Background: Regionalization of care to primary stroke centers (PSC) may improve outcomes for stroke patients. We evaluated the current access of Medicare stroke patients to PSC, and its potential impact on mortality. Methods: We performed a retrospective cohort study of a 100% sample of Medicare fee-for-service claims data for patients admitted with stroke in 2008-2009, with one-year follow-up through 2010. Population weighted centroids were created, helicopter pad locations were identified, and driving distances were calculated based on real road network data. Driving and flying speeds, dispatch, scene, and pre-hospital times were estimated using validated models, adjusted for population density. The association of 30-day mortality with travel times, and treatment at a PSC was investigated using multivariable regression models. Results: During the study period, 510,822 patients (mean age 79.6 years, 59.7% females) had a stroke. There was significant regional variation in our cohort, (Figure) with 8.6% of stroke patients having ground access to a PSC within 30 minutes, 14.9% from 30 to 45 minutes, 11.1% from 45 to 60 minutes, 43.9% from 1 to 4.5 hours, and 21.5% over 4.5 hours. The latter group could be limited to 0.1% of stroke patients, if existing helicopter services were used optimally. 164,485 (32.2%) patients received treatment in a PSC, and had modestly decreased mortality (OR, 0.97; 95% CI, 0.95-0.99). For this group, actual travel time to the PSC was not associated with mortality for patients within one hour of the PSC. On the contrary, travel times from 1 to 4.5 hours (OR, 1.15; 95% CI, 1.08-1.22), and over 4.5 hours (OR, 1.41; 95% CI, 1.29-1.54) were associated with increased mortality. Conclusions: There is significant regional variation in access to PSC for elderly stroke patients, with a potential impact on outcomes. Optimal use of helicopter services may address these disparities. Funding: NIH (P01-AG19783, and U01-AG046830-01).