OBJECTIVE: To summarize the evidence on content and structural validity of 17 patient-reported outcome measures (PROMs) to measure physical functioning in LBP.
STUDY DESIGN AND SETTING: MEDLINE, EMBASE, CINAHL, PsycINFO, SportDiscus, and Google Scholar were searched (February 2017). Records on development, and studies assessing content validity or unidimensionality in patients with LBP were included. Two reviewers defined eligible studies and assessed their methodological quality with updated COSMIN standards. Evidence was synthesized for three separate aspects of content validity: relevance, comprehensiveness and comprehensibility, and for unidimensionality; a modified GRADE approach was applied to evidence synthesis.
RESULTS: High quality evidence showed that 24-item Roland Morris Disability Questionnaire (RMDQ-24) is a comprehensible but not comprehensive PROM. Low to very low quality evidence underpinned the content validity of the other PROMs. Unidimensionality was: sufficient for Brief Pain Inventory pain interference subscale (moderate quality evidence); inconsistent for RMDQ-23, Oswestry Disability Index 2.1a (ODI 2.1a), and Quebec Back Pain Disability Scale (QBPDS, moderate quality); insufficient for RMDQ-24, ODI 1.0 and RMDQ-18 (high quality), and Short Form 36 physical functioning subscale (SF36-PF, moderate quality).
CONCLUSION: The content validity of PROMs to measure physical functioning in LBP is understudied. Structural validity of several widely used PROMs is problematic.