Generating Evidence on Deprescribing Safety

点击翻译按钮获取中文摘要

PROJECT SUMMARY: GEODES R01 AG070047-01 Older adults with multiple chronic conditions (MCC), especially those with Alzheimer’s Disease and Related Dementias (ADRD), have an increased risk of polypharmacy, adverse drug events, treatment burden, and cognitive changes from medication side effects. Deprescribing is the process of reducing or stopping inappropriate medications or medications unlikely to be beneficial supervised by a health-care professional. Patients, caregivers for people with ADRD, and clinicians list potential adverse drug withdrawal events (ADWE) following medication discontinuation as a key barrier to deprescribing. ADWE are not well reported by deprescribing trials, and no large-scale interventions have reported prevalence or predictors of ADWE. Further, ADWE presenting in ambulatory care are rarely captured. This lack of evidence is largely due to inadequate methods for measuring ADWE: Identifying and measuring ADWE requires knowing the timing of medication discontinuation and understanding whether subsequent symptoms or diagnoses constitute ADWE. Our experience with safety monitoring for the OPTMIZE intervention of deprescribing education in ADRD (R33AG057289, MPIs Bayliss / Boyd), a pragmatic, cluster randomized trial to increase deprescribing awareness among patients with ADRD, their family, and primary care clinicians, has exemplified these methodologic gaps and provided the impetus for the proposed work. The project will leverage knowledge and methods from safety monitoring for the ADRD population in OPTIMIZE, combined with rich longitudinal clinical and pharmacy data from the Kaiser Permanente Colorado (KPCO) EHR and Virtual Data Warehouse (VDW) to develop and test methods to generate robust evidence on deprescribing safety. Our study populations will be older adults with MCC and the OPTIMIZE population with dementia plus MCC. Aim 1 will identify individuals who experience medication discontinuation and subsequent ADWE using record review and adjudication. Aim 2 will develop and validate a text mining tool to identify medication discontinuation applicable to multiple classes of medications. Aim 3 will apply the text mining tool in combination with common ADWE diagnoses in two separate cohorts to demonstrate a data-based approach to identifying medication discontinuation and possible ADWE: a sample of older adults with MCC and the OPTIMIZE population of individuals with dementia plus MCC. Aim 4 will model patient and clinical factors predictive of ADWE in the ADRD and MCC populations. Generating adequate evidence to support safe and effective deprescribing for ADRD and MCC populations requires new data-based methods and approaches.