Abstract 2545: Does Participation in a Stroke Registry Improve Tpa Administration Rates?

Background: The purpose of the Georgia Coverdell Acute Stroke Registry (GCASR) is to improve the quality of patient care. GCASR conducts regular quality improvement activities to educate hospital staff and improve systems and processes. Administration of intravenous tissue plasminogen activator (IV tPA) is standard treatment for eligible acute ischemic stroke patients and can dramatically improve outcomes. Purpose: To determine whether GCASR hospitals were more likely to administer tPA to acute ischemic stroke patients than non-GCASR hospitals. Methods: Hospitalization data from acute care hospitals in Georgia was provided by the Georgia Hospital Association for November 2005 through December 2009. Acute ischemic stroke patients receiving tPA were identified using ICD-9 codes (433 and 434), procedure codes (9910), and healthcare common procedure system codes (J2997). A hospital was defined as a GCASR facility if it was actively participating in the registry at the time of patient hospitalization. A generalized estimating equation with robust variance estimation was applied using the SAS GLIMMIX procedure. “Hospital” was treated as a random variable. Relative risks for receiving tPA were estimated and adjusted for demographics, co-morbidities, hospital size, urbanicity, and length of stay. Results: A total of 55,403 patients were admitted with a principal diagnosis of acute ischemic stroke during the study period, and two percent (1,231) received tPA. Three percent of patients (871) seen at registry facilities received tPA, compared to 1.4% (360) of those seen at non-GCASR facilities. Age, gender, race, length of stay, hospital size, and participation in the registry all predicted tPA administration, either at or near significant levels (p-values from <0.0001 to 0.0646). Although IV tPA administration has increased over time in both hospital groups, patients treated at GCASR facilities were more likely to receive tPA after controlling for confounders (OR=1.64; 95% CI: 0.97-2.78), which approached significance (p=0.0646). Approximately 340 fewer people would have received tPA had all study patients been treated at non-GCASR facilities. Conclusions: Although all Georgia hospitals have improved their rate of tPA administration over time, GCASR hospitals maintained a higher rate than non-GCASR hospitals. This may be due in part to the quality improvement activities that registry facilities participate in and the assistance they receive. These results support the stroke registry model as a method of improving stroke patient care and outcomes.