The effect of an educational program on antipsychotic prescribing was assessed in two Canadian long-term care centers (LTCC). In each center (Center A residents, n = 258 and Center B residents, n = 191, with dementia at program inception), the rate of change in the odds of using antipsychotics in residents was estimated using mixed-effects logistic regression during a 6-month program period and a 4-month postprogram period, with baseline proportions of use estimated during the 6 months prior to the program. Preprogram proportions of antipsychotic use were 41.6% and 46.2%, respectively. Antipsychotic use decreased during the program in both centers: (odds ratio with 95% CI: 0.943 per week [0.921, 0.965] and 0.969 per week [0.944, 0.994], respectively). During the postprogram period, antipsychotic use increased in Center A (1.039 per week [1.007, 1.072]) but decreased progressively in Center B. The study results suggest the need to implement an ongoing educational program in LTCC.