Physician Antipsychotic Overprescribing Letters and Cognitive, Behavioral, and Physical Health Outcomes among People with Dementia: A Secondary Analysis of a Randomized Clinical Trial

Abstract

Importance: Antipsychotics, such as quetiapine, are frequently prescribed to people with dementia to address behavioral symptoms but can also cause harm in this population. Objective: To determine whether warning letters to high prescribers of quetiapine can successfully reduce its use among patients with dementia and to investigate the impacts on patients' health outcomes. Design, Setting, and Participants: This is a secondary analysis of a randomized clinical trial of overprescribing letters that began in April 2015 and included the highest-volume primary care physician (PCP) prescribers of quetiapine in original Medicare. Outcomes of patients with dementia were analyzed in repeated 90-day cross-sections through December 2018. Analyses were conducted from September 2021 to February 2024. Interventions: PCPs were randomized to a placebo letter or 3 overprescribing warning letters stating that their prescribing of quetiapine was high and under review by Medicare. Main Outcomes and Measures: The primary outcome of this analysis was patients' total quetiapine use in days per 90-day period (the original trial primary outcome was total quetiapine prescribing by study PCPs). Prespecified secondary outcomes included measures of cognitive function and behavioral symptoms from nursing home assessments, indicators of depression from screening questionnaires in assessments and diagnoses in claims, metabolic diagnoses derived from assessments and claims, indicators of use of the hospital and other health care services, and death. Outcomes were analyzed separately for patients living in nursing homes and in the community. Results: Of the 5055 study PCPs, 2528 were randomized to the placebo letter, and 2527 were randomized to the 3 warning letters. A total of 84881 patients with dementia living in nursing homes and 261288 community-dwelling patients with dementia were attributed to these PCPs. There were 92874 baseline patients (mean [SD] age, 81.5 [10.5] years; 64242 female [69.2%]). The intervention reduced quetiapine use among both nursing home patients (adjusted difference, -0.7 days; 95% CI, -1.3 to -0.1 days; P =.02) and community-dwelling patients (adjusted difference, -1.5 days; 95% CI, -1.8 to -1.1 days; P