Development and Validation of a Frailty Index Using Claims Data for Pharmacoepidemiologic Studies in Older Adults
Funded Grant
Overview
Affiliation
View All
Overview
description
DESCRIPTION (provided by applicant): This application is submitted by Dae Hyun Kim, MD, MPH, ScD in response to RFA-AG-15-016, the K08 Paul B. Beeson Clinical Scientist Development Award in Aging. Dr. Kim is a board-certified geriatrician and epidemiologist at Brigham and Women's Hospital and Instructor in Medicine at Harvard Medical School. Dr. Kim has a solid background in clinical geriatric medicine, epidemiologic methods, and aging epidemiology. His prior research focused on epidemiologic evaluations of functional outcomes using population-based and hospital- based cohort studies, from which he gained insight on how the assessment of frailty could improve the care of older adults. Because frailty is a state of decreased physiologic reserve and increased vulnerability to adverse health outcomes after a stressful event, frail older patients are more likely than non-frail ones to experience adverse outcomes after certain treatments. Therefore, treatment selection should be personalized based on individuals' frailty level so as to maximize benefits and minimize harms. Pharmacoepidemiology in geriatric population remains an underdeveloped field that has great potential to provide important evidence on drug effectiveness and safety to guide treatment choices in frail older adults. Dr. Kim proposes an independent re- search career in geriatric pharmacoepidemiology with a long-term goal of establishing a method for incorporating the key but neglected dimension of frailty in assessing the effectiveness and safety of treatments in older adults. To achieve this, Dr. Kim proposes a 3-year program of career development and mentored research. Within the highly productive and supportive research environment of the Pharmacoepidemiology Division of the Brigham and Women's Hospital Department of Medicine and the larger Harvard community, he will work closely with Drs. Jerry Avorn, Sebastian Schneeweiss, Robert Glynn, and Lewis Lipsitz who will serve as his men- tors in aging and pharmacoepidemiology research. He will acquire expertise in advanced methods and build his leadership potential to become a national leader in geriatric pharmacoepidemiology. The objective of his mentored research is to develop and validate a frailty index using large datasets of health care utilization (or claims data) for pharmacoepidemiologic studies in older adults. Such observational studies are often the main sources of evidence to guide treatments in frail older adults who are poorly represented in clinical trials. How- ever, the lack of detailed clinica information on frailty in such datasets limits the ability to balance frailty be- tween treatment groups (confounding by frailty). In addition, the clinically important question of whether treatmen outcomes differ between frail and non-frail adults remains unanswered (effect modification by frailty). The significance of developing a generalizable algorithm to assess frailty in claims datasets is that we will be able to achieve balance in frailty and assess potential variation in treatment effects by frailty level in claims- based comparative effectiveness and safety research. This proposed research is innovative because it goes beyond the existing methods of confounding adjustment by directly quantifying frailty in claims datasets through application of a widely accepted frailty model in aging research (the cumulative deficit frailty index pro- posed by Rockwood et al). The derived frailty index will make it possible to measure how treatment effective- ness and safety differ between frail and non-frail adults, which cannot be examined using the existing methods. To achieve the objective, Dr. Kim will accomplish the following 2 specific aims. Aim 1 is to derive a cumulative deficit frailty index using claims in the Medicare Current Beneficiary Survey dataset, and then test its validity in an independent Health and Retirement Study-Medicare dataset. The working hypothesis is that individuals with higher claims-based frailty index scores are more likely to have clinical characteristics of frailty, fall, disability, and death than those with lower scores. Aim 2 is to validate the usefulness of this claims-based frailty index in comparative effectiveness and safety evaluation of new oral anticoagulants versus warfarin in Medicare beneficiaries with atrial fibrillation. In this real-word example, we will test the improved adjustment for confounding by frailty by reproducing the results from clinical trials by using the claims-based frailty index. To assess potential effect modification by frailty, I will test the working hypothesis that the overall benefits of new oral anticoagulants versus warfarin in reducing stroke, death, and major bleeding events are greater in frail individuals than in non-frail individuals. Ultimately, successful accomplishment of these aims is expected to have a significant impact on future research and clinical care. The proposed research will yield a novel approach to measure frailty in health care utilization datasets that complement the existing methods in defining medical product effectiveness and safety in older adults. In addition, the substantive knowledge gained from this research will guide future research on database research as well as individualized treatment choice about anticoagulation therapy by frailty status in older adults with atrial fibrillation. Finally, the proposed career development program and mentored research will position Dr. Kim to successfully compete for an NIH R01 award and become a national leader in geriatric pharmacoepidemiology research.
