Abstract 195: Length of Stay in High Risk Percutaneous Coronary Intervention in US: A 5 year Contemporary Experience
Background: High Risk Percutaneous Coronary Intervention (PCI) is increasingly being performed with the availability of hemodynamic support. The aim of this study was to determine the predictors of length of stay (LOS) for high risk PCI in US. Methods: We explored the Healthcare Cost and Utilization Project’s Nationwide Inpatient Sample (NIS) using the ICD9 procedure code of 36.07 and 36.06 for PCI. NIS is largest all-payer dataset that represents 20% of all US hospitals. We included patients who had PCI from 2005 through 2010 who also underwent Percutaneous Circulatory Assist Device (PCAD) or Intra-aortic Balloon Pump (IABP) placement during the same hospital admission. Severity of comorbidities was defined by Deyo modification of Charlson’s Comorbidity Index (CCI). Hospitals were identified by a unique hospital identification number and hospital volume was determined by calculating the total number of PCI performed by an institution on year to year basis. Complications were based on Patient Safety Indicators (PSI) recognized by Agency for Health Care Research and Quality to monitor in hospital complications. We examined the predictors of LOS by a mixed effects linear regression model including patient demographics, admission characteristics, CCI quartiles with first quartile as a reference, hospital PCI volume quartiles, IABP or PCAD use and periprocedural complications. Hospital ID was incorporated as random effects in the model. Results: A total of 26,300 High Risk PCIs (weighted n = 130,151) were available for analysis. Factors associated with increased LOS were the use of IABP as compared to PCAD (+0.86 days, p=0.03), occurrence of any complication (+4.67 days, P < 0.001), high CCI (+2.5 days for CCI=2 and +4.1 days for CCI≥3, p<0.001 for both), teaching hospital (+0.96 days, p <0.001), presence of myocardial infarction (MI) or shock (+0.55 days, p = 0.002) and highest quartile of hospital PCI volume (+0.86 days, p<0.001). Factors associated with decreased LOS included private insurance (-0.9 days, p < 0.001) and self-pay or no insurance (-0.89 days, p<0.001). Conclusion: In our observational study based on a large database, use of IABP as compared to PCAD, occurrence of complications, CCI, teaching hospital, presence of MI or shock and high PCI volume were associated with increased LOS & having private insurance and self pay or no insurance was associated with decreased LOS.