We studied whether ambulatory utilization review (UR) alters how many patients internal medicine residents refer to subspecialists, and whether the effect persists without reinforcement. We compared referral rates of residents from a firm that held UR meetings (intervention firm residents,n=20) with those of residents from a firm that did not (control firm residents,n=21). We then compared referral rates of 17 intervention firm residents while they were participating in UR with their rates after not participating for at least 4 weeks. Intervention firm residents submitted 30% fewer referrals than control firm residents (9% vs 13%,p=.05). However, the effect was short-lived; after 4 weeks without UR, intervention firm resident referral rates were similar to control firm referral rates.