Wearable Devices Help Assess Friedreich's Ataxia Severity Better Than In-Clinic Assessments

Sensor derived metrics demonstrated accurate measurements of disease severity and motor dysfunction in Friedreich's ataxia (FRDA) according to a study published in Communications Medicine. Furthermore, incorporation of the metrics into machine learning models improved predictions of disease severity.

FRDA is a neurodegenerative disorder characterized by progressive loss of coordination and movement due to frataxin protein deficiency. Traditional clinical assessments, such as the modified Friedreich's Ataxia Rating Scale (mFARS), are subjective and often insensitive to subtle changes in the condition, leading to challenges in accurately monitoring disease progression and evaluating therapeutic interventions. The Friedreich Ataxia Rating Scale Activity of Daily Living (FA-ADL) has also been documented to be insensitive to change and the scale is not immune to subjectivity.

Perhaps more importantly, the limitations of these tests impact clinical trial outcomes.The investigatory drug candidates may be well tolerated and achieve high response rates in the testing cohort, but the evaluated endpoints are influenced by mFARS, and its shortcomings may mean the difference between FDA approval and rejection.

Ram Mishra, Ph.D.

Because of the shortcomings of the clinical assessments, Ram Mishra, Ph.D., a senior researcher at BioSensics, and team investigated the utility of wearable technology in monitoring the symptoms and disease progression of FRDA. The team sought to establish the clinical value of real-time data from wearable sensors for FRDA patients and provide a more effective method for an objective overall disease assessment.

The study enrolled 39 ambulatory patients with FRDA. Their physical activity and upper limb function were monitored over a seven-day period using wearable sensors. The researchers collected data on lower and upper extremity functions during daily activities and correlated these sensor-derived metrics with established clinical measures (such as mFARS and FA-ADL) and biological biomarkers (GAA and FXN levels). The biological markers were sourced from clinical charts of the participants.

The PAMSys pendant and PAMSys ULM sensors provided a wide range of real-time data for analysis.The sensors measured posture, sitting, standing, walking, locomotion and postural transitions, such as moving from sitting to standing. For research purposes, the study’s authors only considered days when the sensors were worn by the subjects for at least 18 hours.

Statistical analyses revealed significant correlations with moderate to high effect sizes between the sensor-derived metrics and both clinical FRDA measures and biological outcomes. Notably, the team employed several machine learning (artificial intelligence) models to predict disease severity, which showed that incorporating wearable sensor data improved model performance in terms of predictive accuracy. The results indicated that wearable sensors provide reliable data that correlate significantly with existing clinical assessments, thus enhancing the understanding of FRDA.

The findings of this study highlight the potential of at-home wearable monitoring as a valuable tool for assessing disease severity and monitoring motor dysfunction in patients with FRDA. The ability to gather continuous, objective data on physical activity and upper limb movements allows for a more comprehensive understanding of patients' daily functional capacities, which is crucial for management and therapeutic evaluation. This approach addresses the limitations of traditional clinical assessments that often capture a narrow snapshot of a patient’s condition.

Incorporating machine learning models into the analysis further enhances the predictive power regarding disease progression, suggesting that such technologies can assist in personalizing patient care. The clinical validity of using these wearable devices could enhance how FRDA and, potentially, other neurodegenerative diseases are monitored and treated.

Future research could expand on these findings by examining the long-term benefits of continuous monitoring and validating the results in larger, more diverse patient populations.