FDA starts taking steps towards a new, tailored regulatory framework for artificial intelligence-based medical devices

The Food and Drug Administration (FDA) have released a discussion paper on their proposed regulatory framework for artificial intelligence-based medical devices. The FDA consultation is open until 3rd June 2019.

Under the FDA’s existing guidance, they predict that many changes made to software as medical device driven by artificial intelligence and machine learning would be subject to a pre-market review, thus prompting the FDA to re-imagine a regulatory approach for these devices. The proposed framework may enable the FDA and manufacturers to evaluate and monitor a software product from its pre-market development to post-market performance.

As technology and science advance, with artificial intelligence and machine learning we can expect to see earlier disease detection, greater accuracy in diagnosis, more targeted therapies and improvements in personalised medicine.

Artificial intelligence algorithms are software that can learn from and act on data. These algorithms are already being used to aid in screening for diseases and to provide treatment recommendations, for example the FDA have authorised an artificial intelligence based device for detecting diabetic retinopathy. The FDA have also authorised an artificial intelligence based device for alerting providers of a potential stroke in patients.

The FDA have stated that the goal of the framework is to ensure that ongoing algorithm changes follow pre-specified performance objectives and change control plans, use a validation process that guarantees improvements to the performance, and safety and effectiveness of the artificial intelligence software. A further aim is to include real-world monitoring of performance once the device is on the market to ensure safety and effectiveness are maintained.

To read the FDA’s proposed regulatory framework for modifications to artificial intelligence/machine learning based on software as a medical device, please click here.