Abstract 203: Emergent Brain Imaging for Acute Ischemic Stroke in Veterans Health Administration Hospitals

Objective: Tissue plasminogen activator (tPA) is under-utilized in Veterans Health Administration medical centers (VAMCs); delays in brain imaging may be a significant barrier. Our primary objective was to describe door-to-imaging time (DIT) patterns among veterans with acute ischemic stroke (IS). We identified patient-level predictors of faster imaging times and decomposed variation in DIT attributable to hospital and patient-level factors. Methods: Detailed medical record reviews were done on 5,000 acute IS patients admitted to any VAMC in 2007; this analysis included those with emergent brain imaging (CT/MRI within 6 hours). We used descriptive statistics to report DIT patterns and a series of random-intercept hierarchical linear regression models to identify predictors of DIT and to decompose variation in DIT. Results: Among the 2,681 acute IS patients emergently imaged in a VAMC, median DIT was 67.7 minutes (min) (IQR, 37.1-115.8 min). Among the 83 patients who were eligible for tPA, the median DIT was 45.9 min (IQR, 28.4-72.1 min) and 22% met the DIT<25 min guideline. Arrival from clinic and increased onset-to-arrival time were independently associated with slower DIT, whereas blood pressure on arrival >185/110 mm Hg was associated with faster DIT (Table). In the model without patient-level factors, 7.2% of variation in DIT was attributable to hospital. Adding patient-level predictors to the model explained 18.8% of the variation in DIT, but 6.4% of the variation remained attributable to case-mix-adjusted hospital variation. Despite this clinical substantial hospital variation, the low IS caseload at most hospitals made it impossible to reliably identify high- and low-performing facilities. Conclusion: There remains room for improvement in DIT for VAMC acute IS patients. Variation is attributable to patient and hospital factors, however, low case IS loads at most hospitals prevented reliable discrimination between high and low-performing centers.