Prescribing trends and associated outcomes of antiepileptic drugs and other psychotropic medications in US nursing homes: A mixed-method investigation incorporating a secondary data analysis of multiple CMS datasets and a clinician survey

Abstract

Abstract Background: Pilot data suggests off-label unmonitored antiepileptic drug (AED) prescribing for behavioral and psychological symptoms of dementia is increasing, replacing other psychotropic medications targeted by purposeful reduction efforts. This trend accelerated during the COVID-19 pandemic. Although adverse outcomes related to this trend remain unknown, preliminary results hint harms may be increasing and concentrated in vulnerable populations. Methods and analysis: A mixed-methods study, including a retrospective secondary data analysis and a national clinician survey, will investigate 1) the appropriateness of AED and other psychotropic prescribing in US nursing homes; 2) characteristics and outcomes associated with this prescribing; and 3) how these phenomena may be changing under the combined stressors of the pandemic and medication reduction initiatives. Quantitative data analysis will draw on data elements, linked at an individual level, from CMS’s Minimum Data Set, Medicare Part D, MedPAR, Outpatient, and Public Use Files. Data will be organized into quarterly cohorts of nursing home residents for the years 2009-2021. Each cohort will incorporate individual level data including: demographics, health status, disease variables, psychotropic medication claims, comprehensive nursing home health outcomes, hospital and ER adverse events, and long-stay facility details, including staffing resources and COVID-19 statistics. To help explain and validate findings, we will conduct a national qualitative survey of nursing home prescribers regarding their knowledge and beliefs surrounding changing approaches to dementia care and associated outcomes. Discussion: This study will detail appropriate and inappropriate AED use and related outcomes in nursing homes, as well as describe disparities in long-stay sub-populations treated or not treated with psychotropics. It will delineate the impact of the pandemic in combination with national policies on dementia management and outcomes. We believe this mixed-method approach, including processes that link multiple CMS datasets at an individual level and survey relevant stakeholders, can be replicated and applied to evaluate a variety of patient-oriented questions in diverse clinical populations. Trial Registration: Pending