The aim of this study was to assess concurrent validity between activPAL and activPAL3 accelerometers in a sample of 53 community-dwelling older adults ≥ 65 years. Physical activity (PA) was measured simultaneously with activPAL and activPAL3 while performing scripted activities. The level of agreement between both devices was calculated for sitting/lying, standing, and walking. In addition, PA was measured over one week using activPAL to estimate the expected agreement with activPAL3 in real life. Overall agreement between activPAL and activPAL3 was 97%. Compared with activPAL, the largest disagreement was seen for standing, with 5% categorized as walking by activPAL3. For walking and sitting/lying, the disagreement was 2%, respectively. The expected daily differences between activPAL3 and activPAL were +15.0 min (95% CI: 11.3ߝ18.8) for walking and +29.5 min (95% CI: 6.2–52.7) for standing. ActivPAL and activPAL3 showed good agreement in older adults. However, if using these devices interchangeably, observed differences might still bias results.
Klenk, Büchele, Kaufmann, Peter, and Rothenbacher are with the Institute of Epidemiology and Medical Biometry, Ulm University, Ulm, Germany. Klenk and Lindemann are with the Department of Geriatrics and Geriatric Rehabilitation, Robert-Bosch Hospital, Stuttgart, Germany. Laszlo and Kobel are with the Division of Sports and Rehabilitation, Department of Internal Medicine II–Cardiology, Ulm University Medical Center, Ulm, Germany.