Factorial validity and gender invariance of the Beck Depression Inventory -- Second Edition (BDI-II) and the Multidimensional Pain Inventory (MPI) in individuals with chronic pain

Abstract

The factorial validity and gender invariance of the Beck Depression Inventory -- Second Edition (BDI-II) and the Multidimensional Pain Inventory (MPI) were examined in large samples of women and men with chronic pain. These measures are commonly employed in the assessment of individuals with chronic pain. Moreover, the BDI and the MPI Interference subscale are recommended by the Initiative on Methods, Measurement, and Pain Assessment in Clinical Trials (IMMPACT) committee for use in pain research. In Article I, four competing models of the BDI-II factor structure were examined and confirmatory factor analysis supported the conceptualization of depression as a singular latent construct, within a hierarchical factor structure consisting of three first-order factors -- Negative Attitude, Performance Difficulty, and Somatic Elements. Factor structure, item-total correlations, and correlations between subscale means and subjective pain experience, support the inclusion of somatic items despite concerns regarding their overlap with pain symptoms. Given partial measurement gender invariance, an examination of mean gender differences was warranted. The scores of women and men were similar. In Article II, models of the MPI factor structure [Kerns et al., 1985 with admendments by Rudy, 1989 and Deisinger et al., 2001] were examined. The Interference factor was cross-validated except for one problematic item, and is gender invariant. When configured according to the Deisinger Model, the Punishing Responses and Pain Severity factors provide valid, gender equivalent, information. The Support, Solicitous and Distracting Responses factors operated differently by gender. Although neither model is optimal for Section III, the Kerns Model is best. The Activities Away From Home and Social Activities factors are gender invariant, but Outdoor Work and Household Chores are not. Revisions to Sections II and III are recommended. With its focus on construct validity and gender equivalence, this dissertation complements recent calls for empirically supported measurement and is consistent with gender-fair research initiatives. Results impact chronic pain clinicians and researchers.