Background Aboriginal Canadians are considered to be at increased risk of injury. The de facto standard for measuring injury risk factors among Aboriginal Canadians is to compare hospitalisation and mortality against non-Aboriginal Canadians, but this may be too broad an approach for injury prevention and public health if it over-generalises injury risk.

Methods Data from this study are drawn from the 2001–5 British Columbia Trauma Registry and British Columbia Coroner's Service. Observed and expected hospitalisations and mortality rates on reserves were assessed against three different spatial aggregations of non-reserve reference populations. Data analysis was conducted in a geographical information system using a Poisson probability map.

Results A total of 47 (9.6%) of 487 reserves in British Columbia contained at least one person who was hospitalised or died as a result of serious injury during the study period. Of these, two reserve populations represented 20% (n=19) of all injury morbidity events and 30% (n=22) of all mortality events.

Conclusion Evidence from this study suggests that community-based rather than provincial-based injury reporting is less likely to over-generalise the burden of injury among Aboriginal communities. Community-based surveillance enables researchers to identify why severe unintentional and intentional injury continues to burden some communities but not others and avoids the potentially demoralising and stigmatising effects of current surveillance practices.