Abstract 289: Hospital Volume of LVAD Procedures Determines Patient Outcome

Background: Left ventricular assist devices (LVADs) provide mechanical circulatory support to patients with end-stage heart failure. The use of these devices in the United States has been increasing since the FDA approved the first device in 1994. There are no published studies that have evaluated the relationship between LVAD procedural volume and hospital mortality, despite large variation across hospitals in the volume of LVAD procedures performed. This study sought to explore whether a correlation exists between hospital and surgeon’s procedural volumes and patient outcomes, and also to identify a critical threshold. Methods: We conducted a retrospective cross-sectional analysis of all patient discharges from UHC member hospitals from January 2008 through June 2012 after an insertion of an LVAD during their hospitalization. Patients were identified from UHC’s Clinical Database/Resource Manager (CDB/RM) on the basis of the principal or secondary International Classification of Diseases Ninth Revision, Clinical Modification ( ICD-9-CM) procedure code 37.66. The primary outcome was all cause mortality. Results: There were 87 hospitals that admitted at least 1 patient for an LVAD procedure during the study period (77.5 percent males, mean age 54.3). The mean length of stay was 42.1 days and a mean total cost of $299,067. We identified variation of in-hospital mortality by hospital LVAD procedure volume quartile. Quartile 1 included hospitals performing 1-9 procedures (38.8% mortality), quartile 2 performed 10-46 procedures (18.1% mortality), quartile 3 performed 55-97 procedures (12.8% mortality), and the fourth quartile performed 107-319 procedures (16.1% mortality) during the study period. Categorical variables were compared with the Chi-Square Test, and continuous variables were compared with t-tests. There was significant variation in the mortality for almost all study variables including age, gender, admission severity of illness, and admission risk of mortality, and the variation persisted by volume quartile. Conclusion: Initial results suggest that there is a correlation between hospital LVAD procedure volume and in-hospital mortality. LVADs are becoming an increasingly common treatment method for patients with end-stage heart failure and are either awaiting transplant or will receive the device as the final method of therapy. Identifying critical volume thresholds could improve outcomes and ultimately improve the efficiency and value of care. Implications: Identifying mortality associated with LVAD procedures at these hospitals will provide patients and physicians with more information when seeking treatment options for heart failure. This study may also have health policy implications for cardiac treatment by implementing guidelines that LVAD hospital and surgeon programs must adhere to.