OBJECTIVE: To determine the relationship between ziprasidone starting dose and persistence among patients diagnosed with schizophrenia. METHODS: Adult Medicaid recipients diagnosed with schizophrenia and having ziprasidone prescription claims between July 1, 2001 and September 30, 2003 were categorized by starting dosage: low (20-60mg) n=517; medium (61mg–119mg) n=339; and high (120-160 mg) n=341. Persistence was measured using refill patterns, allowing 15-day gaps between expected refill dates, and compared across starting doses using Chi-Square tests. Multivariate logistic analysis explored the simultaneous impact of age, gender, race, and year of treatment initiation in addition to starting dose. RESULTS: Discontinuation rates across the study period (maximum 30 months) were greater for patients initiated with low (p=0.001) and medium dose (p=0.02) than for high dose patients. Discontinuation rates were not statistically different for low and medium doses. Discontinuation rates at 365, 180, and 90 days were higher for low dose than high dose (p<0.05) but not significantly different between low and medium or medium and high doses. These results were similar in the multivariate models. CONCLUSIONS: Schizophrenia patients started on high doses of ziprasidone have greater persistence up to 2 ½ years than those who start on low doses.