Pre-Conference Workshops will include a coffee break. Lunch is not included.
Measuring error without error
It is imperative for all measurement instruments to be reliable (provide consistent measurements) and valid (measure what it is meant to measure). Lack of reliability and validity can lead to incorrect conclusions, not replicable research findings and the impossibility, for everyday users of the measurement instrument, to assess changes in behavioural and physical information. This workshop aims to equip researchers studying physical behaviour with the statistical knowledge necessary to plan and perform reliability, agreement and validity studies of measurement instruments providing quantitative information, such as performance tests. At the end of the workshop, participants should be able to 1. make the distinction between the concepts of reliability, agreement and validity. These words are too often used interchangeably in the literature leading to confusion and inappropriate statistical analyses; 2. plan a reliability, agreement or validity study given practical constraints (e.g. number of subjects), particularities of the measurement instrument and study aims; 3. determine which statistical measure(s) is (are) the most appropriate depending on the aim of the study and the study design; 4. interpret the results of the study, communicate the results to an audience with minimum statistical knowledge, and acknowledge the limitations of the statistical approach chosen.
Intra-abdominal pressure and daily living: How a simple measurement can provide insight into physical behaviour and health
Janet Shaw, University of Utah – Kinesiology
Monika Leitner, Bern University of Applied Sciences
Robert Hitchcock, University of Utah
Intra-abdominal pressure (IAP) provides insight into physical behaviors associated with trunk muscle activity and ventilatory patterns. Recent advancements in measurement have allowed for comprehensive assessment of IAP during human movement. Given the specialized nature of investigating IAP, however, its role in health and disease is not broadly understood. The overall goal of this workshop is to introduce participants to the physiologic measure of IAP, which has been linked to physical activity levels and types as well as to the clinical presentation of pelvic floor dysfunction, most notably urinary incontinence. The presenters incorporate knowledge and perspectives from exercise physiology, physical therapy and rehabilitation, and bioengineering. They will present their experience with developing sensor technology, measuring IAP in a variety of activities and settings, analyzing and interpreting IAP data, and using knowledge of IAP when conducting clinical work in pelvic floor health.
We need to talk. The absolute basics of getting your device data and other date-time stamped data sources to talk to each other
Elisabeth Winkler, University of Queensland
Philippa Dall, Glasgow Caledonian University
Matthew Buman, Arizona State University
The workshop goals and objectives are:
- To empower device users with the basic programming skills and confidence to generate new, meaningful measures by combining their detailed device data with other date-time stamped information (e.g., another device, diary-reported sleep/wake times; smartphone data)
- To achieve this in a manner that is clearly applicable to a variety of devices and other date-time stamped information (e.g., activPAL, actiGraph, smartphone data, diary?)
- To provide attendees with take-home materials they can use with their own data on their own timeframe in the languages and formats of common software packages that they may be using currently or in the future (e.g., Excel, MS Power BI, SAS, STATA, R?)
Creating context specific accelerometer-based measures
Jasper Schipperijn, University of Southern Denmark
Tom Stewart, Auckland University of Technology
Determining the context in which physical activity or sedentary behavior occurs is important. For example, recent studies have shown that having a job with a high amount of MVPA is related to negative health outcomes, whereas more minutes of MVPA during leisure time or transport are associated with positive health outcomes.
Colleting contextual information using diaries, observations, global positioning systems (GPS) or other devices, and combining it with accelerometer measures, can provide context specific accelerometer-based measures. Recent methodological and technical developments have made it possible to combine contextual and accelerometer data efficiently and create the needed context specific accelerometer-based measures.
The goal of this workshop is to teach participants how to collect contextual information in their accelerometer studies and demonstrate tools that can help create context specific accelerometer-based measures.
Beyond daily totals, using novel analysis of raw-accelerometer data to generate clinically important outcomes of physical behaviour in free-living populations
Andrew Kerr, University of Strathclyde
Kate Lyden, PAL Technologies Ltd
Nicholas Smith, PAL Technologies Ltd
The goal of this workshop is to provoke discussion and reflection on the importance of objective measures of free-living physical behaviour for clinical research and patient focussed decision making. Our objectives are to cause attendees to reflect that accelerometers can be both very effective in measuring how little someone is doing (sitting and sleep) and also how much (physical activity). Looking beyond the commonly reported daily totals for physical activity and sedentary behaviour we will explore the challenges around estimating time in bed, its use as a proxy for sleeping time and how self-report of sleep time impacts on measures of sedentary behaviour. We will reflect on how acceleration profiles of time in bed vary across the lifespan and what might be clinically interesting about these profiles. Body-worn accelerometers are well suited to the quantification of stepping, the major component of daily physical activity. It is commonly accepted that subjects have preferred walking intensities (cadence) but how important is the ability to vary cadence for both community and household stepping tasks? We will look at the patterns of step accumulation and how this varies both within and between individuals. A major challenge in using accelerometers with some clinical populations is the difficulty in quantifying slow stepping and the relative importance of these steps in allowing independent participation in self-care activities. In some clinical populations those household stepping bouts associated with self-care may be the hardest to measure but also the most clinically important.
Just D.O. It. Using Video-based Direct Observation to Assess Physical Behaviour
John Sirard, University of Massachusetts Amherst
Sarah Keadle, California Polytechnic State University
David Bassett, University of Tennessee Knoxville
Scott Strath, University of Wisconsin Milwaukee
The Goal of this workshop is to develop a preliminary consensus document describing best practices and procedures for performing video-based direct observation
The specific Objectives are to:
- Expose attendees to different DO systems currently in use through brief formal presentations and documentation (handouts, electronic docs)
- Provide video-examples of coding scenarios including clear cut transitions and more ambiguous behaviors/postures/intensities/steps. Have participants code select portions of videos.
- Identify important issues for accurate and reliable coding in different populations (children, healthy adults, functionally limited/older adults).
- Identify primary outcome variables (or metrics) that could be consistent across systems.
- Discuss key challenges and limitations of DO coding
How good are our criterion measures for accelerometer developments and validations?
Kong Chen, National Institute of Diabetes and Digestive and Kidney Disease
Scott Crouter, University of Tennessee
Guy Plasqui, Maastricht University
Sarah Kozey Keadle, California Polytechnic State University
Accelerometers are widely used to predict energy expenditure and/or physical activity types. We often focus on accuracy and precision of accelerometers for such predictions, but many users do not understand how these two parameters are obtained. During the development of these accelerometers, from the hardware to prediction models, and when they are validated, “criterion measures” are used. Methodologies commonly used as “gold-standards” are indirect calorimetry and direct observations. How accurate and precise are these standards and how are they determined (what do they rely on as the criteria), and how these could impact the accelerometry predictions? In this symposium, each presenter will highlight how criterion measures are calibrated and validated, and how they should be optimized for accelerometry studies, with focus on their advantages and limitations. The overall goal is for the audience to have a better understand these so-call “gold-standards” when choosing to use them for their own development or validation studies.
AlPHABET: taxonomy of daily physical behaviours, consensus phase 2
Sebastien Chastin, Glasgow Caledonian University, Ghent University
Jorunn Helbostad, Norwegian University of Science and Technology
Andreas Holtermann, National Research Centre for the Working Environment
Matthew Buman, Arizona State University
AlPHABET is an open science project set up to develop a common taxonomy (naming and cataloguing) for classification, harmonisation and storage of objective tracking sensor data of human physical behaviour in daily life. The development will be through an international consensus process. Study and monitoring of daily physical behaviour is a fast emerging science which benefits from the growth in data science and wearable technologies. However, it is limited the lack of shared, standardised and integrated classification system to ascribe meaning to digital data. Each discipline has developed different ontologies and classifications resulting in difficulties in comparing, harmonising and combining knowledge and evidence. Significant progress will be sped up if we can develop a consistent way of describing and classifying human daily activities, physical actions and movements. The first step in developing a universal taxonomy took place at ICAMPAM 2017. This workshop aims at continuing this effort. It will take the form of a debate and interactive session together with a digital consensus process. Report on the current status of the project will be given. This will be followed by new perspectives.
Application of actigraphy and wearable devices in pharmaceutical research
Jaiwei Bai, Johns Hopkins University
Vittorio Illiano, Novartis Pharma AG
Eve Pickering, Pfizer, Inc
In recent years, various types of actigraphy or wearable devices have been deployed in both observational and clinical studies. While such devices are becoming more and more mainstream in academic studies, they remain rare in the pipeline of pharmaceutical research. Many pharmaceutical companies are working toward supporting label claims with actigraphy, building the necessary expertise via their own pilot studies or in collaboration with universities. The session highlights three of the actigraphy-related challenges in clinical research: how to sensitively detect longitudinal behavior change, how to identify whether the accelerometer was worn by the same person throughout the study, and how to replace gold standard assessments with actigraphy-derived endpoints.
Lab Tours will be comprised of touring three areas: Metabolic Research Unit, CAREN, Wearable Healthcare. Groups of 15 people will spend 30 minutes at a time in each area before moving on to other areas. Space is limited so book early!