P Buzkova1, J I Barzilay2, K J Mukamal3. 1. Department of Biostatistics, University of Washington, CHS CC, Bldg. 29, Suite 310, 6200 NE 74th Street, Seattle, WA, 98115, USA. buzkova@u.washington.edu. 2. Kaiser Permanente of Georgia, Division of Endocrinology and the Division of Endocrinology, Emory University School of Medicine, Atlanta, GA, USA. 3. Department of Medicine, Beth Israel Deaconess Medical Center, Brookline, MA, USA.
Abstract
The Fine-Gray method is often used instead of Cox regression to account for competing risks of death in time-to-event analyses for non-fatal outcomes. A series of examples using well-known risk factors of hip fracture in an older cohort with substantial competing mortality demonstrates that the Fine-Gray approach can yield estimates that implausibly contradict long-established associations, while Cox regression preserves them. Cox regression is generally preferred for risk factor-outcome associations even in the presence of competing risk of death. INTRODUCTION: Factors like age, sex, and race are associated not only with risk of hip fracture but also with mortality. Substantial misunderstanding remains regarding the appropriate statistical approach to account for the competing risk of mortality. METHODS: In the Cardiovascular Health Study, an ongoing cohort study of 5888 older adults, we followed participants for incident hip fracture from their 1992-1993 visit through June 2014. We contrasted the conventional cause-specific Cox analysis, which censors individuals at the time of death, with the Fine-Gray (FG) approach, which extends participant follow-up even after death, to estimate the association of well-established demographic and clinical factors with incident hip fracture. RESULTS: For age, current smoking and sex, Cox and FG methods yielded directionally concordant but quantitatively different strengths of association. For example, the Cox hazard ratio (HR) for a 5-year increment in age was 1.74 (95% CI, 1.61-1.87), while the corresponding FG HR was 1.16 (1.09-1.24). In contrast, the FG approach estimated a stronger association of hip fracture with sex. The two approaches yielded nearly identical results for race. For diabetes and kidney function, the estimates were discordant in direction, and the FG HRs suggested effects that were in the opposite direction of well-understood and widely accepted associations. CONCLUSIONS: Cause-specific Cox models provide appropriate estimates of hazard for non-fatal outcomes like hip fracture even in the presence of competing risk of mortality. The Cox approach estimates hazard in the population of individuals who have not yet had an incident hip fracture and remain alive, which is typically the group of clinical interest. The Fine-Gray method estimates hazard in a hypothetical population that can yield misleading inferences about risk factors in populations of clinical interest.
The Fine-Gray method is often used instead of Cox regression to account for competing risks of death in time-to-event analyses for non-fatal outcomes. A series of examples using well-known risk factors of hip fracture in an older cohort with substantial competing mortality demonstrates that the Fine-Gray approach can yield estimates that implausibly contradict long-established associations, while Cox regression preserves them. Cox regression is generally preferred for risk factor-outcome associations even in the presence of competing risk of death. INTRODUCTION: Factors like age, sex, and race are associated not only with risk of hip fracture but also with mortality. Substantial misunderstanding remains regarding the appropriate statistical approach to account for the competing risk of mortality. METHODS: In the Cardiovascular Health Study, an ongoing cohort study of 5888 older adults, we followed participants for incident hip fracture from their 1992-1993 visit through June 2014. We contrasted the conventional cause-specific Cox analysis, which censors individuals at the time of death, with the Fine-Gray (FG) approach, which extends participant follow-up even after death, to estimate the association of well-established demographic and clinical factors with incident hip fracture. RESULTS: For age, current smoking and sex, Cox and FG methods yielded directionally concordant but quantitatively different strengths of association. For example, the Cox hazard ratio (HR) for a 5-year increment in age was 1.74 (95% CI, 1.61-1.87), while the corresponding FG HR was 1.16 (1.09-1.24). In contrast, the FG approach estimated a stronger association of hip fracture with sex. The two approaches yielded nearly identical results for race. For diabetes and kidney function, the estimates were discordant in direction, and the FG HRs suggested effects that were in the opposite direction of well-understood and widely accepted associations. CONCLUSIONS: Cause-specific Cox models provide appropriate estimates of hazard for non-fatal outcomes like hip fracture even in the presence of competing risk of mortality. The Cox approach estimates hazard in the population of individuals who have not yet had an incident hip fracture and remain alive, which is typically the group of clinical interest. The Fine-Gray method estimates hazard in a hypothetical population that can yield misleading inferences about risk factors in populations of clinical interest.
Entities:
Keywords:
Competing risk; Cox regression; Fine-Gray approach; Hip fracture; Mortality
Authors: Andrea L C Schneider; Emma K Williams; Frederick L Brancati; Saul Blecker; Josef Coresh; Elizabeth Selvin Journal: Diabetes Care Date: 2012-12-17 Impact factor: 19.112
Authors: Charles Faselis; Joel A Nations; Charity J Morgan; Jared Antevil; Jeffrey M Roseman; Sijian Zhang; Gregg C Fonarow; Helen M Sheriff; Gregory D Trachiotis; Richard M Allman; Prakash Deedwania; Qing Zeng-Trietler; Daniel D Taub; Amiya A Ahmed; George Howard; Ali Ahmed Journal: JAMA Oncol Date: 2022-10-01 Impact factor: 33.006
Authors: Kyoung Min Kim; Li-Yung Lui; Jane A Cauley; Kristine E Ensrud; Eric S Orwoll; John T Schousboe; Steven R Cummings Journal: J Bone Miner Res Date: 2020-03-19 Impact factor: 6.741
Authors: David S Owens; Traci M Bartz; Petra Buzkova; Daniele Massera; Mary L Biggs; Selma D Carlson; Bruce M Psaty; Nona Sotoodehnia; John S Gottdiener; Jorge R Kizer Journal: Heart Date: 2021-06-02 Impact factor: 7.365
Authors: Stein Atle Lie; Anne Marie Fenstad; Stein Håkon L Lygre; Gard Kroken; Eva Dybvik; Jan-Erik Gjertsen; Geir Hallan; Håvard Dale; Ove Furnes Journal: JB JS Open Access Date: 2022-02-23