Project Summary About 40-45% of people living with Alzheimer’s disease and related dementias (ADRD) reside in a skilled nursing facility (SNF). Behavioral and psychological symptoms of dementia (BPSD) occur in ~80% of older adults with ADRD living in an SNF. Antipsychotic medications (APMs) are the most commonly used pharmacological treatment for BPSD. Because APMs are associated with numerous adverse events, clinical guidelines recommend that their use should be limited to managing acute episodes and discontinued as soon as possible. However, studies have shown that APMs are often used in individuals with ADRD for sustained periods (≥6 months). Small randomized controlled trials (RCTs) comparing withdrawing vs. continuing APMs used for BPSD have yielded conflicting and confusing results that suggested deprescribing APMs had little or no benefits for adverse events. These RCTs were clearly underpowered, and they severely underrepresented frail and complex older adults with ADRD in routine care. There was also a lack of non-randomized studies addressing this critical knowledge gap because deprescribing APMs for behavior disturbance is highly informed by symptom severity, and confounding by disease severity can be very difficult to control unless detailed clinical information is available for research. Our objective is to assess the health effects of different APM deprescribing strategies for managing BPSD in an SNF. To provide solid evidence guiding the deprescribing process, we will assess the effects of discontinuing APMs with vs. without gradual dose reduction and different rates of dose tapering. We will integrate Medicare claims data with electronic health records (EHR), Minimum Data Set (MDS), and Certification and Survey Provider Enhanced Reporting (CASPER), covering >370,000 older adults with ADRD living in an SNF from 2013 to 2026. We will employ the clone-censor-weight approach, high-dimensional machine-learning-aided proxy adjustment methods, external adjustment, and instrumental variable analysis to minimize measured and unmeasured confounding. We will address three specific aims: 1) To evaluate the prescribing and discontinuation patterns and determine the barriers to APMs deprescribing among older adults with BPSD in an SNF. 2) To determine comparative health outcomes of different discontinuation strategies vs. continuation of APMs used for BPSD in older adults who reside in an SNF. 3) To determine the treatment effect heterogeneity by key clinical phenotypes when comparing continuation vs. different discontinuation strategies of APMs used for BPSD in older adults who reside in an SNF so that such deprescribing decisions can be tailored according to patient characteristics. The impact of this proposal is high because it will generate direct evidence to inform optimal management of psychotropic medications in older adults with ADRD living in an SNF. It will also yield a scalable analytical framework specializing in comparative safety ...