Psychometrics of inflammatory back pain criteria in the US population

ABSTRACT: Objectives: Evaluation of axial spondyloarthritis relies on capturing various inflammatory back pain (IBP) features represented in different criteria. These criteria are likely to be interpreted differently across demographic subpopulations. We examined IBP criteria dimensionality and differential item functioning (DIF) among different age, gender, and race groups using a nationally representative sample. Methods: We utilised the National Health and Nutrition Examination Survey (NHANES) 2009-2010 dataset, a US nationally representative data collection with demographic weighting. The Arthritis Questionnaire (ARQ) within NHANES included questions from published Calin, European Spondyloarthropathy Study Group, and Berlin criteria (8a and 7b). Confirmatory factor analysis (CFA) was performed to assess the dimensionality of IBP criteria. DIF analysis based on Item Response Theory was applied to evaluate differences in criterion performance across age (20-49 years vs >50 years), gender (men vs women), and race (white vs non-white) groups. Results: The study included 1511 patients with complete ARQ IBP questionnaires. CFA indicated that 1-factor models were generally preferred for all IBP criteria. DIF analysis revealed age-related differences in several IBP items, with non-uniform DIF observed except in the Berlin 8a criteria. In contrast, gender differences were noted primarily in the Berlin 8a lower back pain item, and racial differences were minimal. Conclusions: Findings support the predominant use of a single underlying construct in IBP criteria, validating their clinical application. However, demographic factors, especially age and gender, introduce significant variability in responses, necessitating tailored approaches in clinical assessment and further research to confirm these findings across broader populations.