This research examined the malleability of physicians' knowledge about psychiatric disorders (and their recommended treatments) based on a brief educational intervention. Primary care physicians were randomly assigned to a single-session intervention dealing with selected DSM-IIIR affective and anxiety disorders or to a wait-list control group. Posttest (vs. pretest) measures indicated that experimental physicians showed greater increases than controls in diagnostic accuracy on relevant disorders, specifically, major depression, dysthymic disorder, and agoraphobia with panic attacks. Treatment recommendations were also influenced, specifically, for major depression and panic disorder. Interestingly, treatment recommendations were also influenced for certain somatic disorders not addressed in the intervention, suggesting "halo" effects for these disorders. Overall, the intervention was effective in increasing diagnostic knowledge and altering treatment recommendations. Implications for research on psychiatric interventions among physicians are discussed.