Advancing the objective measurement of physical activity and sedentary behaviour context
thesisposted on 20.06.2017 by Adam Loveday
In order to distinguish essays and pre-prints from academic theses, we have a separate category. These are often much longer text based documents than a paper.
Objective data from national surveillance programmes show that, on average, individuals accumulate high amounts of sedentary time per day and only a small minority of adults achieve physical activity guidelines. One potential explanation for the failure of interventions to increase population levels of physical activity or decrease sedentary time is that research to date has been unable to identify the specific behavioural levers in specific contexts needed to change behaviour. Novel technology is emerging with the potential to elucidate these specific behavioural contexts and thus identify these specific behavioural levers. Therefore the aims of this four study thesis were to identify novel technologies capable of measuring the behavioural context, to evaluate and validate the most promising technology and to then pilot this technology to assess the behavioural context of older adults, shown by surveillance programmes to be the least physically active and most sedentary age group. Study one Purpose: To identify, via a systematic review, technologies which have been used or could be used to measure the location of physical activity or sedentary behaviour. Methods: Four electronic databases were searched using key terms built around behaviour, technology and location. To be eligible for inclusion papers were required to be published in English and describe a wearable or portable technology or device capable of measuring location. Searches were performed from the inception of the database up to 04/02/2015. Searches were also performed using three internet search engines. Specialised software was used to download search results and thus mitigate the potential pitfalls of changing search algorithms. Results: 188 research papers met the inclusion criteria. Global positioning systems were the most widely used location technology in the published research, followed by wearable cameras and Radio-frequency identification. Internet search engines identified 81 global positioning systems, 35 real-time locating systems and 21 wearable cameras. Conclusion: The addition of location information to existing measures of physical activity and sedentary behaviour will provide important behavioural information. Study Two Purpose: This study investigated the Actigraph proximity feature across three experiments. The aim of Experiment One was to assess the basic characteristics of the Actigraph RSSI signal across a range of straight line distances. Experiment Two aimed to assess the level of receiver device signal detection in a single room under unobstructed conditions, when various obstructions are introduced and the impacts these obstructions have on the intra and inter unit variability of the RSSI signal. Finally, Experiment Three aimed to assess signal contamination across multiple rooms (i.e. one beacon being detected in multiple rooms). Methods: Across all experiments, the receiver(s) collected data at 10 second epochs, the highest resolution possible. In Experiment One two devices, one receiver and one beacon, were placed opposite each other at 10cm increments for one minute at each distance. The RSSI-distance relationship was then visually assessed for linearity. In Experiment Two, a test room was demarcated into 0.5 x 0.5 m grids with receivers simultaneously placed in each demarcated grid. This process was then repeated under wood, metal and human obstruction conditions. Descriptive tallies were used to assess the signal detection achieved for each receiver from each beacon in each grid. Mean RSSI signal was calculated for each condition alongside intra and inter-unit standard deviation, coefficient of variation and standard error of the measurement. In Experiment Three, a test apartment was used with three beacons placed across two rooms. The researcher then completed simulated conditions for 10 minutes each across the two rooms. The percentage of epochs where a signal was detected from each of the three beacons across each test condition was then calculated. Results: In Experiment One, the relationship between RSSI and distance was found to be non-linear. In Experiment Two, high signal detection was achieved in all conditions; however, there was a large degree of intra and inter-unit variability in RSSI. In Experiment Three, there was a large degree of multi-room signal contamination. Conclusion: The Actigraph proximity feature can provide a binary indicator of room level location. Study Three Purpose: To use novel technology in three small feasibility trials to ascertain where the greatest utility can be demonstrated. Methods: Feasibility Trial One assessed the concurrent validity of electrical energy monitoring and wearable cameras as measures of television viewing. Feasibility Trial Two utilised indoor location monitoring to assess where older adult care home residents accumulate their sedentary time. Lastly, Feasibility Trial Three investigated the use of proximity sensors to quantify exposure to a height adjustable desk Results: Feasibility Trial One found that on average the television is switched on for 202 minutes per day but is visible in just 90 minutes of wearable camera images with a further 52 minutes where the participant is in their living room but the television is not visible in the image. Feasibility Trial Two found that residents were highly sedentary (sitting for an average of 720 minutes per day) and spent the majority of their time in their own rooms with more time spent in communal areas in the morning than in the afternoon. Feasibility Trial Three found a discrepancy between self-reported work hours and objectively measured office dwell time. Conclusion: The feasibility trials outlined in this study show the utility of objectively measuring context to provide more detailed and refined data. Study Four Purpose: To objectively measure the context of sedentary behaviour in the most sedentary age group, older adults. Methods: 26 residents and 13 staff were recruited from two care homes. Each participant wore an Actigraph GT9X on their non-dominant wrist and a LumoBack posture sensor on their lower back for one week. The Actigraph recorded proximity every 10 seconds and acceleration at 100 Hz. LumoBack data were provided as summaries per 5 minutes. Beacon Actigraphs were placed around each care home in the resident s rooms, communal areas and corridors. Proximity and posture data were combined in 5 minute epochs with descriptive analysis of average time spent sitting in each area produced. Acceleration data were summarised into 10 second epochs and combined with proximity data to show the average count per epoch in each area of the care home. Mann-Whitney tests were performed to test for differences between care homes. Results: No significant differences were found between Care Home One and Care Home Two in the amount of time spent sitting in communal areas of the care home (301 minutes per day and 39 minutes per day respectively, U=23, p=0.057) or in the amount of time residents spent sitting in their own room (215 minutes per day and 337 minutes per day in Care Home One and Two respectively, U=32, p=0.238). In both care homes, accelerometer measured average movement increases with the number of residents in the communal area. Conclusion: The Actigraph proximity system was able to quantify the context of sedentary behaviour in older adults. This enabled the identification of levers for behaviour change which can be used to reduce sedentary time in this group. Overall conclusion: There are a large number of technologies available with the potential to measure the context of physical activity or sedentary time. The Actigraph proximity feature is one such technology. This technology is able to provide a binary measure of proximity via the detection or non-detection of Bluetooth signal: however, the variability of the signal prohibits distance estimation. The Actigraph proximity feature, in combination with a posture sensor, is able to elucidate the context of physical activity and sedentary time.
Loughborough University, School of Sport, Exercise and Health Sciences.
- Sport, Exercise and Health Sciences