LAUSR

We respectfully disagree with Subbiah et al. in their criticism of our study.1 To overcome the lack of dedicated International Classification of Diseases codes for Parkinson disease (PD) psychosis, we accrued patients with PD and included several patient characteristics of psychosis. Although residual confounding is always possible in observational research, we disagree that we compared groups with vastly different baseline characteristics. We accrued both pimavanserin users and nonusers among long-term care residents with PD to minimize selection bias at the design stage. In addition, we used inverse probability of treatment weighting (IPTW) to control for confounding at the analytic stage. After IPTW, pimavanserin users and nonusers were well balanced for observed patient characteristics, including other antipsychotic use.