Identification and Evaluation of Self-Report Physical Activity Instruments in Adults With Osteoarthritis: A Systematic Review

Abstract

OBJECTIVE: To identify and evaluate the measurement properties of self-report physical activity instruments suitable for patients with osteoarthritis (OA). METHODS: We conducted a comprehensive 2-stage systematic review using multiple electronic databases, from inception until July 2018. In the stage 1 review, we sought to identify all self-report physical activity instruments used in individuals with joint pain attributable to OA in the foot, knee, hip, or hand. In the stage 2 review, we searched for and appraised studies investigating the measurement properties of the instruments identified. In both stages of the review, we screened all articles for study eligibility criteria, completed data extraction using the Qualitative Attributes and Measurement Properties of Physical Activity questionnaire checklist, and conducted methodology quality assessments using a modified COSMIN (COnsensus-based Standards for the selection of health Measurement INstruments) checklist. Measurement properties for each physical activity instrument were evaluated and combined, using narrative synthesis. RESULTS: In the stage 1 review, we identified 23 unique self-report physical activity instruments. In the stage 2 review, we identified 54 studies that evaluated the measurement properties of 13 of the 23 instruments identified. Instrument reliability varied from inadequate to adequate (intraclass correlation coefficient ≥0.7). Instrument construct and criterion validity assessment showed small to moderate correlations with direct measures of physical activity. Instrument responsiveness was assessed in only 1 instrument and was unable to detect changes in comparison to accelerometers. CONCLUSION: Although many instruments were identified as being potentially suitable for use in patients with OA, none demonstrated adequate measurement properties across all domains of reliability, validity, and responsiveness. Further high-quality assessment of self-report physical activity instruments is required before such measures can be recommended for use in OA research.