Smart Watches Can Help Detect Parkinson’s Disease Earlier: Study - Decrypt
06/21/2024 18:49”Off-the-shelf” wearables linked to a smartphone app could be useful in measuring disease progression, according to new research.
A recent study found that mainstream, consumer-grade gadgets—specifically, older Apple Watches and iPhones—may help doctors detect signs of Parkinson’s disease earlier, accelerating care options for a neurological disorder that affects 10 million people worldwide.
To better equip healthcare providers with a way to identify and detect the progression of the disorder as early as possible, a team of researchers from the University of Rochester Medical Center conducted a 12-month study of 82 individuals with early-stage, untreated Parkinson’s and 50 other age-matched individuals as a control group.
The research results were published last week in npj Parkinon’s Disease, an open-access, peer-reviewed journal affiliated with Nature. It was an update to results reported last year.
The study states that the team wanted to use “accessible technology,” and they settled on the Apple Watch 4 or 5 and the iPhone 10 or 11. The study was conducted between June 2019 and December 2020, and the Apple hardware included models older than the top-of-the-line versions available at the time.
The smartwatches were used to gather health and fitness data for the BrainBaseline app installed on the iPhone. Participants simultaneously wore research-grade wearable “Opal” sensors placed on the sternum, lower back, and shoes.
The “off-the-shelf” wearables and the purpose-built Parkinson’s Disease app measured several factors, including walking speed, step distance, arm swing, speech patterns, finger tapping, as well as other inputs.
“We found numerous valuable digital measures derived from a commercial smartwatch and smartphone,” the study noted. “The response of these metrics to medications remains to be established. However, this study [brings] us closer to having meaningful digital measures for future use in Parkinson’s Disease clinical trials.”
The key study findings showed that walking problems got worse over a year in people with early Parkinson’s Disease. Other measured movements—such as arm swing, walking speed, and step length—decreased as well for the Parkinson’s Disease cohort.
These changes were measured using the wearables in real-world settings on a daily basis for each participant, and showed more pronounced differences than traditional doctor's tests, the study noted.
Study participants within the Parkinson’s group also had more hand tremors detected over time with the devices and app, showing that they experienced more time each day with tremors after a year. This change was more noticeable than what doctors usually measure, according to the published paper.
While there were some challenges cited that arose from the technology and timing limitations related to COVID as well, the researchers concluded that their premise was validated and warranted more research.
Consumer wearable and fitness technology have also advanced since the study was performed, with smart watches and smart rings adding numerous new sensors and gathering more metrics.
The study was led by Dr. Jamie Adams, associate professor at the University of Rochester’s Department of Neurology and Movement Disorders.
“These findings reinforce what other studies have shown,” Adams said in 2023. “Digital devices can differentiate between people with and without early Parkinson’s and are more sensitive than traditional rating scales for some measures of Parkinson’s disease.”
According to the Parkinson’s Foundation, roughly 90,000 Americans are diagnosed with the neurological disorder each year, which affects physical movement and is the second most common neurodegenerative condition behind Alzhiemer’s. While Parkinson’s currently has no cure, its symptoms can be subtle and tricky to diagnose.
Adams did not immediately respond to a request for comment from Decrypt.
Edited by Ryan Ozawa.
Generally Intelligent Newsletter
A weekly AI journey narrated by Gen, a generative AI model.