) through the Emergency Department (ED), and 2) through Transfer center. Several perceptions exist about transfer patients being more complex, and association with worse outcomes, or similar outcomes as compared to ED admits, and hence need for optimization of transfer decision and process. Unfortunately, very limited data is available comparing the outcomes of these patients. In this retrospective study, we test the hypothesis, if patients admitted to the hospital by Transfer Center experience worse outcomes compared to those admitted from ED when matched for age, gender, Charlson comorbidity and admitting service.
Methods: A retrospective cohort study, to include patients admitted to Mayo Clinic Hospitals, Rochester, MN, between July 2019 to June 2021, with and without Admissions and Transfer center (ATC) involvement was designed. IRB approval was obtained. Demographic variables of age, gender, Charlson comorbidity, admitting service, and outcome variables of hospital length of stay (LOS), in-hospital mortality, 30-day mortality (after discharge), and readmissions (within 30 days of discharge) were retrieved from electronic health record and validated. Patients were categorized into two groups, one with ATC involvement and the other admitted from ED. The two groups were matched for age, gender, Charlson comorbidity and admitting service. Kruskal Wallis test (continuous variable) and Chi Square test (categorical variables) were performed to compare the two groups in univariate analysis. Conditional logistic regression models were utilized to compare the outcome variables of mortality, 30-day mortality, and readmissions with the two groups, and LOS as covariates. A Poisson regression model was utilized to estimate difference in LOS across two groups.
Results: Of a total of 73,707 eligible patients, 38,214 were included for analysis after matching for age, gender, Charlson comorbidity and admitting service with 19,107 in each group. 5,163 in ATC group and 30,330 in ED group were deemed as outliers and excluded from matched analysis. A median age of 63 years, 52.8% male gender, and a mean Charlson comorbidity score of 4.0 were noted in both the groups. Table 1 outlines the comparison of different outcome variables in univariate analysis and Table 2 shows the outcomes of conditional logistic regression analysis. Poisson regression analysis demonstrates the LOS in ATC group is 1.27 times higher than ED group.
Conclusions: Patients admitted at a tertiary care hospital through the Transfer Center were noted to have poor hospital outcomes compared to patients admitted through the ED, when matched for age, gender, Charlson comorbidity and admitting service. Additional studies evaluating the etiologies for these poor outcomes and, identifying appropriate predictor variables need to be conducted to facilitate appropriate intervention and mitigation of these poor outcomes in this population.