Integrating accelerometer-based physical behaviour measurements in Danish public health surveillance: feasibility study of administrative burden

Background: Physical behaviour is typically measured using questionnaires, but this methodology has inherent biases. Accelerometers have been used supplementarily, but it is costly at scale. The Motus system, employing wireless and cloud-based technology, has demonstrated feasibility in work environment surveillance. The system has an associated smartphone application for users to deliver contextual information, (de)activate their sensor and upload data.

Purpose: This study evaluates Motus’ feasibility in public health surveillance of physical behaviour, focusing on recruitment, participation, and administrative burden.

Methods: Web-based respondents of the Danish National Health Survey (DNHS)-2023 were invited to assess physical behaviour using Motus. Feasibility was assessed as administrative burden (time, effort, resources) and participation (acceptance, compliance, wear time). A subsample wore the sensor for two weeks to assess Hawthorne’s effect. Participant characteristics were obtained through national registries and the DNHS-2023.

Results:
Of 6,993 invited, 1,617 (23.1%) accepted and 1,050 (64.9%) completed the study. We had a 75.3% return-rate of sensors; 93.1% returned among finishers, 44.7% returned among dropouts. Participants in the accelerometer study exhibited higher education (long higher education: 22.4% vs. 16,8%), met guidelines for physical activity more often (52.2% vs 45.1%), smoked less (daily/occasional smokers: 8.5% vs 13.5%) and reported better health (excellent, very good, good: 81.9% vs. 78.7%) than DNHS-2023 respondents. Study administration totalled 267 hours, averaging 9.9 minutes per participant. Tasks included equipment packing (7.8%), digital sensor-participant linking (23.2%), disinfection upon return (4.1%) and communication with participants (64.8%).In the two-week subsample, no significant differences were observed in physical behaviour between first and second week of measurement.

Conclusions: Public health surveillance of physical behaviour using Motus system seems feasible. However, before it can be used for data collection in large scale public health surveillance, solutions for minimizing sensor loss, administrative hours and selection bias are needed.