There is much controversy about the use of computed tomography (CT) for patients with minor head injury. We aimed to develop a highly sensitive clinical decision rule for use of CT in patients with minor head injuries. We carried out this prospective cohort study in the emergency departments of ten large Canadian hospitals and included consecutive adults who presented with a Glasgow Coma Scale (GCS) score of 13-15 after head injury. We did standardised clinical assessments before the CT scan. The main outcome measures were need for neurological intervention and clinically important brain injury on CT. The 3121 patients had the following characteristics: mean age 38.7 years); GCS scores of 13 (3.5%), 14 (16.7%), 15 (79.8%); 8% had clinically important brain injury; and 1% required neurological intervention. We derived a CT head rule which consists of five high-risk factors (failure to reach GCS of 15 within 2 h, suspected open skull fracture, any sign of basal skull fracture, vomiting >2 episodes, or age >65 years) and two additional medium-risk factors (amnesia before impact >30 min and dangerous mechanism of injury). The high-risk factors were 100% sensitive (95% CI 92-100%) for predicting need for neurological intervention, and would require only 32% of patients to undergo CT. The medium-risk factors were 98.4% sensitive (95% CI 96-99%) and 49.6% specific for predicting clinically important brain injury, and would require only 54% of patients to undergo CT. We have developed the Canadian CT Head Rule, a highly sensitive decision rule for use of CT. This rule has the potential to significantly standardise and improve the emergency management of patients with minor head injury.