Computer Modeling and Simulation: FDA Outlines Credibility Assessment Framework

With the increasing use of computer modeling and simulation (CM&S) in medical device submissions, the U.S. Food and Drug Administration (FDA) released draft guidance last month detailing a framework based on the risks to assess the credibility of these models.

In the 36-page guide, the agency outlines a nine-step generalized framework for assessing the credibility of computer modeling in a regulatory submission and provides additional detail on key concepts involved in the assessment.

Although the guidelines are largely based on the FDA-recognized American Society of Mechanical Engineers (ASME) V&V 40 standard Assessing the credibility of computer modeling through verification and validation: application to medical devicesthe framework it provides is more general.

“ASME V&V 40 assumes the ability to perform traditional validation activities, i.e. comparing model predictions with well-controlled validated experiments. For computer models used in regulatory submissions, there are often many different sources of evidence that are available to support the credibility of the model,” such as clinical studies, model calibration results, and population-level validation results.

The guidelines apply to computer modeling or simulations used to support pre-market approval (PMA) applications, humanitarian device exemptions (HDE), investigational device exemptions (IDE), notifications pre-market (510(k)) and de novo requests. The FDA writes that such models can be used “in a variety of ways,” including silicone to test devices or “to influence algorithms in software embedded in a device”. Other applications include “silicone clinical trials”, in which CM&S methods are used to evaluate device performance using a virtual cohort to complement real clinical trials, and qualified medical device development tools based on CM&S.

The FDA says the guidelines provide a “risk-based framework that can be used in assessing the credibility of computer modeling and simulation” and are necessary because “regulatory submissions often lack a clear rationale for why the models can be considered credible for the context of use (COU).”

“In all cases, it is necessary to demonstrate that the computer model is credible,” writes the FDA. “Demonstrating the credibility of a model involves various activities that include verification, validation, quantification of uncertainty, analysis of applicability, as well as assessment of adequacy.”

The agency notes that the guidance focuses specifically on first principles-based computational models, such as physics-based or mechanistic models, rather than statistical or data-based models that would be used in computational models. machine learning or artificial intelligence.

“These guidelines are intended to help improve the consistency and transparency of CM&S evidence review, increase confidence in the use of CM&S in regulatory submissions, and facilitate better interpretation of CM&S evidence submitted in reviewed regulatory submissions. by FDA staff,” the agency writes.

The FDA notes that the use of CM&S, or silicone methods, is “well established and growing rapidly” and can elucidate “important information not available from traditional methods”. live Where in vitro evaluations” while streamlining development and reducing pre-marketing costs.

Advice