In Vitro Diagnostics Become Increasingly “Smarter” Under EUA

IVD submissions incorporating artificial intelligence and smartphone-based apps to combat COVID-19

May 27, 2021

Since March 2020, the U.S. Food & Drug Administration (FDA) has been holding weekly town hall-style meetings where the in vitro diagnostics (IVD) community is invited to ask clarifying questions about the emergency use authorization (EUA) process for SARS-CoV-2 diagnostic tests. In recent meetings, the conversation has included elements relevant to developers working with supportive technology, either in the form of artificial intelligence algorithms or smartphone-based apps. Key takeaways from the FDA’s guidance and stance on submissions that include these technical elements can be found below:

  • FDA welcomes artificial intelligence algorithms in a number of different applications, but these algorithms must be verified and validated before authorization. This includes providing a change plan detailing how the sponsor will assess and demonstrate the appropriateness of any modifications the algorithm is expected to undergo after it is placed on the market as well as change protocols to ensure continued satisfactory performance. The change plan is valuable because it ensures any modifications it covers will not impact the safety, effectiveness, or performance of the product. With the change plan in place, FDA may decide that the changes are adequately assessed by the change protocols and potentially that no interaction beyond communicating the intent to modify and the resulting performance metrics is necessary before enacting the modification on market. Change plans can also help to identify those modifications that may require amendments or supplements to the EUA submission to FDA for review before they can be enacted. Exponent has experience crafting these sorts of documents and supporting efforts to satisfy this guidance.
  • Smartphone-based apps are expected to monitor the impact of operating system updates on the models supported by the developer. Developers should use the change plan to mitigate potential issues by including app modifications targeting incompatibility and obsolescence in their submissions.
  • More information on software change plans and protocols can be found in Chapter 6 of Software Engineering: Artificial Intelligence, Compliance, and Security, a reference guide written by Exponent consultants and their collaborators.

More information:


How Exponent Can Help

Did you know approximately 66% of serological test EUA submissions independently reviewed by the National Cancer Institute (NCI) for the FDA are ultimately declined? The IVD industry is shifting towards sophisticated software to combat the pandemic, but FDA regulations in areas such as artificial intelligence are still in their infancy. Developers are increasingly reliant on those with significant software experience to properly design the necessary verification and validation plans for software authorization. With the low success rate and the long wait times currently stifling developers’ path to market, it is more important than ever to get your submission right the first time.

Exponent’s multi-disciplinary team of software, in vitro diagnostic, medical device, and pharmaceutical industry professionals develop safety information for products and processes that comply with pertinent standards, regulations, and best practices. With expertise in Software as a Medical Device (SaMD) regulations, artificial intelligence, CLIA compliance, quality system establishment and FDA assessments, Exponent can assist in reviewing submission packages, auditing labs, or performing due diligence and validation/technical design of the assay, instrument, or collection kit. Our offerings also include reviewing diagnostic data to assess the completeness of the package for submission according to any of the EUA pathways or 510(k)/De Novo submissions.

AUTHORS