At the a2 National Symposium: Empowering Innovation in AI/Tech + Aging, Senior Research Scientist Jen Blankenship, PhD, presented our ongoing work with the Center for Human Health and Performance at the University of Massachusetts Amherst to develop digital clinical measures that capture how patients with Alzheimer’s disease function in their real-world environments.
Read more about the pilot study below and download the poster.
Alzheimer’s disease is a leading cause of death, with very few treatments to cure or slow disease progression. Aggressive research efforts are underway to develop disease-modifying drugs for AD. However, identifying efficacy endpoints that are meaningful to patients, clinically relevant, and sensitive to treatments is a current challenge.
The Unmet Need
Preserving functional independence during disease progression is a major goal for patients with Alzheimer’s disease. Most clinical trials capture functional independence with established patient- or clinician-reported outcome assessments. However, there are inherent limitations to these assessments:
- They are burdensome to patients
- They are captured episodically
- They do not comprehensively capture aspects of health that matter to patients
Wearable sensors have the potential to substantially advance the way function is assessed by directly capturing how patients behave in their own environments. Given that walking is a central aspect of functional independence, real-world digital measures of walking behavior could enable holistic assessments of Alzheimer’s disease. Existing algorithms to derive measures of walking from wearable inertial sensors are developed using data from healthy, younger populations. There is a need to develop algorithms to derive measures of walking behavior in populations with a slow and altered gait, such as older adults with Alzheimer’s disease.
Pilot Study Goals
To address the existing evidence gaps, VivoSense is launching a pilot study funded by the Massachusetts AI & Technology Center for Connected Care in Alzheimer’s Disease (MassAITC) and the National Institute on Aging (NIA). The purpose of this pilot study is to develop machine learning algorithms to derive features of gait and walking behavior from wearable inertial sensors in patients with Alzheimer’s disease.
Pilot Study Methods
In this study, 30 older adults (≥ 65 years) with and without mild Alzheimer’s disease will perform a variety of walking tasks while wearing sensors at several body locations in a laboratory setting. Motion capture systems will collect the truth data needed to develop and validate machine learning algorithms that derive measures of real-world gait and walking behavior in Alzheimer’s disease. To establish real-world feasibility and acceptability, 2-weeks of at-home monitoring with wearable activity monitors will assess how patients function at home.
VivoSense’s long-term goal is to qualify a real-world digital measure of functioning for patients with Alzheimer’s disease with the FDA. Accurate and reliable measurement of real-world walking behavior is a critical first step towards developing a digital measure of functional independence for patients with Alzheimer’s disease.
Funding Acknowledgements: National Institute on Aging grant P30AG073107