Wearable sensors can more accurately track Parkinson’s disease progression than traditional observation

A recent study from Oxford University suggests that by using a combination of wearable sensor data and machine learning algorithms, the progression of Parkinson's disease can be monitored more accurately than in traditional clinical observation.

Moreover, this approach may not only improve predictions about disease progression but also allow for more precise diagnoses.

The study

The Oxford study involved 74 patients with Parkinson's who were monitored for disease progression over a period of 18 months. The participants wore wearables with sensors in different regions of the body - on the chest, at the base of the spine, and on each wrist and foot.

These sensors — which had gyroscopic and accelerometric capabilities — kept tabs on 122 different physiological measurements and tracked the patients during walking and postural sway tests. Kinetic data was then analyzed by custom software programs using machine learning.

The sensor data collected by the wearables were compared to standard MDS-UPDRS assessments, which are currently considered the gold standard.

Whereas the traditional test "did not capture any change" in the study, the sensor-based analysis "detected a statistically significant progression of the motor symptoms," according to the researchers.

More precise data on the progression of Parkinson's isn't a cure, but the incorporation of metrics from wearables could help researchers confirm the efficacy of novel treatment options.

The context

Parkinson's disease is a neurological condition that affects motor control and movement. Although there is currently no cure, early intervention can help delay the progression of the disease in patients.

Diagnosing and tracking the progression of Parkinson's disease currently involves a neurologist using the Movement Disorder Society-Unified Parkinson's Disease Rating Scale (MDS-UPDRS) to assess the patient's motor symptoms by assigning scores to the performance of specific movements. However, because this is a subjective, human analysis, classification can be inaccurate.