Bi-Factor Models in Health-Related Quality-of-Life Measurements

• Activity Number: 255
• Type: Contributed
• Date/Time: Monday, August 5, 2013 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Zugui Zhang*+
• Companies: Christiana Care Health System
• Keywords: Bi-factor Models ; exploratory factor estimation ; confirmatory factor analysis ; and higher-order factor models ; target rotations ; Schmid-Leiman transformations

Abstract

Bi-factor models have been increasingly used in recent years as an alternative approach to testing multifaceted constructs. A bi-factor model is comprised of a general factor that accounts for the commonality shared by the facets, and multiple group factors, each of which accounts for the unique influence of the specific component over and above the general factor. In this study, we will first review three exploratory factor estimation methods as well as the conceptual differences among these three methods: bi-factor rotations, target rotations, and Schmid-Leiman transformations. we will then present bi-factor models within the framework of confirmatory factor analysis and examine the conceptual differences between bi-factor models and higher-order factor models. Finally, we will introduce the potential applicability of bi-factor models in health related quality of life measurements. Two multi-domain instruments, SAQ and RAND_36, will be used to assess quality of life via a bi-factor model in the COURAGE, a randomized trial evaluating the outcomes of optimal medical therapy with or without percutaneous coronary intervention.