Comparing of reliability, construct validity and responsiveness of the IPAQ-SF and pase in adults with osteoarthritis

Smith, R.D.; Dziedzic, K.S.; Quicke, J.G.; McHugh, G.A.; Healey, E.L.

Abstract

Purpose: Physical activity (PA) is recommended as a core treatment for all adults with osteoarthritis (OA), yet actual levels of physical activity in this population are low. Valid and reliable measures of PA levels that are responsive to change in adults with OA are essential to accurately assess interventions aiming to increase PA. Although the International Physical Activity Questionnaire - Short Form (IPAQ-SF) and the Physical Activity Scale for the Elderly (PASE) are commonly used in OA research, there is currently a lack of evidence that demonstrate the validity, reliability and responsiveness of these self-report instruments in adults with OA. The aim of this study was to evaluate and compare the test-retest reliability, construct validity and responsiveness of the IPAQ-SF and PASE in adults with OA.

Methods: This study was a secondary analysis of the MOSAICS cluster randomised controlled trial (ISRCTN number ISRCTN06984617) baseline and 3-month follow up questionnaire data; the sample (n=525) included adults aged 45 years and over consulting primary care with OA of the knee, hip, hand or foot. Total IPAQ-SF scores were reported in median and interquartile range (IQR) and PASE scores were reported in mean and standard deviation (SD). To evaluate construct validity of both instruments baseline IPAQ-SF and PASE scores were compared using Pearson’s correlation for total scores and Spearman’s rank correlation (for subdomains). Test-retest reliability and measurement error was assessed in a subsample (n=401) that reported remaining stable in their physical activity between baseline and 3-month follow up, reliability was evaluated using a random effects Intra-class correlation for absolute agreement (ICC). Measurement error was assessed using standard error of measurement (SEM), smallest detectable change (SDC) and limits of agreement (LOA). Responsiveness was assessed in a subsample (n=90) that reported not participating in exercise at baseline but participated in exercise at 3-month follow up using effect size (ES), standard responsive measurement (SRM) and response ratio (RR).

Results: There were 310 (55% female) participants who completed the IPAQ-SF and PASE at baseline and used in this analysis. Mean age of participants was 67.2 (10.5 SD). Both the IPAQ-SF (median 1440 METS-1mins-1week, 206-3600 IQR, ranges 0-16398) and PASE (mean 145.1, 79.5 SD, score ranges 0-400) total scores correlated strongly (r=0.56), subdomain correlations found small correlations in moderate activities (rs =0.34) and vigorous activities (rs =0.39) compared to sitting activities (rs =0.46) and walking activities (rs =0.57). The PASE showed greater reliability compared to the IPAQ-SF (ICC=0.68; 0.61-0.74 95%CI & ICC=0.64; 0.55-0.72, respectively). In the PASE; SEM=46.7, SDC=129.6 and LOA ranges=-130 to 112, suggesting relatively large measurement error when considering PASE maximal scoring range of 0-400 (Figure 1.). The IPAQ-SF; SEM=3532.2 METS-1mins-1week, SDC= 9790.8, LOA ranges=-3942 to 4509 METS-1mins-1week, suggesting large measurement error in weekly energy expenditure (Figure 2.). Responsiveness was poor in both measures with ES of -0.14 and -0.16, SRM of -0.21 and -0.21 and RR of 0.12 and 0.09 for the IPAQ-SF and PASE, respectively.

Conclusions: Currently, there are no recommendations on selecting instruments that measure PA in adults with OA. Overall the IPAQ-SF and PASE appear limited in reliability, measurement error and responsiveness in adults with OA. This study demonstrated that both instruments investigated correlated strongly with each other, suggesting the IPAQ-SF and PASE are similar in terms of measuring total PA levels in our sample. A limitation of our study was that we could not test construct validity against an objective measure such as accelerometery. The PASE contains several subdomains that the IPAQ-SF doesn’t, which may explain the weaker correlations in the subdomains. The general findings on the IPAQ-SF and PASE in terms of reliability and measurement error in our study are in line with a previous research into the measurement properties of the IPAQ-SF and PASE self-report instruments measuring PA in adults with OA. The measurement properties of the IPAQ-SF and PASE should inform decision making of researchers and clinicians when selecting measures of PA, particularly when comparing different levels of PA and monitoring PA changes over time in those with OA.

Citation

Smith, R., Dziedzic, K., Quicke, J., McHugh, G., & Healey, E. (2020). Comparing of reliability, construct validity and responsiveness of the IPAQ-SF and pase in adults with osteoarthritis. Osteoarthritis and Cartilage, 28(S1), S428. https://doi.org/10.1016/j.joca.2020.02.663

• Journal Article Type: Conference Paper
• Conference Name: 2020 OARSI World Congress on Osteoarthritis, 30 April-3 May 2020
• Conference Location: Messe Wien Exhibition & Congress Center, Vienna, Austria
• Online Publication Date: Apr 20, 2020
• Publication Date: 2020-04
• Deposit Date: Jun 21, 2023
• Journal: Osteoarthritis and Cartilage
• Print ISSN: 1063-4584
• Publisher: Elsevier
• Peer Reviewed: Peer Reviewed
• Volume: 28
• Issue: S1
• Pages: S428
• DOI: https://doi.org/10.1016/j.joca.2020.02.663
• Keywords: Orthopedics and Sports Medicine; Biomedical Engineering; Rheumatology