OBJECTIVE: The objective of this study was to estimate rates of venous thromboembolism (VTE) and inpatient healthcare utilization following total hip replacement (THR) or total knee replacement (TKR) surgery. METHODS: Using a retrospective cohort approach and inpatient data from a large Midwestern integrated healthcare organization, we evaluated all patients who underwent a THR or TKR procedure between January 1, 1998 and September 30, 2000 and who received warfarin or low-molecular-weight heparin (LMWH) prophylaxis against deep vein thrombosis (DVT) during the initial hospitalization. VTE rates (based on ICD-9-CM diagnoses for pulmonary embolism or DVT) during the index stay were estimated along with the associated incremental length of stay (LOS) and total charges. Time to VTE occurrence within 90 days following index hospitalization also was evaluated along with the associated LOS. RESULTS: We identified 1411 patients who underwent THR (N=605) or TKR surgery (N=806). For both procedures, warfarin was used 57% of the time versus 36% for LMWH, while 7% of patients received both. The mean LOS (±SD) for the initial hospitalization was 4.4 ± 2.0 days and mean charges were $22,150 ± $12,250. The rate of VTE during the initial hospitalization was 1.1% (15/1411) leading to an extra 3.7 ± 0.5 days (p<0.01) in hospital and $11,269 ± $3,168 in additional charges. Following discharge, another 14 patients (1%) were readmitted for treatment of VTE with a mean LOS of 6.1 ± 7.8 days. Most of the readmissions for VTE (10 of 14) occurred within 30 days. CONCLUSIONS: We found that approximately 2% of patients receiving DVT prophylaxis in this high-risk population were diagnosed with thromboembolic events, which resulted in a significant health care burden and cost extending at least 90 days beyond surgery. Whether this failure rate represents inadequate anti-coagulation or other risk factors requires further study.