Abstract: PO1695
Tailoring the Beers Criteria for Mortality Risk Stratification Among Older Adults Initiating Hemodialysis
Session Information
- Advances in Geriatric Nephrology
October 22, 2020 | Location: On-Demand
Abstract Time: 10:00 AM - 12:00 PM
Category: Geriatric Nephrology
- 1100 Geriatric Nephrology
Authors
- Hall, Rasheeda K., Durham VA Medical Center, Durham, North Carolina, United States
- Muzaale, Abimereki, Johns Hopkins University, Baltimore, Maryland, United States
- Segev, Dorry L., Johns Hopkins University, Baltimore, Maryland, United States
- McAdams-DeMarco, Mara, Johns Hopkins University, Baltimore, Maryland, United States
Background
American Geriatrics Society’s Beers Criteria lists potentially inappropriate medications (PIMs) that carry more risk of harm than benefit in older adults, but PIM risks may differ in kidney failure. To tailor the Beers Criteria, we developed a novel mortality risk score for older patients initiating hemodialysis.
Methods
We assembled a USRDS cohort of 39,098 adults aged ≥65 initiating hemodialysis (2013-2014) and enrolled in Medicare Part D by 90 days post-initiation. We used Part D claims days’ supply to quantify prescription length of Beers Criteria PIMs. In a training cohort (60% sample), we identified which of 38 PIM classes were associated with mortality using Cox modeling; censoring for loss of Medicare coverage, modality change, or 9/1/2015. Models were adjusted for demographics, initiation year, comorbidities, drug dependence, smoking status, inability to ambulate, and institutionalization. PIMs classes that were associated with mortality were summed to create a PIM count risk score. We used Cox models to estimate the association of PIM count risk score (time-varying) with mortality in training and validation cohorts.
Results
The training cohort (n=23,521) had mean age 75 years, 43% women, 21% black, and 75% (n=17,706) had ≥2 PIM fills/month. We identified 15 PIMs (HR >1) to include in the risk score (Figure). Patients with ≥2 fills/month (vs. no fills) were more likely institutionalized (13.8 vs. 10.1%), non-ambulatory (20.8 vs. 17.2%), and have cardiovascular disease (62.4 vs. 50.7%). Compared to those with no fills, there was increased mortality risk among those with 1 fill (aHR=1.32;1.25-1.39) and ≥2 fills (aHR=1.66;1.56-1.75).
Conclusion
We identified 15 of 38 PIM classes associated with mortality to yield a novel PIM count risk score. This score facilitates tailoring the Beers Criteria to promote age-appropriate prescribing in older adults initiating hemodialysis.
PIM Classes and Mortality Risk
Funding
- NIDDK Support