Application of Exploratory Factor Analysis and Item Response Theory to Validate NHANES ADL Scale in Patients Reporting Rheumatoid Arthritis

Abstract

Background: Several instruments are used for measuring functional limitations among rheumatoid arthritis (RA) patients. However, these instruments are incongruously assessed for their psychometric properties. The National Health and Nutritional Examination Survey (NHANES) uses a generic questionnaire to assess the activities of daily living (ADL) to measure functional limitations among its participants. The psychometric properties of the NHANES-ADL scale were evaluated using a patient examination and survey data. Methods: NHANES-ADL scale was assessed for its internal consistency and factor structure. Scale reliability was assessed with Cronbach’s alpha reliability coefficient. Principal component analysis with Promax rotation was used to obtain factor structure. Confirmatory factory analysis was used to calculate fit indices. The graded item response theory model was used to estimate item discrimination, difficulty, and test information. Results: Our sample included 1132 individuals with RA. Exploratory factor analyses of 19-item NHANES ADL scale produced one factor solution and accounted for 35% of variance. The Cronbach alpha of this scale was 0.92. The results of graded item response model indicated items performing well discriminating high and low level of functional ability. A higher slope (α) reflected stronger ability of items to discriminate across the continuum. Conclusions: The NHANES ADL scale showed good reliability, single dimensionality, and validity in RA patients. Studies should explore its test-retest reliability and its ability to reliably measure functional change over time in the future.