Leveraging large-scale administrative claims data to evaluate prescription opioid use, risks, and outcomes in older adults living with HIV

点击翻译按钮获取中文摘要

PROJECT SUMMARY The purpose of this Mentored Research Scientist Development Award (K01) is to provide support for me to become an independent investigator with a multidisciplinary research program in HIV at the intersection of prescription drug use, comorbidities, and aging. In this project, I propose to leverage existing administrative claims data from Medicare beneficiaries ≥65 years of age from 2007-2018 to determine how high-risk prescription opioid use, e.g. high dose use (≥120 mg/day) or prolonged use (≥90 days), affects health outcomes and health care utilization in older people living with HIV (PLWH). Building on my strong foundation in HIV epidemiology and HIV comorbidities across the lifespan, the training from this K01 will allow me to (1) obtain proficiency in the analysis of large-scale, longitudinal administrative claims data; (2) attain content expertise in prescription drug utilization research and pharmacoepidemiology; and (3) develop expertise in the application of machine learning methods. My career development plan includes specific coursework, seminars, conferences, directed readings, and tailored mentoring from a multidisciplinary team comprised of experts in administrative claims data, substance use epidemiology, pharmacoepidemiology, and machine learning methods. Rutgers University provides an exceptional environment for completion of this training, with research support and infrastructure for analyzing large, multiyear datasets (including Medicare claims) and conducting high-impact HIV research. The proposed research is significant given that a substantial gap exists in understanding how prescription opioid use affects a growing population of aging PLWH who commonly report chronic pain, and have multiple comorbidities, increasing polypharmacy, and increased risk for untoward drug-drug interactions. This research is critical to target appropriate prevention and treatment programs to optimize health outcomes for this population. To fill this gap, specific aims are to: (1) Assess the associations between HIV infection and a) adverse health outcomes (e.g. falls/fractures, dementia, mortality) and b) health care utilization (e.g. emergency department use, inpatient hospitalizations, and outpatient visits) among older adults, and estimate the interaction between HIV infection and high-risk prescription opioid use on adverse outcomes and utilization; (2) Among older PLWH, estimate drug-drug interactions between high-risk prescription opioid use and specific antiretroviral drug classes or sedatives (e.g. benzodiazepines) on risk of adverse health outcomes and health care utilization; and (3) Examine the feasibility of applying existing machine learning approaches to predict adverse health outcomes and high health care utilization among older PLWH based on patient profiles and opioid prescription patterns in a large administrative claims database. Findings from this project will generate valuable new information and tools to support clinical decisions and more precisely target prevention and treatment interventions to improve health for older PLWH who are prescribed opioids, and directly inform an R01 application to study drug-drug and drug- disease interactions between widely used prescription drugs and common comorbidities among older PLWH.