Background/Purpose: An accurate assessment of physical activity (PA) is critical to manage rheumatoid arthritis (RA). Accelerometry is an objective measure of PA but is not widely used clinically. The Nurses Health Study Physical Activity Questionnaire (NHSPAQ) is a brief, simple self-report measure used extensively to assess PA in adults with cancer and other chronic illnesses.  Validity of the NHSPAQ has not been determined for estimation of PA in RA. This study examines the validity of the NHSPAQ for adults with RA when compared to accelerometry estimates and data from performance tests

Methods: 32 adults with RA were sampled from a large tertiary care hospital-based arthritis clinic registry, consented and participated in a 1 week accelerometry trial. Medical and demographic data were collected including: age, gender, disease duration, disease activity (RADAI and DAS-CRP3), education, medications, and co-morbidities. At intake, participants completed the NHSPAQ, performed a self-paced 20-m Walk Test and Timed Step Test. Subjects were given an accelerometer to wear for 7 consecutive days and completed a second NHSPAQ at the end of the week. Descriptive statistics characterized the sample. Metabolic equivalents (METs) were derived from the NHSPAQ, accelerometers and the Timed Step Test using standardized algorithms. NHSPAQ validity was assessed by correlating NHSPAQ METs, accelerometer METs and METs derived from performance tests. Bland-Altman plots compared METs derived from the NHSPAQ and accelerometers. Posthoc power calculations were conducted

Results: 78% of subjects were female (mean age=62.1 years (SD=11.2). The mean disease duration of 21 years (SD=10). On average, RA disease was moderately active at intake (mean RADAI = 2.6 (SD=2) and 74% of subjects were taking biologics. The mean timed walk was 16.2 s (SD=3.9) and Timed Step Test METs was 4.7 (SD=0.9). Average weekly physical activity as determined by accelerometer was 33.3 METs (SD=23). A moderate correlation existed between NHSPAQ METs at one week and accelerometer METs (r = .67, p = .0001). Timed Step Test METs had a low correlation with self-reported physical activity levels (NHSPAQ METs) at one week (r = .42, p =.03). No significant correlation was found between disease activity, step test performance and NHSPAQ METs at intake. Bland-altman plots revealed METs derived from the NHSPAQ are more reliable among subjects with moderate to high physical activity levels. Posthoc power calculations suggested the study was appropriately powered