Full Citation
Title: Competing classes confront competing risks: unraveling mortality inequities with parametric g-computation
Citation Type: Journal Article
Forthcoming?: Yes
ISBN:
ISSN:
DOI:
NSFID:
PMCID:
PMID:
Abstract: Measuring health inequities across social groups is crucial for allocating public- health resources, and for designing policy and organizing strategies to mitigate inequities and improve population health.1,2 Measuring absolute (versus relative) inequities in risks (versus rates) of adverse outcomes may be most useful for resource allocation and meaningful to stakeholders.3–6 Nonetheless, defining and measuring absolute inequities in risks of adverse outcomes can be challenging in settings with competing events, which are events, like cancer mortality, that preclude the occurrence of the event of interest, like cardiovascular-disease mortality.5,7 For example, the naïve Kaplan-Meier estimator targets the conditional risk: the cumulative probability of the event of interest in a hypothetical population in which competing events are eliminated but the hazard of the event of interest is unaltered. Meanwhile, the Aalen-Johansen (AJ) estimator targets the unconditional risk: the cumulative probability of the event of interest among the population observed at baseline, with no change to the hazard of competing events. While either type of risk may be suitable depending on the research question, the unconditional risk targeted by the AJ estimator may be more interpretable and policy relevant, as it corresponds to the realistic setting in which only population members who do not experience competing events can experience the event of interest, rather than to a hypothetical world with vastly different disease burdens.
User Submitted?: No
Authors: Jerzy Eisenberg-Guyot, Audrey Renson
Periodical (Full): American Journal of Epidemiology
Issue:
Volume:
Pages: 1-15
Data Collections: IPUMS Health Surveys - NHIS
Topics: Health, Population Health and Health Systems
Countries: