Measurement Properties of the Brief Pain Inventory-short Form (BPI-SF) and Revised Short McGill Pain Questionnaire Version-2 (SF-MPQ-2) in Pain-related Musculoskeletal Conditions: A Systematic Review

Objective To systematically locate, critically appraise and summarize clinical measurement research addressing the use of BPI-SF and SF-MPQ-2 in pain-related musculoskeletal (MSK) conditions. Methods We systematically searched four databases (Medline, CINAHL, EMBASE & SCOPUS) and screened articles to identify those reporting the psychometric properties (e.g. validity, reliability) and interpretability (e.g. minimal clinically important difference) of BPI-SF and SF-MPQ-2 as evaluated in pain-related musculoskeletal conditions. Independently, two reviewers extracted data and assessed the quality of evidence with a structured quality appraisal tool and the updated COSMIN guidelines. Results Twenty-six articles were included (BPI-SF, n=17; SF-MPQ-2, n=9). Both tools lack reporting on their cross-cultural validities and measurement error indices (e.g. standard error of measurement). High quality studies suggest the tools are internally consistent (α=0.83-0.96), and they associate modestly with similar outcomes (r=0.3-0.69). Strong evidence suggest the BPI-SF conforms to its two-dimensional structure in MSK studies; the SF-MPQ-2 four-factor structure was not clearly established. Seven reports of high-to-moderate quality evidence were supportive of the BPI-SF known group validity (n=2) and responsiveness (n=5). One report of high quality established the SF-MPQ-2 responsiveness. Discussion Evidence of high to moderate quality supports the internal consistency, criterion-convergent validity, structural validity and responsiveness of the BPI-SF and SF-MPQ-2 and establishes their use as generic multidimensional pain outcomes in musculoskeletal populations. However, more studies of high quality are still needed on their retest reliability, known group validity, cross-cultural validity, interpretability properties and measurement error indices in different MSK populations.