Individualisation of Physical Activity Measurement Using Wearable Technology

Warner, Ashley

Abstract

The benefits of performing regular physical activity are well documented for reducing the risk of comorbidities such as cardiovascular disease. Physical activity intensity is an important component associated with such benefits, with National Health Services advocating for moderate intensity as the minimum requirement to meet physical activity targets. For many years moderate physical activity intensity has been objectively quantified using absolute measures of physical activity intensity, with the most popular being metabolic equivalents (METs). Although the MET provides an objective measure of intensity, it does not account well for inter-individual differences in cardiorespiratory fitness. In contrast, relative measures of physical activity such as % heart rate reserve (HRR) account much better for inter-individual cardiorespiratory fitness. However, the primary limitation in using relative measures of physical activity intensity to date has been the ability to use these measures at a population level.
One solution for using relative measures at a population level is wearable technology. Sales of wearable devices are growing rapidly, with the most popular wearable device being the smart watch, and the most popular smart watch being the Apple Watch. The Apple Watch has the capacity to measure heart rate using plethysmography (PPG) sensors when placed on the wrist. Wearable devices now have the capability to measure heart rate with good accuracy, especially at rest, low, and moderate intensity. Thus, relative measures of physical activity monitoring and prescription are now becoming widely available to the general population. Therefore, the aims of this thesis were to (1) examine the agreement and relationship between %HRR and METs, (2) examine how well %HRR as measured by the Apple Watch compares to criterion intensity (VO2R) and the native Apple Watch Activity app for measuring moderate intensity while walking, (3) examine the plausibility of guiding walking cadence to meet relative moderate intensity via metronome and haptic feedback from the Apple Watch, (4) examine the perceptions of participants surrounding the use and knowledge of moderate intensity and physical activity guidelines, and (5) examine the perceptions of participants on wearable technology.
The first study (Chapter 4), a systematic review with meta regression sought to assess the agreement or disagreement between absolute and relative intensity measures of physical activity. A total of 15 papers were incorporated in the systematic review, where a comparison between METs and relative values (%HRR, %HRmax, %VO2R, %VO2max, HRindex) of intensity was undertaken. The results identified agreement for how intensity was classified (i.e., light, moderate, or vigorous) between relative and absolute intensity values in 60% of the trials. Raw data was obtained from three authors incorporating 3 studies and 290 participants. A Bayesian random intercept logistic regression was conducted to examine the agreement between relative and absolute intensity, showing agreement in 43% of all trials. Two studies had identical relative variables totalling 240 participants included in the random intercept regression. The best performing model was a log-log regression which showed that for every 1% increase in METs, %HRR increased by 1.12% (95% CI: 1.10–1.14). Specifically, the model predicts at 3 METs (lower bound of moderate intensity), %HRR was estimated to be 33% (95% CI: 18–57), highlighting the discrepancies between absolute and relative measures of physical activity intensity.
The main aim of the first experimental study (Chapter 5) was to compare a newly developed Apple Watch app (MVPA app) using %HRR as a measure of moderate-to-vigorous physical activity (MVPA) compared to the native Apple Watch Activity app (the green ‘exercise’ ring). The purpose of the investigation was to determine how accurately each method measures moderate intensity walking against the criterion measure of intensity, %VO2R. There were also several secondary aims, including obtaining the required walking cadence to reach criterion moderate intensity, 40% VO2R, comparing the Apple Watch’s ability to measure resting heart rate compared to a criterion laboratory measure and examining subjective measures (RPE, thermal sensation, breathlessness, and the talk test) ability to estimate moderate intensity. Bayesian posterior distributions indicated that at the inflection point (the treadmill speed at which the Apple Watch green ring advanced), the native Apple Watch app mean was 33% VO2R (CI: 31% to 36%) the MVPA app was 43% VO2R (95% CI: 40% to 44%). A Bayesian regression model identified %HRR and walking cadence estimated by the Apple Watch best described by a non-linear smooth regression, predicted that a mean walking cadence of ~123 steps.min-1 was required to reach 40% HRR. Mean (SD) cadence at 3 METs was 104 (9) steps.min-1. Bayesian regression models identified the subjective measures to be poor indicators of moderate intensity. At an RPE of 3 (moderate) the %HRR point estimate was 41% (95% CI: 23 – 57). Thermal sensation at 8 (slightly warm) was 39% HRR (95% CI: 22% - 58) and breathlessness of 3 (moderate) was 42% HRR (95% CI: 25 – 60). The uncertainty in subjective measures is highlighted by the large credible intervals spanning at least two intensity classifications.
The main aim of the second experimental study (Chapter 6) was to examine the use of cadence cues to guide walking intensity to criterion moderate intensity (40% HRR) via the Apple Watch using metronome and haptic cues. These trials were compared to a self-paced condition whereby participants were instructed, via NHS guidelines to moderate intensity, to walk at a cadence they perceived to be moderate based on the intrinsic, subjective instruction advised from NHS guidelines. Synchronisation between foot contact and the metronome cue were observed in a subsample of 30 participants, with the criteria for synchronisation being if foot contact was made within -0.1 to 0.1 seconds of the metronome cue. Results indicated that the metronome cue guided 81% of participants to criterion moderate intensity, whereas haptic cues guided 68% of participants to criterion moderate intensity. In the self-paced trials, 50% of participants walked at a cadence high enough to elicit criterion moderate intensity. Synchronisation with the metronome cue in the subsample of participants was achieved in 43% of foot contacts. The Bayesian posterior distributions identified participants ability to match their cadence tempo with the metronome cue tempo to be poor, with the mean effect (95% CI) being 138 (CI: 136 to 140) beats.min-1, while cadence was 126 (124 to 129) steps.min-1. Similarly, the haptic cue was administered at a mean effect of 140 beats.min-1 (CI: 137 to 143) whereas cadence in the haptic cue trials was 122 (ci: 120 – 125) steps.min-1.
The first qualitative study (Chapter 7) gathered insights into participants’ perceptions of NHS physical activity guidelines. A subsample of 23 participants who had participated in the laboratory trials agreed to partake in semi-structured interviews. A semi structured interview guide covering questions related to NHS physical activity guidelines, physical activity intensity and perceptions of wearable technology aiding physical activity was developed, with interviews conducted following the completion of each participants’ laboratory trial. Five main themes arose from the data. Collectively, the themes indicate that it is likely a large percentage of the population do not understand government guidelines for moderate intensity physical activity and are unaware of how moderate intensity should make them feel when being performed. Additionally, only three participants correctly identified the cardiovascular component of ‘150 minutes of moderate intensity or 75 minutes of vigorous physical activity a week’ described in UK NHS guidelines for physical activity. To generalise, peoples’ understanding of physical activity guidelines is poor.
Finally, the second qualitative study (Chapter 8) identified perceptions of wearable technology and the application of nudge theory to guide physical activity intensity, based on the trials performed in Chapter 6. Semi-structured interviews were conducted with the same group of 23 participants. The results indicated that participants in general would be open to the use of nudge theory to promote physical activity if it was of benefit to their cardiovascular health. The use of a metronome as a cadence cue was met with some conflicting opinions. If used persistently to guide walking cadence several participants highlighted it may be irritating to the user. Additionally, a large proportion of participants found the haptic cue to be too weak to feel during arm swing, which is a limitation that wearable device manufacturers may wish to address if they were to use haptic cues to guide physical activity intensity.
The findings from this thesis provide evidence for the benefits of relative intensity when monitoring and prescribing physical activity intensity. Moreover, these benefits are now widely available at a population level through wearable devices. The results identify the large discrepancies between relative and absolute intensity and as such, relative measures offer a more accurate method of monitoring physical activity at both an individual and, now arguably, at a population level. Furthermore, it is evident people do not understand National Health Service guidelines for physical activity. Collectively these findings have important implications for physical activity practitioners, health professionals, and researchers who seek to enhance physical activity outcomes at a population level.

Citation

Warner, A. (2024). Individualisation of Physical Activity Measurement Using Wearable Technology. (Thesis). University of Hull. https://hull-repository.worktribe.com/output/4927514

• Thesis Type: Thesis
• Deposit Date: Nov 26, 2024
• Publicly Available Date: Nov 28, 2024
• Keywords: Sports science
• Public URL: https://hull-repository.worktribe.com/output/4927514
• Additional Information: School of Sport, Exercise and Rehabilitation Science; University of Hull
• Award Date: Mar 25, 2024