Abstract 18846: Impact of Bleeding on Length of Stay and Total Cost in Atrial Fibrillation Patients

Abstract

Objectives: To determine the impact of hospital acquired bleeding assessed in patients admitted for atrial fibrillation (AF) on length of stay (LOS) and total hospitalization cost.

Methods: The Premier research database was queried to identify adult inpatients discharged between 2008 and 2011 who had a primary diagnosis code of AF and in whom bleeding was not present on admission. Bleeding was identified using ICD-9 diagnosis codes, charge codes for fresh frozen plasma or blood, or charge codes for transfusion of blood or blood components. Patients considered extreme outliers for LOS (greater than 1 standard deviation [SD] above the mean) were excluded from the study. Descriptive statistics including the outcomes of LOS and total hospitalization cost were calculated. In the univariate comparison of patients with and without bleeding, t-tests for continuous variables and chi-squared tests for categorical variables were used (alpha = 0.05). Gamma regression adjusting for demographics, comorbidities and antiplatelet/anticoagulant use were used to produce adjusted estimates for mean LOS and total hospitalization cost among patients with and without bleeding.

Results: There were 143,415 patients, average age of 69.4 (SD 14.5) that met the study criteria. Across the entire population, mean LOS was 3.5 (SD 3.1) days, and mean total cost was $7,498 (SD 8,557). Overall, 3,058 (2.1%) patients had a bleed during their hospitalization. After adjustment, the mean estimated LOS was 6.68 (95% CI 6.52-6.84) days in the bleed group and 3.28 (95% CI 3.27-2.29) days in the no bleed group. The adjusted mean estimated total hospitalization cost was greater in the bleed group $16,621 (95% CI 16,178-17,074) vs. $6,996 (95% CI 6,969-7,023) in the no bleed group.

Conclusions: Occurrence of hospital acquired bleeding in patients with AF is associated with a significant increase in the LOS and total hospitalization cost. In this study, adjustment for demographic and comorbidity differences as well as the use of antiplatelets and anticoagulants did not change the magnitude or direction of these associations, though there may be residual confounding. In this cohort, the effect of bleeding accounted for approximately $30 million dollars in cost and 10,397 hospitalization days.