Comparison of outcomes for African Americans, Hispanics, and Non-Hispanic Whites in the CATIE study.

OBJECTIVE: Medication outcome literature in schizophrenia across racial-ethnic groups is sparse, with inconsistent findings. The Clinical Antipsychotic Trials of Intervention Effectiveness (CATIE) study provided an opportunity for exploratory analyses of racial-ethnic outcomes. The study objective was to examine race-ethnicity outcomes for CATIE's main outcome (study discontinuation) and secondary outcomes.

METHODS: CATIE participants included whites (non-Hispanic) (N=722), African Americans (N=506), and Hispanics (N=170). Survival analyses and mixed-effects regression modeling were conducted, with adjustment for baseline sociodemographic differences and baseline scores of the secondary outcomes.

RESULTS: Racial-ethnic groups had unique patterns of outcomes. Hispanics were much more likely to discontinue for lack of efficacy from perphenazine (64% versus 42% non-Hispanic whites and 24% African Americans) and ziprasidone (71% versus 40% non-Hispanic whites and 24% African Americans); Hispanics' quality of life also declined on these medications. Non-Hispanic whites were more likely to discontinue for lack of efficacy in general (averaging olanzapine, quetiapine, and risperidone discontinuation rates). African Americans were less likely to continue after the first phase (32% continuing versus 40% for non-Hispanic whites and 41% Hispanics). Discontinuations were driven by research burden, personal issues, and unspecified loss to follow-up. Non-Hispanic whites had higher depression scores during the follow-up period. African Americans had fewer side effects.

CONCLUSIONS: CATIE results did not show disparities favoring non-Hispanic whites. CATIE may have provided state-of-the-art treatment and thus reduced disparate treatments observed in community clinics. African Americans discontinued even after consideration of socioeconomic differences. Why perphenazine and ziprasidone may be less effective with Hispanics should be explored.