Introduction

The regulation of artificial intelligence (AI) medical devices has become a pressing issue. The US Food and Drug Administration (FDA) has adopted a firm-based approach for regulating lower-risk medical devices. This method focuses on the quality systems of the manufacturing firm rather than the mechanics of individual products. A recent publication in JAMA Forum discusses how this approach can be applied to AI medical devices.

The Firm-Based Approach to Regulation

The FDA’s firm-based approach shifts regulatory focus from the product itself to the company’s overall development process. This method emphasises the importance of a robust quality management system. By ensuring consistent production and control, the FDA aims to enhance the safety and effectiveness of medical devices. Routine inspections and audits verify the firm’s procedures and operational controls. This approach ensures better protection for public health.

Legislation and AI Medical Devices Regulation

The VALID Act (Senate Bill 2209 and House Bill 4128) is currently before Congress. It codifies the firm-based approach to regulation and enables the FDA to oversee the methods used to develop and validate AI technology. The AI used in medical devices analyses vast amounts of data to generate clinical insights. These devices undergo constant modification as new information becomes available. The VALID Act aims to modernise the FDA’s oversight of in vitro diagnostics and strike the right balance for the agency’s regulation of laboratory-developed tests.

Challenges and Risks

AI models have evolved to manage certain patient interactions directly. For instance, ChatGPT has demonstrated substantial semantic medical knowledge and the ability to perform medical reasoning. However, these models involve complex layers of computations and extensive data processing. This results in predictions with typically opaque reasoning pathways. One risk is that these systems can return wrong answers, especially when trained on unreliable datasets. The FDA cannot fully comprehend how AI medical devices generate every possible answer.

An example of the firm-based approach’s effectiveness can be seen with the Apple Watch. In 2018, the FDA cleared the Apple Watch for detecting irregular heart rhythms. Apple provided a substantial amount of data proving its reliability in real-world settings. This validation shifted the regulatory focus from dissecting the product’s hardware to evaluating the company’s overall development process. The success of the Apple Watch clearance illustrates the potential benefits of applying a firm-based approach to AI medical devices.

Future Directions and Reforms

Under a firm-based approach, innovators could bring new products to market and update existing ones without undergoing the same premarket review in every case. This modern approach would make the introduction of new innovations far more efficient. Additional reforms could establish a framework for third-party certification of some lower-risk AI devices. Independent organisations meeting criteria set by the FDA would assist in evaluating and certifying these devices. This flexibility would enable the agency to tap into external expertise and streamline the FDA’s review process.

Conclusion

To unlock the potential of AI medical devices for patients, regulatory policies need to address their unique characteristics. The FDA’s traditional regulatory approach might prove infeasible in this context. Congress must enact new laws to codify these modern concepts that are essential for regulating this new technology.