After the black box warning: predictors of psychotropic treatment choices for older patients with dementia.

OBJECTIVES: This study aimed to evaluate factors associated with the selection of pharmacological treatments often given as first-line treatments to elderly patients with neuropsychiatric symptoms associated with dementia. It also evaluated patterns of medication usage over time in the year preceding and three years after the U.S. Food and Drug Administration issued a black box warning for antipsychotic usage.
METHODS: A retrospective cohort consisted of 19,517 Veterans Affairs patients with diagnosed dementia and a new outpatient start with an antipsychotic agent (haloperidol, olanzapine, quetiapine, or risperidone) or valproic acid and its derivatives between May 1, 2004, and September 30, 2008. Patient and facility characteristics were examined for their association with the new starts of these medications.
RESULTS: Trends in the rate of fills for psychotropic medications varied, with yearly increases in the use of quetiapine, haloperidol, and valproic acid and decreasing use of olanzapine and risperidone. Predictors of haloperidol use included a new start in nonpsychiatric settings, prior benzodiazepine use, and any prior-year hospitalization. Anxiety disorder and major depression were predictive of not receiving haloperidol. Parkinson's disease was predictive of quetiapine use, whereas bipolar disorder was predictive of valproic acid use. Older age was predictive of use of antipsychotics rather than valproic acid. Urban facilities were less likely to use olanzapine, and significant regional variations were seen.
CONCLUSIONS: Important patient and facility characteristics were associated with initiating different psychotropic agents among elderly dementia patients. In addition, the rate of use and the factors predictive of use varied across the study years.