Objective: With the increasing focus on reducing medical errors and the aging of the physician workforce has come growing concern for cognitive impairment among physicians. This study sought to establish and validate an approach to detecting neuropsychological impairment among physicians.Method: The neuropsychological test performance of 30 physicians referred clinically for neuropsychological evaluations was compared to that of 39 normal community-practicing urologists. We derived 9 key variables from the cognitive and motor tests as dependent variables. Impairment among the clinically-referred doctors was operationalized as scoring ≤5th percentile of the community physicians on at least 3, 4, 5, or 6 of the 9 variables. Using this approach, all clinically-referred physicians were classified as either "impaired" or "ambiguous."Results: A cutoff of ≥5 impaired test scores provided the best balance among competing models. Using this criterion, 14 of the clinically-referred doctors (46%) were impaired and 16 remained ambiguous. The impaired physicians: (1) were older, (2) were more often suspected of having a neurodegenerative disorder, and (3) were more likely to have discontinued practicing medicine. These findings serve as initial validation of our methodology.Conclusions: Using conservative criteria derived from normal community physicians, we could identify a substantial subgroup of clinically-referred physicians who are unambiguously neurocognitively impaired. Replication and refinement of our method with larger samples are recommended, as are the development of specialty-specific criteria for impairment.