Premium
Application of Traditional and Emerging Methods for the Joint Analysis of Repeated Measurements With Time‐to‐Event Outcomes in Rheumatology
Author(s) -
Arbeeva Liubov,
Nelson Amanda E.,
Alvarez Carolina,
Cleveland Rebecca J.,
Allen Kelli D.,
Golightly Yvonne M.,
Jordan Joanne M.,
Callahan Leigh F.,
Schwartz Todd A.
Publication year - 2020
Publication title -
arthritis care and research
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 2.032
H-Index - 163
eISSN - 2151-4658
pISSN - 2151-464X
DOI - 10.1002/acr.23881
Subject(s) - medicine , event (particle physics) , repeated measures design , time point , context (archaeology) , osteoarthritis , proportional hazards model , event data , joint (building) , dependency (uml) , statistics , data mining , computer science , artificial intelligence , mathematics , architectural engineering , engineering , biology , aesthetics , paleontology , philosophy , physics , alternative medicine , pathology , quantum mechanics , analytics
Objective The goal of this paper is to describe approaches for the joint analysis of repeatedly measured data with time‐to‐event end points, first separately and then in the framework of a single comprehensive model, emphasizing the efficiency of the latter approach. Data from the Johnston County Osteoarthritis (JoCo OA ) Project will be used as an example to investigate the relationship between the change in repeatedly measured body mass index ( BMI ) and the time‐to‐event end point of incident worsening of radiographic knee OA that was defined as an increased Kellgren/Lawrence grade in at least 1 knee over time. Methods First, we provide an overview of the methods for analyzing repeated measurements and time‐to‐event end points separately. Then, we describe traditional (Cox proportional hazards model [Cox PH ]) and emerging (joint model [ JM ]) approaches, both of which allow combined analysis of repeated measures with a time‐to‐event end point in the framework of a single statistical model. Finally, we apply the models to JoCo OA data and interpret and compare the results from the different approaches. Results Applications of the JM (but not the Cox PH ) showed that the risk of worsening radiographic OA is higher when BMI is higher or increasing, thus illustrating the advantages of the JM for analyzing such dynamic measures in a longitudinal study. Conclusion Joint models are preferable for simultaneous analyses of repeated measurement and time‐to‐event outcomes, particularly in the context of chronic disease, where dependency between the time‐to‐event end point and the longitudinal trajectory of repeated measurements is inherent.