Activity trackers have been more popular among a generally young and physically active population in the last decade. In addition to this group, older adults (65 and up) may benefit from activity trackers. According to Dutch norms, just 37% of older individuals in the Netherlands were properly physically active in 2018. Activity monitors can help with this by providing insight into physical activity levels, raising awareness, and inspiring older persons to be more physically active. 

Several research have found that as outcome factors for physical activity, older persons are particularly interested in step count and amount of physical behavior. Recent research has found that consumer-grade activity trackers do not accurately capture step count and physical behavior at modest walking speeds, which are common during activities of daily living (ADL) such as household chores. The majority of consumer-grade activity trackers do not include older adults as a target population and do not alter their algorithms accordingly, which explains the reduced validity. 

An older adult might prefer to carry an activity tracker in their trouser pocket, according to a recent qualitative study. As a result, according to the presented technique by Bijnens et al.(2019), the customizable algorithm was optimized to estimate step count and dynamic, standing, and sedentary time for older persons and a pocket worn activity tracker. 

Goal of this study 

The goal of this study was to confirm optimum algorithm parameter settings for step count and physical behavior in older persons during ADL using a pocket-worn activity tracker. Second, the performance of the optimized method was compared to three reference applications for a more accurate interpretation of the results. 


The modified algorithm parameter settings were evaluated and compared to the algorithm that the customizable classification algorithm is based on, as well as two activity trackers, in a cross-sectional validation study. Based on ADL, a participant-determined sequence activity regimen was created. Participants were given the option of choosing the order and duration of a series of daily activities to replicate free-living. 

This investigation used the MOX Activity Logger (MOX; Maastricht Instruments, Maastricht, NL) (Maastricht Instruments BVa, 2020). The MOX was clipped into the front trouser pocket to maintain a fixed orientation of the device in relation to the upper leg’s axial motion. 


Based on a test combination of PE, APE, correlation coefficients, and paired sample t-test, this study found that the optimized algorithm parameter settings (MOXMissActivity) can more accurately measure step count and physical behavior expressed as dynamic, standing, and sedentary time in older adults wearing an activity tracker in their trouser pocket during ADL than the MOXAnnegarn, activPAL, and Fitbit. 

MOXAnnegarn, activPAL, and Fitbit had lesser validity in the current investigation when compared to the gold standard and the MOXMissActivity. These target group and wear location specific classification techniques are obviously inapplicable outside of their respective environment. The activPAL’s findings are consistent with a recent study that found low validity in older persons during short stepping bouts and activities with low walking speeds, such as shuffling, picking, transitions, and kneeling. Fitbit’s underestimating of dynamic time can be explained by the definition of active minutes it uses: 10 minutes of sustained moderate-to-intense activity (>3 MET). It’s plausible to presume that daily activities weren’t performed with the same intensity and/or for as long during this regime. 

(a) Percentage Error and (b) Absolute Percentage Error for dynamic, standing, and sedentary time. PE and APE for dynamic time are presented in blue, for standing time in black and for sedentary time in brown. The Fitbit Alta HR measures dynamic time only, therefore no data for standing and sedentary time are presented. 

 When compared to reference applications with generic activity tracker algorithms, this study found that the optimized algorithm parameter settings can more accurately estimate step count, dynamic, standing, and sedentary time in older adults with a normal gait pattern wearing an activity tracker in their trouser pocket during a participant-determined sequence activity protocol with activities of daily living. 

Related products 

MOX1 Activity logger for physical activity assessment. 

The MOX1 is a validated accelerometer-based activity logger that seamlessly measures and records high resolution raw acceleration data up to 7 days. By using the IDEEQ software featuring our proprietary algorithms, objective measurements of human physical activity can be classified and quantified. The waterproof system design and biocompatible adhesive ensures easy application and comfortable wear by the subject. 

How can we help you with your research?  

Maastricht Instruments creates equipment in the field for accelerometry. We provide support for Clinical research, E-health applications, accelerometry algorithms and data processing and analyses. Consult us about our accelerometry products, MOX1, MOX2, MOX3, MOX5 and BACE. or eHealth applications: Hospital Fit and Miss Activity 

Article reference 

Ummels D, Bijnens W, Aarts J, Meijer K, Beurskens AJ, Beekman E. The Validation of a Pocket Worn Activity Tracker for Step Count and Physical Behavior in Older Adults during Simulated Activities of Daily Living. Gerontol Geriatr Med. 2020 Sep 30;6:2333721420951732. doi: 10.1177/2333721420951732. PMID: 33088850; PMCID: PMC7545746.