Darcy Ummels

General discussion | 147 7 reduced physical activity levels and different walking patterns 10 ‐ 14 and rely on activities of daily living as their primary sources of physical activity. 15 Within the validity studies in this dissertation, people using walking aids (e.g. a cane or walker) were exclued because they have differnt gait patterns compared to people who do not use walking aids. 16 Therefore, the results from the validity studies are probably not transferable to people who use walking aids. The algorithm developed in phase two could potentially be adapted to people who use walking aids in the same way as was done for people with low walking speeds in general and/or during activities of daily living. This could be relevant for a large group of older adults and (older) adults with a chronic disease, but also for people who are recovering in hospitals or rehabilitation centres. Another target population might be ‘white ‐ collar’ workers, since their occupations are generally sedentary, requiring low levels of physical activity. 17 The user interface was designed according to the preferences and needs of older adults, but the main aim was to make a simple user interface. In principle, an uncomplicated user interface could be used by any target group with a preference for simplicity. Another consideration regarding the target group was that the group could have had a selection bias; participants in the studies could have been more interested in technology and possibly more skilled at using technology. This might have affected the results of the development study and feasibility studies. However, using technology is becoming more popular among older adults. In 2019, 76% of older adults were using social media, with 73% online on a daily basis and 62% using the internet to retrieve health information. 18,19 Therefore, the results of the development study and feasibility study might be generalisable for a growing group of older adults, both now and for years to come. Determining the validity of activity trackers One strength of this dissertation is the setup of the validation study; it used not only activities of daily living, but also a widely accepted gold standard. In several other studies, activity trackers were validated in lab settings using a gold standard but not a free ‐ living protocol with activities of daily living (e.g. using a treadmill). Alternatively, a free ‐ living protocol was used, but validation took place without comparison to a gold standard. 20–24 In the work for this dissertation, the research team simulated free living as far as possible, still using the gold standard to validate the activity trackers. The team also used other commercially available activity trackers as references, and by having several references, the validity of the activity trackers could be more accurately interpreted. One recent study proposed a protocol for determining the validity of the step count of consumer wearables and smartphones. 1 Most of the criteria for this