Daniel K Jung

AI healthcare newsletter

Wearables, combined with ML, can now track Parkinson’s Disease Progression

Using Wearable Devices

Wearable devices offer the ability to track kinematic data of patients. The advantages of these devices is the capability to continue data collection outside the premises of a hospital, and the ability to have the data collection range span 24/7. 

The most famous examples of wearable devices for consumers are devices such as the fitbit, or the apple watch. These devices collect data about the wearer around the clock, and are used to provide information such as: step counts, sleep patterns, heart rates, etc. All of these endpoints are outputted with a sensor to collect the data, and a machine learning algorithm to decipher the data to meaningful results.

The researchers at University of Oxford aimed to establish whether they can utilize the kinematic data to estimate the clinical rating scale of Parkinson’s Disease progression, as well as monitor the motor symptom progression longitudinally.

Importance of Monitoring Progression

This novel way of monitoring the progression of Parkinson’s Disease is important for a couple key reasons: it allows clinicians to gain more insight in assessing whether treatment plans are working or not. And having a more robust methodology to log their motor symptoms allows for implications in clinical trials as well. 

Professor Chrystalina Antoniades mentioned how the current rating scale used to assess symptoms of Parkinson’s Disease has elements of subjectivity, and that different clinicians may provide different scores. As strongly as the field of medicine stands by Evidence-based Medicine, finding objective ways to quantify symptoms is an important part of removing variability from treating patients.

The future directions of Remote Monitoring Devices

Ultimately, I consider this as another ‘win’ for innovation in the Remote Monitoring Devices space. 

The traditional primary care model, that we still practice today, is having patients follow-up once every 3 months and receiving updates from them. Patients only come in earlier if ‘they’ believe something is wrong. We only get a ‘snapshot’ of what their quality of life is like only 4 times every year. 

By effectively utilizing the modern technology we have available for us today, we can change that. Along with seeing them once every 3 months, we can have remote monitoring devices, backed by machine learning algorithms that monitor the datapoints continuously, 24/7. 

“ML algorithms essentially serves as a ‘personal’ nurse that provides the ability to alert the physician if their datapoints look abnormal.”

We can now guide patients to come into the office for further evaluation, quicker, using our own clinical expertise, rather than having the patients decide to come in after it’s gotten severe.

The future of remote monitoring devices is bright, and I see it as the technology that will dynamically shift how the next generation of physicians (us) will be practicing.