The U.S. Food and Drug Administration (FDA) has issued an announcement that they will be extending the period for public comment on a proposed guidance draft that would reevaluate the way that computer models can be used to gain FDA approval for medical devices.
The guidance in question is officially called 86 FR 72969 or “Assessing the Credibility of Computational Modeling and Simulation in Medical Device Submissions.” The guidance was originally published on Dec. 23, 2021, with an intended 60-day comment period. This period was extended when, according to the FDA’s submission to the Federal Register, they received “a request for an extension to allow interested persons additional time to submit comments.”
This guidance was created because, according to the FDA, “Regulatory submissions often lack a clear rationale for why models can be considered credible for the context of use.”
This guidance was proposed to replace that unclear rationale with “a risk-based framework” that will provide clear guidance for the use of computational models and physics-based simulations. This guidance is not intended to challenge the credibility of statistical or data-driven models.
These computer models are currently used to support medical devices in the submission processes for a number of FDA approvals and clearances. According to the FDA, these models are accepted in applying for:
- Premarket Approval (PMA)
- Humanitarian Device Exemptions (HDE)
- Investigational Device (IDE) status
- Premarket Notifications (510(k)s)
- De Novo requests
- Qualification of Medical Device Development Tools (MDDTs)
With the improved clarity of the proposed guidance, the FDA hopes to make the determination of reliable data more effective.
The four main uses of these computational models, called CM&S models in the proposed guidance are:
- In Silico Device Testing
- CM&S used in medical device software
- In Silico Clinical Trials
- CM&S-based qualified tools
Each of these applications is being evaluated by the FDA in order to determine what makes credible use of CM&S and what does not.
In Silico testing refers to testing a device in a simulated environment, such as a patient model or a specific set of hypothetical scenarios to demonstrate a device’s durability. The FDA states that In Silico models could be used to demonstrate a device’s effectiveness in “representative in vivo conditions.”
The use of CM&S in medical device software is to test the software’s functionality or put a program through a specific set of functions. This could include using a digital model to simulate how a patient may respond to a surgery during the planning stage before the operation or testing a monitoring software’s capacity to detect subtle changes in a patient’s inputs.
In Silico clinical trials evaluate a medical device’s effectiveness in a virtual patient population to either bolster a smaller live clinical trial, or to selectively include or exclude certain traits. The FDA guidance notes that these models can be helpful in detecting potential issues that may not have occurred in the study group but are more common in the indented population that will use the device. However, the FDA stresses that In Silico clinical trials are a tool to supplement live, In vivo trials and are not meant to replace them.
CM&S-based qualified tools refer to tools created to evaluate or develop medical devices under the MDDT program. If they are accepted by the FDA, they become a non-clinical assessment model (NAM) for predicting device safety, effectiveness, or performance of medical devices.
The FDA stated that in order for a CM&S to be considered credible it must demonstrate five characteristics: “verification, validation, uncertainty quantification, applicability analysis, as well as adequacy assessment.”
These factors are enumerated in the guidance’s generalized framework and key concepts sections which together take up the majority of the document with more granular analysis. In layman’s terms, the FDA is trying to determine the accuracy of models by putting the models through a variety of evidence-based tests in order to determine whether the results they put out are based on sound, real-world principles and calculations.
- Verification: the process of determining whether the model’s results are the results of a sound mathematical model when applied to the medical device being evaluated
- Validation: uses evidence to demonstrate that the models used are applicable to the real world
- Uncertainty quantification: identifies the factors that could influence the accuracy of the analysis and quantifies how much each factor may obscure the accuracy of the model
- Applicability assessment: tests how likely the model is to correctly predict an outcome based on credibility and evidence
- Adequacy assessment: makes a determination whether or not there is sufficient evidence to believe that a particular model is credibly able to perform its intended function
The FDA has extended the period for receiving public comment on the publicly available draft document until Mar. 24, 2022. Concerned patients, advocates, scientists, engineers, or citizens can download the proposed guidance and comment their own thoughts to the FDA about these new efforts.