The article outlines the key points associated with the changes to software-based medical devices marketed in the US. 

The Food and Drug Administration (FDA or the Agency) has published a draft guidance document dedicated to a predetermined change control plan for Artificial Intelligence/Machine Learning (AI/ML)-enabled device software functions. The document is intended to provide recommendations to be considered in the context of marketing submissions, as well as additional clarifications to be taken into account by medical device manufacturers (software developers) in order to ensure compliance with the respective regulatory requirements. It is important to mention that FDA guidance documents are non-binding in their legal nature, nor are intended to introduce new rules or impose new obligations, but rather to assist in complying with the requirements set forth by the applicable legislation. The authority explicitly states that an alternative approach could be applied, provided such an approach is in line with the current legislation and has been agreed with the authority in advance. 

Regulatory Background 

First of all, the authority confirms its longstanding commitment to develop and apply innovative approaches to the regulation of medical device software and other digital health technologies to ensure their safety and effectiveness. The FDA acknowledges the increasing role of innovative technologies used in healthcare, including the wider use of products utilizing artificial intelligence in general and machine learning technologies (the authority further refers to as machine learning-enabled device software functions or ML-DSFs). The purpose of the present guidance is to describe the approach to be applied in the context of the development of medical devices based on novel technologies while ensuring their compliance with any and all applicable regulatory requirements in terms of safety, quality, and effectiveness. In this document, the FDA pays special attention to the regulatory matters related to the use of ML models trained by ML algorithms. 

The authority further acknowledges potential benefits for the healthcare industry in general from the wider use of ML-enabled technologies, especially when it takes to processing large volumes of data. According to the guidance, examples of ML applications in medicine include earlier disease detection and diagnosis, the development of personalized diagnostics and therapeutics, and the development of assistive functions to improve the use of devices with the goal of improving user and patient experience. 

The authority also recognizes that the process of ML-DSFs development is iterative in its nature. The purpose of the present guidelines is to introduce a least burdensome approach to be followed with respect to modifications to ML-enabled software while ensuring their continued safety and proper performance. In particular, the document pays special attention to the information the authority expects to be included in a Predetermined Change Control Plan (PCCP) being a part of a marketing submission for an ML-DSF. The said document should provide additional information regarding the expected modifications, as well as the appropriate methodology to be used by a medical device manufacturer when developing, implementing, and validating such modifications, and also the way the impact of such modifications will be assessed. The authority also mentions that the PCCP is reviewed as part of a marketing submission to ensure the continued safety and effectiveness of the device without necessitating additional marketing submissions for implementing each modification described in the PCCP. 

As the very first stage, in 2019 the authority has published a discussion paper dedicated to the matter, describing in detail the approach suggested by the FDA with respect to modifications to AI/ML-based products. The said approach has been further improved based on the new information becoming available to the FDA. The authority also takes into consideration the risk categorization principles developed by the International Medical Device Regulators Forum (IMDRF), a voluntary association of national regulating authorities in the sphere of medical devices collaborating for further improvement of the existing regulatory framework. 

The main feature making ML-based products different from other SaMD is their ability to learn through data instead of merely following the predefined approach. This is closely related to the ability to improve their performance through iterative modifications, including by learning from real-world data. These aspects are expected to be addressed in detail in a PCCP. According to the guidance, the latter includes the following key elements:

  1. A detailed description of the specific, planned device modifications;
  2. The associated methodology to develop, validate, and implement those modifications in a manner that ensures the continued safety and effectiveness of the device across relevant patient populations, referred to as the “Modification Protocol”; and 
  3. An Impact Assessment to describe the assessment of the benefits and risks of the planned modifications and risk mitigations. 

The aforementioned discussion paper received feedback from all the parties involved in operations with software-based medical devices. This includes general comments, as well as responses to 18 special questions raised by the authority. Moreover, the authority has conducted numerous public meetings and workshops dedicated to the matter. During these workshops, the authority has collected additional feedback and suggestions from the industry to be considered when developing a new regulatory approach addressing the specifics of ML-enabled software functions.

These days the authority receives an increasing number of marketing submissions associated with medical devices utilizing ML technologies and expects this trend to continue in the future. 

In summary, the present guidance describes the basics of the regulatory approach to software-based medical devices utilizing artificial intelligence/machine learning technologies. By virtue of the guidance, the authority highlights the specifics of this type of products, and explains the approach applied. 

How Can RegDesk Help?

RegDesk is a holistic Regulatory Information Management System that provides medical device and pharma companies with regulatory intelligence for over 120 markets worldwide. It can help you prepare and publish global applications, manage standards, run change assessments, and obtain real-time alerts on regulatory changes through a centralized platform. Our clients also have access to our network of over 4000 compliance experts worldwide to obtain verification on critical questions. Global expansion has never been this simple.