Stroke registries can provide information on evidence-based practices and interventions, which are critical for us to understand how stroke care is delivered and how outcomes are achieved. The Registry of Canadian Stroke Network (RCSN) was initiated in 2001 and has evolved over the past decade. In the first two years, we found it extremely difficult to obtain informed consent from the patient or surrogate which led to selection biases in the registry. Subsequently (2003 onwards), under the new health privacy legislation in Ontario, Canada, the RCSN was granted special status as a "prescribed registry" which allowed us to collect data on all consecutive patients at the regional stroke centres without consent. The stroke data was encrypted and all personal contact information had been removed, therefore we could no longer conduct follow- up interviews. To obtain patient outcomes after discharge, we linked the non-consent-based registry database to population-based administrative databases to obtain information on patient mortality, readmissions, socioeconomic status, medication use and other clinical information of interest. In addition, the registry methodology was modified to include a periodic population-based audit on a sample of all stroke patients from over 150 acute hospitals across the province, in addition to continuous data collection at the 12 registry hospitals in the province. The changes in the data collection methodology developed by the RCSN can be applied to other provinces and countries.