Workshops are free of charge for all registered conference participants.
Quantifying and Visualising Physical Behaviour: An alternative to energy expenditure estimation in evaluation of physical activity interventions.
Authors and affiliations
Malcolm Granat, School of Health Sciences, University of Salford, Salford, UK
Kate Lyden, PAL Technologies Ltd, Glasgow, Scotland, UK
It has been suggested that physical activity is about “the relationship between human beings and their environment” and the “strengthening of that relationship”. However, the primary physical activity outcome has invariably been energy expenditure, with definitions of different aspects of physical activity (Sedentary Behaviour, MVPA, compliance with guidelines etc.) based on levels of energy expenditure. It is proposed that the patterns of robustly defined physical activities, Physical Behaviour, can provide an alternative construct to energy expenditure estimation.
The understanding and quantification of free-living Physical Behaviour is important not only in determining the relationship between these behaviours and health but also in planning interventions and developing public health messages. Physical Behaviour is made up of free-living activities whose patterns reflect the response to a range of factors and constraints. These can be external influences (e.g. from vocational or environmental), health related, or free choice. It is proposed that the type of activities (with their associated movements) and the choice of when these are performed, i.e. their pattern in time, are the building blocks of free-living Physical Behaviour. Using body-worn sensors we can classify the primary activities of lying, sitting, standing, walking, car transportation and cycling in the free-living individual. Physical Activity interventions are aimed at changing this behaviour and these interventions may, indirectly, change energy expenditure.
Hence the primary effect of an intervention is not operating at the level of changing energy expenditure but changing behaviour, behaviour being the pattern of activities.
There are two main goals of this workshop. Firstly is to demonstrate how, using event-based analysis, we can develop person-centred outcomes from the pattern of the individual’s activity and how from these patterns it is possible to derive detailed information on walking bouts and commuting (walking, cycling, transportation). Secondly, this workshop will address ways in which data analysis and visualisation can be made person-focused, bringing the needs of the participant or patient to the centre.
Malcolm Granat and Kate Lyden will open the workshop with a structured discussion using exemplar data to outline the key concepts of the analysis method. This discussion will outline how we use event-based analysis to quantify specific features of interest from the data.
There will then be an interactive discussion using “live” data, and sample data from different populations, on how the activities of lying, cycling and car travel can be derived from body-worn accelerometer data alone.
We will then demonstrate how we can analyse these activities, volumes and patterns, to build up a picture of the individual’s Physical Behaviour and look at ways in which these patterns can be quantified.
The workshop participants will be divided into groups each of which will be given a data set from different populations who use different modes of transport to commute to and from work. Each group will analyse the data and produce summary outcomes on the volumes and patterns of these activities. Inferences will then be made, based on this information, to provide the “Physical Behaviour story”.
Finally, we will demonstrate novel methods of visualisation of Physical Behaviour.
- To understand how accelerometer data from body-worn sensors can be interpreted to classify periods of lying, cycling and car travel.
- To construct appropriate models of Physical Behaviour using the newly classified activities of cycling and car travel to address specific research questions.
- To use event-based analysis of lying, sitting, upright, walking, cycling and car travel activities to provide context rich information.
- To understand how different methods of data visualisation can be used to enhance our understanding of Physical Behaviour.
There will be three specific interactive features of this workshop.
- A structured discussion on how we can use event-based analysis to derive outcomes based on the patterns of activities.
- Interactive analysis of sample data to show how we can derive the new outcomes of lying, cycling and car travel from body-worn accelerometer data.
- Group working, groups of approximately four participants per group, to analyse sample data and the different methods of visualisation of Physical Behaviour data from different individuals/populations. The groups will produce “Physical Behaviour stories” from these sample data sets and will construct interventions to change behaviour based on this data.