Background: Cognitive impairment is a common complication of multiple sclerosis, even in early stage disease, with significant impacts on life quality and social interaction. However, its detection is highly test-dependent. Objective: To validate a recently described screening tool, the ARCS, for detecting cognitive impairment in a multiple sclerosis population. Methods: The ARCS administers tests of executive function, memory, visual spatial construction and language via an audio device to unsupervised patients who write their responses for later scoring. Some 127 patients with a wide variety of disease course and severity were assessed by ARCS, of whom 87 also completed the Paced Auditory Serial Addition Test (PASAT) and 45 underwent formal (‘gold standard’) neuropsychological testing. Results: Compared with PASAT, we found that the ARCS showed better sensitivity (86% versus 68%) at equivalent specificity (71%) for detection of impairment in any cognitive domain, and superiority in the detection of memory and executive impairments. Acceptance and completion rates for the ARCS were as good or better than for the PASAT. Conclusions: ARCS is sensitive, well-tolerated, easy to administer and facilitates comprehensive cognitive assessment in less than 5 min of clinician time. It has several advantages over the PASAT for detecting cognitive impairment in patients with multiple sclerosis.