Transient Change In Steps And MVPA Due To Activity Monitoring In Adults With Knee Osteoarthritis

Abstract

Accurate measurement of physical activity (PA) is vital for clinical trials and epidemiological studies that assess important health outcomes. While objective monitoring has become a standardized approach, it is possible that people may alter their behavior due to awareness of being monitored, i.e., the Hawthorne effect. While 7 days of wear time is recommended for valid measurement, it is not known to what extent PA may be influenced by the Hawthorne effect. PURPOSE: To investigate the differences in PA, i.e., steps/day and minutes/day in Moderate to Vigorous intensity PA (MVPA) across each wear day of an Actigraph monitor. METHODS: We used data from the Osteoarthritis Initiative (OAI), a large cohort study of adults with or at risk for knee OA. An Actigraph GTM1 accelerometer was used at the 48-month follow-up visit to measure steps/day and minutes in MVPA (MVPA/day) for each day of wear. For this analysis, participants must have worn the device for 7 days and for ≥10 hours/day. To determine differences in steps/day and MVPA/day between each wear day, we used generalized estimating equations with Day 1 as our reference, adjusted for weekday vs weekend. RESULTS: 1511 participants (age = 65.2 ± 9.1 years, sex = 53.6% female, BMI = 28.2 ± 4.7 kg/m2) were included. For steps/day, participants walked 192 fewer steps on Day 4 (95% CI = [-369.6, -14.5]), 335 fewer on Day 5 [-510.7, -158.6], 275 fewer on Day 6 [-449.9, -98.9], and 238 fewer on Day 7 [-403.0, -73.1] than Day 1 (6632.1 [6438.0, 6826.2], which met statistical significance). For MVPA/day, we observed similar trends (Fig. 1). CONCLUSION: For both steps/day and MVPA/day, Days 1-3 of Actigraph wear were similar while Days 4-7 were significantly lower than Day 1. This suggests the Hawthorne effect may have influenced PA the first 3 days as participants may have increased their PA during that time. However, while statistically significant, without an MCID for PA in adults with knee OA, these results may not represent meaningful differences.