Abstract T P315: Intra-Arterial Treatment: Long-Term Mortality and Effect of Hospitals

Abstract

Objectives: To first describe long-term mortality in patients receiving any IA treatment for stroke compared to patients receiving IV thrombolysis. Second, given the wide variability in delivery of IA treatments, to test the hypotheses that long-term mortality after IA treatment would vary at the hospital level and that high volume hospitals would have lower mortality. Methods: All Medicare beneficiaries with a principal diagnosis of ischemic stroke (ICD-9 433.x1, 434.x1, 436) admitted via emergency department or transfer were identified using MEDPAR files from 2007-2010. IA procedures were identified using physician billing codes (CPT code 37184-6, 37201, 75896) via linked Medicare Carrier files. Long-term mortality was compared using Kaplan-Meier curves for patients receiving IA, IV or combined therapy. To estimate hospital variability in mortality after IA or combined therapy we used a multi-level logistic regression model adjusting for demographics, vascular risk factors and comorbidities with a random hospital-level intercept. Hospital effects were characterized with the intraclass correlation coefficient (ICC). Secondary analysis adjusted for hospital characteristics: IA volume, stroke volume, bed size, academic status, rural/urban. Results: Median survival after IA was 171 days; after one year-the survival curves for IA and IV treated patients are nearly parallel. (Figure 1) Hospitals had almost no effect on long-term mortality (ICC 0.01) and neither hospital volume nor any other hospital characteristics were associated with mortality. Conclusions: Short-term mortality is significant in IA treated Medicare beneficiaries. Surprisingly, in spite of wide variation in center experience, patient selection, thrombolytic techniques, peri- and post-procedural management, hospitals have no effect on mortality after IA treatment.