Online AI test outperforms primary care physicians in assessing Parkinson’s disease

A test of a new AI tool that can assess the severity of Parkinson’s disease showed that it outperforms primary care doctors with Unified Parkinson Disease Rating Scale certification, but did not outperform neurologists, who fared slightly better than AI.

AI outperforms primary care physicians: ©PaeGAG - stock.adobe.com

Developed by researchers at the University of Rochester, the AI tool allows individuals with Parkinson's disease to evaluate their symptoms within minutes through a simple finger-tapping test. Researchers say this innovation holds great promise in enhancing patient care and accessibility. The study appeared in npj Digital Medicine.

Patients can participate from their homes, tapping their fingers ten times in front of a webcam to assess their motor performance on a scale of 0 to 4.

Traditionally, doctors have relied on patients performing basic motor tasks to assess movement disorders and rate the severity using guidelines like the MDS-UPDRS scale. This AI model streamlines the process by rapidly generating computational metrics such as speed, amplitude, frequency, and period. These metrics adhere to medical guidebooks, ensuring consistency, standardization, and interpretability in assessing the severity of tremors.

The study enlisted 250 global participants diagnosed with Parkinson's disease to perform the finger-tapping task. While the expert neurologists outperformed the AI model, it significantly outperformed primary care physicians, highlighting its potential in primary care settings, researchers said.

The researchers noted that this is an example of how patients who have trouble accessing a neurologist, securing appointments, or traveling to appointments, can use the AI tool from home and still receive quality care, improving health care access and equity.

The researchers believe that their AI-based method can be adapted for other motor tasks, potentially facilitating evaluations of various movement disorders, such as ataxia and Huntington's disease.