FDA competition seeks digital endpoints for Parkinson’s

Miikka Luotio

The FDA is asking digital health technology developers to come up with a solution to measuring a common symptom of Parkinson’s disease that can have a serious impact on patients.

People with Parkinson’s sometimes experience temporary, involuntary periods where they are unable to move their feet when trying to walk – known as freezing of gait (FoG) – which can have a detrimental effect on balance and increase their risk of falling.

According to the Parkinson’s Foundation, freezing generally only lasts a few seconds, but it is one of the more frustrating and dangerous symptoms of the disease, with patients feeling like their feet are glued to the ground.

The exact cause of freezing is unknown, but it sometimes happens when a patient is due for the next dose of dopaminergic drugs, changing direction whilst walking, or transitioning from standing to walking.

To try to get a firmer understanding of the phenomenon, the FDA will launch the Digitally-Derived Endpoints for Freezing-of-Gait Detection (DEFoGD) Challenge later this month, looking for artificial intelligence/machine learning (AI/ML) models that may be able to identify and predict FoG events in Parkinson’s patients.

Kicking off on 28th May, the competition – aimed at medical device developers, tech innovators, software developers, and academic researchers – is looking for digital health innovations that may speed up the development of new and effective approaches to helping patients with the disease.

It’s an important project, as around 38% of Parkinson’s patients have a fall every year, and recurrent falls are an indicator that the disease is progressing, as well as impacting their quality of life and raising the risk of serious injury or death.

The competition – on which more details are available here – is being run by the FDA’s Center for Devices and Radiological Health (CDRH) and the Office of Digital, and ties in with ongoing efforts by the regulator to find and validate digital endpoints that can be used to study diseases and potential new therapies.

The competition will be hosted on precisionFDA, a secure computing platform that connects researchers from around the world and provides them with a library of cloud-based tools to interact with biological datasets, in the name of advancing precision medicine.

In the first stages of the competition, contestants will work on publicly available datasets or their own data to develop the AI models, and the best-performing models will then be evaluated against validated and curated FoG data within the FDA in a second phase, expected to run between September and November 2024.

Photo by Miikka Luotio on Unsplash