Jiannong Liu1, Suying Li1, David T Gilbertson1, Keri L Monda2, Brian D Bradbury2, Allan J Collins3. 1. Chronic Disease Research Group, Minneapolis Medical Research Foundation, Minneapolis, MN. 2. Center for Observational Research, Amgen, Inc, Thousand Oaks, CA. 3. Chronic Disease Research Group, Minneapolis Medical Research Foundation, Minneapolis, MN; Department of Medicine, University of Minnesota, Minneapolis, MN. Electronic address: acollins@cdrg.org.
Abstract
BACKGROUND: Because transfusion avoidance has been the cornerstone of anemia treatment for patients with kidney disease, direct measurement of red blood cell transfusion use to assess dialysis facility anemia management performance is reasonable. We aimed to explore methods for estimating facility-level standardized transfusion ratios (STfRs) to assess provider anemia treatment practices. STUDY DESIGN: Retrospective cohort study. SETTING & PARTICIPANTS: Point prevalent US hemodialysis patients on January 1, 2009, with Medicare as primary payer and dialysis duration of 90 days or longer were included (n = 223,901). All dialysis facilities with eligible patients were included (n = 5,345). PREDICTOR: Dialysis facility assignment. OUTCOME: Receiving a red blood cell transfusion in the inpatient or outpatient setting. MEASUREMENTS: We evaluated 3 approaches for estimating STfR: ratio of observed to expected numbers of transfusions (STfR(obs)), a Bayesian approach (STfR(Bayes)), and a modified version of the Bayesian approach (STfR(modBayes)). RESULTS: The overall national transfusion rate in 2009 was 23.2 per 100 patient-years. Our model for predicting the expected number of transfusions performed well. For large facilities, all 3 STfRs worked well. However, for small facilities, while the STfR(modBayes) worked well, STfR(obs) values demonstrated instability and the STfR(Bayes) may produce more bias. LIMITATIONS: Administration of transfusions to dialysis patients reflects medical practice both within and outside the dialysis unit. Some transfusions may be deemed unavoidable and transfusion practices are subject to considerable regional variation. CONCLUSIONS: Development of an STfR metric is feasible and reasonable for assessing anemia treatment at dialysis facilities. The STfR(obs) is simple to calculate and works well for larger dialysis facilities. The STfR(modBayes) is more analytically complex, but facilitates comparisons across all dialysis facilities, including small facilities.
BACKGROUND: Because transfusion avoidance has been the cornerstone of anemia treatment for patients with kidney disease, direct measurement of red blood cell transfusion use to assess dialysis facility anemia management performance is reasonable. We aimed to explore methods for estimating facility-level standardized transfusion ratios (STfRs) to assess provider anemia treatment practices. STUDY DESIGN: Retrospective cohort study. SETTING & PARTICIPANTS: Point prevalent US hemodialysis patients on January 1, 2009, with Medicare as primary payer and dialysis duration of 90 days or longer were included (n = 223,901). All dialysis facilities with eligible patients were included (n = 5,345). PREDICTOR: Dialysis facility assignment. OUTCOME: Receiving a red blood cell transfusion in the inpatient or outpatient setting. MEASUREMENTS: We evaluated 3 approaches for estimating STfR: ratio of observed to expected numbers of transfusions (STfR(obs)), a Bayesian approach (STfR(Bayes)), and a modified version of the Bayesian approach (STfR(modBayes)). RESULTS: The overall national transfusion rate in 2009 was 23.2 per 100 patient-years. Our model for predicting the expected number of transfusions performed well. For large facilities, all 3 STfRs worked well. However, for small facilities, while the STfR(modBayes) worked well, STfR(obs) values demonstrated instability and the STfR(Bayes) may produce more bias. LIMITATIONS: Administration of transfusions to dialysis patients reflects medical practice both within and outside the dialysis unit. Some transfusions may be deemed unavoidable and transfusion practices are subject to considerable regional variation. CONCLUSIONS: Development of an STfR metric is feasible and reasonable for assessing anemia treatment at dialysis facilities. The STfR(obs) is simple to calculate and works well for larger dialysis facilities. The STfR(modBayes) is more analytically complex, but facilitates comparisons across all dialysis facilities, including small facilities.
Authors: David T Gilbertson; Heng Yan; Hairong Xu; Marvin Sinsakul; Yi Peng; James B Wetmore; Jiannong Liu; Suying Li Journal: Kidney360 Date: 2021-04-09
Authors: Celena B Peters; Jared L Hansen; Ahmad Halwani; Monique E Cho; Jianwei Leng; Tina Huynh; Zachary Burningham; John Caloyeras; Tara Matsuda; Brian C Sauer Journal: EGEMS (Wash DC) Date: 2019-07-03
Authors: Aminu K Bello; Christine M Ribic; Serge H Cournoyer; Mercedeh Kiaii; Martine LeBlanc; Melanie Poulin-Costello; David N Churchill; Norman Muirhead Journal: Can J Kidney Health Dis Date: 2018-06-05