Real-world treatment of schizophrenia is being scrutinized in the Clinical Antipsychotic Trials of Intervention Effectiveness (CATIE). In phase 1, a conventional antipsychotic and several newer, “atypical,” antipsychotics all had high rates of treatment discontinuation. Simulating clinical practice, phase 2 switched these patients to a different antipsychotic, one of several atypicals. Two 18-month trials were undertaken, targeting patients who discontinued treatment because of intolerable side effects and those who quit because of inefficacy. Discontinuation rates were again high, but there were noteworthy differences between drugs. Clozapine was included in one of the two trials, described by McEvoy et al. (p. 600). As expected, it produced less discontinuation (56% versus 71%–93%) and more time in treatment (10.5 months versus 2.7–3.3 months). Agranulocytosis and eosinophilia each developed in one of the 45 patients, confirming the need for safety monitoring. In a trial that excluded clozapine, reported by Stroup et al. (p. 611), discontinuation rates were 64%-84%, but olanzapine and risperidone produced longer median times to discontinuation (6.3 and 7.0 months) than quetiapine and ziprasidone (4.0 and 2.8 months) (see figure above). Ziprasidone had the highest rate of serious adverse events. The results for olanzapne illustrate the need to tailor treatment to the individual-in phases 1 and 2 it produced longer treatment continuation, but the effects on weight, cholesterol, and triglycerides rule it out for certain patients. An editorial by Carol Tamminga on treatment of schizophrenia is on p. 563.