Physical body experiences questionnaire simplified for active aging (PBE-QAG): Validation with Rasch measurement theory

Abstract

Purpose To validate the Physical body experiences questionnaire simplified for active aging (PBE-QAG) with Rasch measurement theory. PBE-QAG measures body awareness during physical activity and includes dimensions of body-mind relationship, body acceptance, and awareness of physical skills and limits. Methods Adults without pain (n=269), with pain (n=61), with mental health conditions (n=200), and with stroke (n=36) were recruited at the Minnesota State Fair, Highland Fest, and in the Brain Body Mind Lab (University of Minnesota) and completed demographic and clinical questionnaires as well as the PBE-QAG. The PBE-QAG has 12 items, with scores ranging between 0 (totally true) to 4 (totally false). A low total score on the PBE-QAG reflects better body awareness. We evaluated item and person fit, targeting, unidimensionality, person separation reliability (PSR), local item dependence (LID), and differential item functioning (DIF) for demographic and clinical characteristics. We compared with Kruskal-Wallis ANOVA the person mean location in four groups: Adults with or without mental health conditions; and whether those groups did body awareness training. Results Unidimensionality and item fit were obtained after deleting 2 and rescoring 5 items. Seven participants did not fit the model (1.23%). There was minimal floor (5.72%), no ceiling effect (0.00%), and no LID. No DIF was greater than 0.50 logits for any of the variables. The Wright-corrected PSR was 0.96. The person mean location was -1.71+/-1.21 logits. Adults with mental health conditions who did not practice body awareness had a higher person mean location [Median (IQR)=0.83(0.89) logits, p<0.0001] versus the other three groups, reflecting lower body awareness. Conclusions PBE-QAG demonstrated good item and person fit, but the targeting is off. Therefore, the current version of PBE-QAG is not recommended for use in the general population. We encourage further validation of PBE-QAG in adults with mental health conditions who do not practice body awareness.

Preliminary Rasch analysis of the Multidimensional Assessment of Interoceptive Awareness in adults with stroke

“…A "threshold" is the point at which adjacent categories have the same likelihood of being selected [66,67]. Disordered thresholds reveal that the logical estimated order of the scale construct is not measured appropriately by the item response categories and that respondents may have had difficulty differentiating between categories and/or that not all categories were used [60,[67][68][69][70][71]. Fit statistics indicate the appropriateness of person and item fit to the Rasch model [67].…”

“…LID appears when a pair of items share content to a greater degree than other items on the assessment [65,76]. Standard residual correlations of 0.20 or greater than the average of the standard residual item correlation indicate the presence of LID [60,77]. The person mean location was 0.05±1.13 logits indicating that the MAIA item difficulty was well targeted for this population.…”

“…Person and item fit residuals greater than +2.50 indicate misfitting items or persons, and less than -2.50 indicate redundant items. Significant p -values for item fit are calculated with Bonferroni correction [60].…”

“…PSR differentiates the person’s ability levels of the trait for research or clinical purposes, with PSR > 0.90 allowing the researcher or clinician to make decisions for individuals and ≥ 0.70 to make group decisions [60,73,74]. Floor and ceiling effects above 15% are considered problematic [60, 72].…”

“…PSR differentiates the person’s ability levels of the trait for research or clinical purposes, with PSR > 0.90 allowing the researcher or clinician to make decisions for individuals and ≥ 0.70 to make group decisions [60,73,74]. Floor and ceiling effects above 15% are considered problematic [60, 72]. The average person location within a range of -0.50 and 0.50 logits of average item location indicates adequate assessment targeting [60,75].…”