Abstract 134: Hospital-Level Variability in the Use of Anticoagulant Strategies in a Real-World Population of Patients with Acute Myocardial Infarction
Circulation-cardiovascular Quality and Outcomes(2013)
摘要
Background: Anti-coagulants (AC) are a key treatment for reducing morbidity and mortality during an acute myocardial infarction (AMI). Each AC agent has risks and benefits and may be optimal in certain patient subgroups, but the extent to which the choice of AC therapy is dependent on the hospital site, as opposed to patient characteristics alone, is unknown. Methods: We examined the use of unfractionated heparin (UFH), low-molecular weight heparin (LMWH), and bivalirudin within the first 24 hours of hospitalization for AMI among patients from 24 US hospitals (83% academic; mean #beds=598 [range 66-1000]; all with PCI capability) in the TRIUMPH registry from 4/05-12/08. Only patients who underwent coronary angiography during hospitalization and were treated with an AC were included in the analyses. Hierarchical multinomial regression was used to examine the predicted probabilities of receiving UFH, LMWH, or bivalirudin across the 24 sites, adjusting for patient characteristics (including estimated risk of bleeding). Site level variation was further explored with a median odds ratio (MOR), which estimates the average difference in odds ratios of 2 hypothetically identical patients treated at 2 random sites within TRIUMPH, using UFH as the reference treatment. Results: Of the 3682 AMI patients (47% STEMI/53% NSTEMI) who underwent angiography and were treated with an AC, 77.1% were treated with UFH during the first 24 hours of their AMI, 18.2% with LMWH, and 4.8% with bivalirudin. The predicted probabilities of AC use by site for a hypothetically identical patient ranged across hospitals from 45-98% for UFH, 1-51% for LMWH, and 1-11% for bivalirudin (Figure). The MORs (95% CI) were 2.97 (95% CI 2.25-5.21) for LMWH vs. UFH and 2.52 (95% CI 1.92-4.80) for bivalirudin vs. UFH, indicating substantial variability of the use of different ACs across hospitals even after accounting for patient-level differences. Conclusion: In a large, multicenter AMI registry, we found that the use of different AC strategies is highly dependent on the site to which a patient presents, even after accounting for differences in patients’ presenting characteristics and treatment strategies.
更多查看译文
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络