Incidence, prescription patterns, and determinants of antipsychotic use in patients with Parkinson's disease.

Empirical data regarding the choice of antipsychotics for the management of psychosis in patients with Parkinson's disease are limited. This study aimed to evaluate the incidence and prescribing patterns of antipsychotics and to determine the predictors associated with the prescribing of typical antipsychotics in patients with Parkinson's disease. This was a retrospective cohort study analyzing data from the National Health Insurance Research Database in Taiwan between January 1, 2000, and December 31, 2006, in which patients with Parkinson's disease (ICD-9-CM codes 332) initially receiving any antiparkinsonian drug (n = 2095) were followed up to evaluate the subsequent use of antipsychotics. Kaplan-Meier statistics and multiple logistic regression were employed to evaluate the cumulative probability of antipsychotic use and determinants of prescribing of typical antipsychotics, respectively. The cumulative probability of initiation of an antipsychotic within 6 years was found to be 51%, and the proportion of patients who began taking an atypical antipsychotic increased from 11.1% in 2001 to 36.1% in 2005. Physician specialty was found to be the most influential predictor of the prescribing of typical antipsychotics: physicians with an internal medicine specialty were 10.62 times more likely (95% confidence interval, 4.64-24.32) to prescribe typical antipsychotics than were neurologists. The use of antipsychotics in Parkinson's disease is common, and the use of typical antipsychotics dominates antipsychotic treatment. Particular attention needs to be paid to improving practice, including efforts that encourage primary care providers to have the appropriate choice of antipsychotics in patients with Parkinson's disease.