FDA’s industry ties scrutinized in investigation of AI medical devices

19 Oct 2020 --- Researchers are calling for the US Food and Drug Administration (FDA) to be more transparent about its industry ties. A cross-sectional study found that parties with financial ties to industry made nearly two-thirds of the comments on the FDA’s proposed regulatory framework for artificial intelligence/machine learning (AI/ML)-based software as a medical device.

“The findings add to an already large body of literature examining the involvement of industry in the healthcare system,” James Smith, lead author and postdoctoral scientist at the Nuffield department of orthopedics, rheumatology and musculoskeletal sciences, University of Oxford, tells NutritionInsight. 

“Two things surprised me in this case: the lack of transparency about industry ties – there is no requirement to disclose them – and the lack of input from academics. This is despite AI/ML medical devices, and medical device regulation more broadly, being very active and hotly debated areas of research,” he adds. 

Industry ties in moderation
However, he stresses that industry ties are not necessarily problematic. Notably, the paper did not find evidence that industry ties have resulted in biased commenting. 

“Industry has an important role to play in healthcare, so I’m not suggesting that it shouldn’t be involved in these discussions. However, it is important to understand the perspective that input is coming from to be able to evaluate potential biases.”

Therefore, Smith would like to see the FDA request a conflict of interest statement with each comment. He also hopes that more scientific evidence will be used. 

“The majority of commenters do not cite scientific evidence, in particular systematic reviews, which are generally considered to be high-quality evidence.”

Finally, he believes that more input from parties that evaluate and use, rather than develop, healthcare technologies could be valuable. This might include academics or people involved in health technology assessment, for example.

“I’m not sure why there is so little academic input, but it may be as simple as a lack of awareness that the FDA seeks input during the development of regulations.”The FDA had proposed a new regulatory framework for modifications to AI/ML and asked for feedback from the public to refine the regulations.

The role of AI and ML
Players across the healthcare space – including in nutrition – are harnessing the new insights that AI and ML can provide. 

Recent examples of industry embracing tech include Abbott reimagining glucose monitoring for the sports nutrition market and Atmo Bioscience’s ingestible gas-sensing capsules that provide real-time insights into gut health and microbiome function.

At the start of the year, Abbott also flagged how implanted medical devices are set to dominate as people look to have increasingly personalized and timely data about their own bodies. 

However, the suitability of traditional medical device regulatory pathways for AI/ML has been called into question because the nature of the technology means it is continually evolving and adapting to improve performance.

Proposals for new regulatory framework
Under the FDA current framework, devices require further review and approval as they evolve. This can be time-consuming and may affect patient safety and interests. 

The FDA, therefore, proposed a new regulatory framework for modifications to AI/ML and asked for feedback from the public to refine the regulations.

Of 125 publicly available comments on the FDA proposal between April 2 2019 to August 8 2019, 63 percent were from parties with financial ties to industry in the sector. For a further 29 percent of comments, the presence or absence of financial ties could not be confirmed. 

The vast majority of submitted comments (86 percent) did not cite any scientific literature, with only 4 percent citing a systematic review or meta-analysis.

Ties within nutrition
While Smith’s investigation only investigated AI/ML medical devices, he points to some previous research looking specifically at the nutritional realm. 

For example, a 2013 study examined whether financial industry funding or the disclosure of potential conflicts of interest influenced the results of published systematic reviews conducted in the field of sugar-sweetened beverages (SSBs) and weight gain or obesity.

The researchers found that systematic reviews with conflicts of interest were five times more likely to present a conclusion of no positive association between SSBs and weight gain than those without them. 

“Industry can have unique subject matter expertise that it is very useful to have access to. Transparency is the important thing. Without knowing who has ties to industry, it is very difficult to evaluate the impact of those ties,” concludes Smith. 

By Katherine Durrell