Artificial Intelligence (AI) and Software as Medical Device
Artificial Intelligence and Medical Devices
With all of the discussion around artificial intelligence (AI) and its effects on the pharmaceutical and medical device industry, it is good to understand the thinking of the FDA on this topic.
AI is an exciting field but is still very new to most in the industry. It is the same for the FDA. They have published several guidance documents and initial thinking on the subject. Please see below or an overview.
The FDA recognizes, as documented on their website, that AI and machine learning have the potential to transform healthcare. Innovative medical device manufacturers are using AI to innovate products to better assist providers and improve patient care.
The FDA medical device regulations were not originally written or designed for adaptive artificial intelligence thus the FDA anticipates that many of the AI software will need premarket review. Recognizing that the adaptive nature of the new software as medical device (SaMD) requires a new total product lifecycle (TPLC) regulatory approach that allows these devices to continually improve while providing effective safeguards.
In April of 2019 the FDA published a discussion paper “Proposed Regulatory Framework for Modifications to Artificial Intelligence/Machine Learning (AI/ML)-Based Software as a Medical Device (SaMD) – Discussion Paper and requested feedback.
The discussion paper proposes a framework for modifications to AI/ML-based SaMD that is based on the internationally harmonized International Medical Device Regulators Forum (IMDRF) risk categorization principles, FDA’s benefit-risk framework, risk management principles in the software modifications guidance, and the organization-based Total Product Lifecycle Approach (TPLC) as envisioned in the Digital Health Software Precertification (Pre-Cert Program).
The proposed TPLC Approach:
- FDA regulatory oversight embracing the iterative improvement power of AI while assuring product safety.
- Assuring that ongoing algorithm changes are implemented according to pre-specified performance objectives.
- Assuring that algorithm changes follow defined change protocols
- Assuring that the validation process is committed to improving the performance, safety and effectiveness of AI/ML software.
- Assuring that real-world monitoring of performance.
The TPLC regulatory framework’s aim is to promote a mechanism for manufacturers to be continually vigilant in maintaining the safety and effectiveness of their SaMD to bring increased benefits to patients and providers.
AI/ML-based software, when intended to treat, diagnose, cure, mitigate, or prevent disease or other conditions, are considered medical devices. The intended use of the AI/ML based SaMD, may exist on a spectrum of impact to patients as categorized by IMDRF SaMD Risk Categorization Framework.
Two factors are used to determine a risk categorization, similar to traditional risk-based approaches used by the FDA: 1) Significance of Information Provided by the SaMD to the Healthcare Decision. 2) State of the Healthcare Situation of Condition. These factors are used to categorize the SaMD into one of four categories.
State of Health Care Situation or Condition | Significance of Information Provided by SaMD to the Healthcare Decision | ||
Treat or Diagnose | Drive Clinical Management | Inform Clinical Management | |
Critical | IV | III | II |
Serious | III | II | I |
Non-Serious | II | I | I |
What about Modifications to current SaMD?
Some modification will not require review as defined in the guidance “Deciding When to Submit a 510K for a Software Change to an Existing Device.” The types of modification will generally fall into three broad categories:
- Performance: Clinical and Analytical Performance
- Inputs: Used by the Algorithm and their Clinical Association to the SaMD Output
- Intended Use: The Intended Use of the SaMD as Describe Above in the IMDRF Risk Categorization Framework
Software Changes may have different impact on users, including patients, healthcare professionals or others:
- Modification Related to Performance with No Change to the Intended Use or New Input Type
- Modifications Related to Inputs, with No Change to the Intended Use
- Modification Related to the SaMD’s Intended Use
In the Pre-Certification approach, the FDA will assess the culture of Quality and Organizational Excellence of a company and have reasonable assurance of the high quality of their software development, testing and performance of their products. The TPLC approach enables the evaluation and monitoring of a software product from its premarket development to post market performance.
The FDA expects every medical device manufacturer to have an established quality system that is geared toward developing, delivering and maintaining high-quality products throughout the lifecycle. Device that rely on AI/ML are expected to demonstrate analytical and clinical validation, the specific types of data necessary to assure safety and effectiveness during the premarket review.
Good article.