Sensor-Based Algorithm Detects Patient Response to Parkinson Therapies

Disabling motor features including tremor, reduced speed, and gait/balance impairment are some of the most prevalent complications of patients with Parkinson disease (PD). PD, a progressive neurological disorder, affects nearly 6 million people globally, a number that is expected to double by 2040.

Further complications in patients with PD are seen when they enter an akinetic (medication OFF) and mobile phases (medication ON), demonstrated in 50% of patients diagnosed within 3-5 years, and 80% of patients diagnosed within 10 years. These fluctuations in motor function present a critical point in terms of managing the disease because it requires continued adjustments in treatment, such as changing the frequency and dosage amount or parameters for deep brain stimulation.

At present, PD motor fluctuations are addressed and identified by conducting clinical examinations or through self-reports from patients, although self-reports can be unreliable. A recent study looked to develop a technology-based system to “provide reliable information about the duration in different medication phases that can be used by the treating physician to successfully adjust therapy.”

Researchers combined an algorithm and sensor-based system that was able to detect medication ON and OFF states in patients with PD using 2 wearable sensors that were placed on the patients most affected wrist and ankle.

The sensors collected movement signals while patients performed 7 daily living activities such as walking or getting dressed while actively in their medication ON and OFF phases. The algorithm was trained using approximately 15% of the data from 4 activities and subsequently tested on the remaining data.

“Our approach is novel because it is customized to each patient rather than a ‘one-size-fits-all’ approach and can continuously detect and report medication ON and OFF states as patients perform daily routine activities,” said Behnaz Ghoraani, PhD, senior author and assistant professor at Florida Atlantic University’s College of Engineering and Computer Science in a statement.

Researchers found that the algorithm was able to detect the response to medication during the participants’ daily activities with an average of 90.5% accuracy, 94.2% sensitivity, and 85.4% specificity. The goal was to be develop a system that can be individualized and trained using collected data during a patient’s first clinical visit.

“Once the algorithm is trained it can readily be used as a passive system to monitor medication fluctuations without relying on patient or physician engagement,” said Ghoraani.

The study authors were able to conclude that the sensor-based algorithm could provide objective and accurate assessments of medication states, which in turn could lead to successful adjustment of each individual’s therapy, and therefore improve care delivery and quality of life of patients with PD.

Reference

Hssayeni M, Burack M, Jimenez-Shahed J, Ghoraani B. Assessment of response to medication in individuals with Parkinson’s disease [published online March 12, 2019]. Med Eng Phys. doi.org/10.1016/j.medengphy.2019.03.002