An overview of confirmatory factor analysis and item response analysis applied to instruments to evaluate primary healthcare

Abstract

This paper presents an overview of the analytic approaches that we used to assess the performance and structure of measures that evaluate primary healthcare; six instruments were administered concurrently to the same set of patients. The purpose is (a) to provide clinicians, researchers and policy makers with an overview of the psychometric methods used in this series of papers to assess instrument performance and (b) to articulate briefly the rationale, the criteria used and the ways in which results can be interpreted. For illustration, we use the case of instrument subscales evaluating accessibility. We discuss (1) distribution of items, including treatment of missing values, (2) exploratory and confirmatory factor analysis to identify how items from different subscales relate to a single underlying construct or sub-dimension and (3) item response theory analysis to examine whether items can discriminate differences between individuals with high and low scores, and whether the response options work well. Any conclusion about the relative performance of instruments or items will depend on the type of analytic technique used. Our study design and analytic methods allow us to compare instrument subscales, discern common constructs and identify potentially problematic items.