To evaluate the psychometric performance of the Ankylosing Spondylitis Quality of Life (ASQoL) scale in patients with non-radiographic axial spondyloarthritis (nr-axSpA) to assess its appropriateness as an outcome measure in future clinical studies.
Patients with active axSpA from a Phase III, randomized, double-blind, placebo-controlled trial (RAPID-axSpA, NCT01087762) were included (N = 325). Modified New York (mNY) classification criteria were used to classify patients as having ankylosing spondylitis or nr-axSpA; those with nr-axSpA were further categorized based on objective signs of inflammation. Psychometric properties of the ASQoL were assessed/documented using a mixture of modern psychometric methods and classical test theory methods. These included exploratory factor analysis and item response theory models to assess the domain structure, test the utility of a single domain relative to subdomains, assess bias, and generate statistics to guide an empirical scoring algorithm. The reliability and validity of scores were evaluated via internal consistency, test-retest reliability, concurrent validity, and known-groups validity. Score responsiveness was assessed via anchor-based clinically meaningful change, supplemented with empirical cumulative distribution function visualizations.
The ASQoL data were defined by four domains. However, a four-domain solution was found to be inferior to a bifactor solution in which the four domains were included within a total domain. Scoring statistics supported a unit-weighted total score. Within the nr-axSpA population with objective signs of inflammation, the ASQoL mean score had adequate reliability, validity, and ability to detect clinically meaningful change.
Our findings suggest that the ASQoL is an appropriate outcome measure in interventional clinical trials in patients with nr-axSpA.