Background: Reducing the economic and healthcare system burden of venous thromboembolism (VTE) is critical, given the estimated annual costs in the US ranging from $13.5-$27.2 billion (2011 US dollars). The EINSTEIN clinical trials showed rivaroxaban (RIVA) to be as effective as warfarin (WAR) for VTE with shorter hospital lengths of stay. This study was undertaken to compare the real-world total hospitalization costs for patients hospitalized for VTE initiating oral anticoagulation treatment with RIVA versus WAR during the hospital stay.
Methods: Hospital admissions from the MarketScan Hospital Drug Database (HDD) from November 2012 to December 2013 were selected and analyzed for adults hospitalized with a primary diagnosis of deep vein thrombosis (DVT) or pulmonary embolism (PE) and initiated on either RIVA or WAR during their admission. Patients who received both RIVA and WAR during the admission, or who had also received apixaban or dabigatran were excluded. WAR patients were matched 1:1 with RIVA patients using a combination of exact matching on diagnosis and propensity score matching by nearest neighbor without replacement. Total hospitalization costs included room rate, laboratory tests, inpatient procedures, pharmacy costs, and all other inpatient services. Generalized linear modeling was employed to adjust for the effect of potential confounders on total hospitalization costs.
Results: Following matching, the final study cohorts each consisted of 1,223 patients with primary VTE. RIVA and WAR patients were well matched on demographic and clinical characteristics, with no statistically significant differences between the treatment groups. Mean age was 64 years, with 52% female. Length of stay was significantly shorter for RIVA-treated patients compared to WAR-treated patients (RIVA 3.7±3.1 days, WAR 5.2±3.7 days; p<0.001). RIVA-treated patients incurred statistically significant lower mean and median hospitalization costs per admission compared to WAR (RIVA $8,688±$9,927 median $5,666; WAR $9,823±$9,319 median $7,286; p=0.004). These results were confirmed with adjusted modeling which showed the predicted mean hospitalization costs for RIVA patients at $8,387 compared to WAR patients at $10,275 (p<0.001), a difference of $1,888 lower per hospitalization for RIVA patients. Notably, RIVA-treated patients incurred significantly lower mean and median room rate costs compared to WAR-treated patients (RIVA $2,098±$2,490, median $1,531; WAR $3,093±$3,341, median $2,343; p<0.001).
Conclusions: Hospitalization costs for a primary diagnosis of VTE were significantly lower by an adjusted average of $1,888 per admission for patients receiving oral anticoagulation treatment with rivaroxaban compared with those treated with warfarin.