Construct validity of the Mini-BESTest in individuals with chronic pain in specialized pain care

Abstract

Background: Balance assessment scales are important clinical tests to identify balance impairments. Chronic pain (> 3 months) is associated with impaired dynamic balance; however, very few balance assessment scales are psychometrically evaluated for the population. The purpose of this study was to evaluate the construct validity and internal consistency of the Mini-BESTest for individuals with chronic pain in specialized pain care.

Methods: In this cross-sectional study, 180 individuals with chronic pain (> 3 months) were assessed with the Mini-BESTest and included in the analyses. For construct validity, five alternative factor structures were evaluated using a confirmatory factor analysis. In addition, we tested the a priori hypotheses about convergent validity with the 10-meter walk test, and divergent validity with the Brief Pain Inventory (BPI): pain intensity, the Tampa Scale of Kinesiophobia-11 (TSK-11), and the Pain Catastrophizing Scale (PCS-SW). Internal consistency was evaluated for the model with the best fit.

Results: A one-factor model with added covariance via the modification indices showed adequate fit indices. In line with our hypotheses, Mini-BESTest showed convergent validity (r_s = > 0.70) with the 10-meter walk test, and divergent validity (r_s = < 0.50) with BPI pain intensity, TSK-11, and PCS-SW. Internal consistency for the one-factor model was good (α = 0.92).

Conclusions: Our study supported the construct validity and internal consistency of the Mini-BESTest for measuring balance in individuals with chronic pain, who were referred to specialized pain care. The one-factor model showed an adequate fit. In comparison, models with subscales did not reach convergence, or showed high correlations between subscales, implying that Mini-BESTest is measuring one construct in this sample. We, therefore, propose using the total score, instead of subscale scores, for individuals with chronic pain. However, further studies are necessary to establish the reliability of the Mini-BESTest in the population.

Keywords: Balance; Chronic pain; Internal consistency; Mini-BESTest; Specialized pain care; Validity.

Fig. 1

Flowchart of the included and excluded data ^a Declined both the Mini-BESTest and the 10 m Walk Test. Mini-BESTest = Mini-Balance Evaluation Systems Test. 10 MWT = 10 m Walk Test, both comfortable and maximum walking speed. BPI = Brief Pain Inventory. TSK-11 = Tampa Scale for Kinesiophobia-11. PCS-SW = Pain Catastrophizing Scale - Swedish version

Fig. 2

Confirmatory factor analysis for Model 1, Model 2, and Model 5 * Confirmatory factor analysis for the original unidimensional model, presented without added covariance (A; Model 1 A) and with added covariance between items 4–5 and 7–9 (B; Model 1B), a first order 4-factor model based on the four sub-sections in the test (C; Model 2), and a first-order 4-factor model without item 7 (D; Model 5). The factor, i.e., the latent construct, is illustrated by an oval. The item, i.e., the indicator, is illustrated by a rectangle. One-headed arrows from the factors to the items represent the factor loading, and the two-headed arrows represent the covariance between the suggested factors. Each measurement error (ε1–14) is presented by the circle below the item. Item 1 - sit to stand; item 2 - rise to toes; item 3 - stand on 1 leg; items 4, 5, 6 - compensatory stepping correction – forward, backward, lateral; item 7 - stance with feet together and eyes open, on firm surface; item 8 - stance with feet together and eyes closed, on foam surface; item 9 - stance with eyes closed, on an inclined surface; item 10 - walk with change in gait speed; item 11 - walk with horizontal head turns; item 12 - walk with pivot turns; item 13 - step over obstacles; item 14 - timed up & go with dual task. APA = anticipatory postural adjustments; RPC = reactive postural control; SO = sensory orientation; DG = dynamic gait; PR = postural response; SG = stability in gait