Our Projects

Problems with everyday mobility, such as walking or driving, are common in older adulthood and can negatively impact health and social functioning. This program of research is focused on using recent advancements in technology to understand early changes in older peoples’ mobility and to identify those who will benefit from further follow-up and early preventative treatment. We will complete four interrelated research projects over the next five years that will address issues specific to early mobility limitation in older Canadians. The first project will focus on understanding the impact of changes in mobility on an older persons’ level of functioning, including consensus on how to best define and measure early mobility problems. The second project will use machine learning techniques applied to data from the Canadian Longitudinal Study on Aging to find the most relevant predictors of early mobility problems. In the third project, we will customize a tracking device (i.e., wristwatch) that is capable of monitoring many different aspects of mobility in the home and community. In a group of 1000 older adults, we will use machine learning alongside other methods to assess the ability of this technology to identify people with differing levels of mobility problems and trajectories. In the fourth and final research project, we will use our findings from the first three research projects to begin the development of a mobility self-monitoring tool for older adults and their caregivers. Ultimately, with this tool, we hope to help older people by preventing or delaying late-life mobility problems through early detection and management.

The main aims of this project are to investigate: 1) the reliability and validity of commercially available activity trackers in older adults 2) confirm accuracy of selected GPS trackers in healthy volunteers 3) user testing of wearables in older adults and 4) feasibility of wearable device protocol for future cohort study. Results will provide guidance in choosing the mobility tracking system for use in a large cohort study of older Canadians.

We will be using interviews and ‘walk-alongs’*  to explore early changes in mobility among older adults.  The results of this study will inform the metrics and outcome measures to be used in the broader research program (CLSA analysis and cohort study) examining trajectories of mobility limitation among older adults. *The walk-along approach is used here, although the methods are expected to be refined depending on level of mobility.

An online consensus exercise will be used to determine how to operationalize and measure early mobility limitation with respect to both physical function and transportation related mobility. This will consist of an online real time consensus meeting with experts in mobility, including both researchers and senior clinicians. The process will be facilitated by a moderator (Queen’s University Executive Decision Centre). This approach differs from a traditional Delphi approach in that participants’ responses are not anonymized, which has the advantage of gaining an expedited consensus. Activities in this project include: conducting a rapid review on measurement of early mobility problems, preparation of material for the online exercise, conducting and writing up the online exercise results. Findings from these projects will inform both the operationalization of measures in the CLSA analysis and the cohort study.

Machine learning (ML) and computational statistics (CS) approaches will be used to analyze data from the Canadian Longitudinal Study on Aging (CLSA) in order to derive and validate analytical models that predict the onset and progression of early mobility limitation among community-dwelling older adults. In our study, it is important to understand how the predictions are made and, more importantly, to understand how the most important variables contribute to the predictions. This is crucial inter alia for communicating results with our SAC panel and other experts to ensure the clinical significance of predictions. Therefore, in our analyses, we will use the most important variables identified at the ML step to build CS models. CS approaches have arisen relatively recently and are enabled by the extraordinary computational power now available. Furthermore, our SAC will be consulted during this entire process to ensure the clinical significance of the results remains front-of-mind.

A multidisciplinary stakeholder advisory group (older adults, clinicians, policy makers) will collaborate with the research team to provide input throughout the entire research process.Members of the SAC will gain knowledge and familiarity with areas of early mobility limitation, machine learning, and technology, and members of the investigative team will gain valuable experience working with and developing relationships with a diverse group of stakeholders. The first step is developing a framework to guide these activities.