Background: Hyperacute stroke is a time-sensitive emergency for which outcomes improve with faster treatment. When stroke systems are accessed via emergency medical services (EMS), patients are routed to hyperacute stroke centres and are treated faster. But over a third of patients with strokes do not come to the hospital by EMS, and may inadvertently arrive at centres that do not provide acute stroke services. We developed and studied the impact of protocols to quickly identify and move “walk-in” patients from non-hyperacute hospitals to regional stroke centres (RSCs). Methods and Results: Protocols were developed by a multi-disciplinary and multi-institutional working group and implemented across 14 acute hospital sites within the Greater Toronto Area in December of 2012. Key metrics were recorded 18 months pre- and post-implementation. The teams regularly reviewed incident reports of protocol non-adherence and patient flow data. Transports increased by 80% from 103 to 185. The number of patients receiving tissue plasminogen activator (tPA) increased by 68% from 34 to 57. Total EMS transport time decreased 17 minutes (mean time of 54.46 to 37.86 minutes, p<0.0001). Calls responded to within 9 minutes increased from 34 to 59%. Conclusions: A systems-based approach that included a multi-organizational collaboration and consensus-based protocols to move patients from non-hyperacute hospitals to RSCs resulted in more patients receiving hyperacute stroke interventions and improvements in EMS response and transport times. As hyperacute stroke care becomes more centralized and endovascular therapy becomes more broadly implemented, the protocols developed here can be employed by other regions organizing patient flow across systems of stroke care.