OBJECTIVE: The aim of this study was to assess and compare the Confusion Assessment Method (CAM) and the Delirium Rating Scale (DRS) in the detection of delirium in the acute stroke setting, when used by a non-psychiatrist doctor. METHODS: Consecutive participants within 4 days of an acute stroke were screened for delirium using the CAM and the DRS. Patients also had a Mini-Mental State Examination at each assessment. Patients were screened weekly for a maximum of 4 weeks. The CAM and DRS were compared against each other with respect to agreement and applicability. RESULTS: Of 110 eligible patients, 82 were recruited over a 7 month period. Delirium developed in 23 patients (28%), 21 of whom developed delirium during week 1. We found high agreement between the CAM and the DRS in the detection of stroke in the acute stroke setting (kappa values 0.97, 0.86, 0.79 and 1 at weeks 1, 2, 3 and 4, respectively). In addition, there was strong correlation between low MMSE scores (MMSE less that 10) and delirium (kappa scores 1.0, 0.82, 0.83 and 1.0 at weeks 1, 2, 3 and 4, respectively). CONCLUSIONS: Delirium is a common complication post-stroke. The CAM is equivalent to the DRS in the acute stroke setting when used by a trained non-psychiatrist. A low MMSE score may have a small benefit in identifying patients that are at risk of having delirium.