Although efficacy of antipsychotic medications is well documented, controlled clinical trials have called into question their effectiveness. We examined the effectiveness of olanzapine and risperidone in schizophrenia in a naturalistic setting.

We used an electronic medical records database at a Veterans Affairs medical center to conduct a retrospective study of all new outpatient medication trials of olanzapine (N= 221) and risperidone (N= 274) over a 2-year period beginning January 1999 in patients diagnosed with schizophrenia or schizoaffective disorder. We defined medication discontinuation as a switch between the two agents (the majority of switches) or self discontinuation by not getting medication supply for over one month.

Sample mean age (± SD) was 48.4 (±11.6) years, 91% were male. Discontinuation rates were high (73%) trending lower in olanzapine (70%) than risperidone (76%) (p= 0.12). Median time to discontinuation was 120 days (95% CI: 105-135), longer for olanzapine (150, 95% CI: 120-180) than risperidone (90, 95% CI: 71-109) (p= 0.04).

Self discontinuation was high (48%) with no significant difference between olanzapine (50%) and risperidone (46%). Switching rate was 25% and more likely to occur in risperidone (30%) than olanzapine (20%) (OR= 1.72, 95% CI: 1.13-2.61).

Effectiveness of antipsychotic medications in schizophrenia may be hampered by high rates of medication self discontinuation in outpatient practice settings. Time to discontinuation suggests olanzapine may be more effective than risperidone. Strategies to address causes of poor adherence should be incorporated in medication algorithms to optimize their effectiveness.