Goodwin Insights October 17, 2022

FDA Issues Final Clinical Decision Support Software Guidance

On September 28, 2022, the U.S. Food and Drug Administration (“FDA” or “the Agency”) issued its long-awaited final guidance, “Clinical Decision Support Software” (the “CDS Guidance”). The CDS Guidance follows the Agency’s September 2019 draft guidance of the same name (the “Draft Guidance”) and seeks to clarify several key concepts for determining whether clinical decision support (“CDS”) software is a medical device.

Specifically, the CDS Guidance provides the Agency’s interpretation of the four criteria established by the 21st Century Cures Act for determining whether a decision support software function is excluded from the definition of a device (i.e., is considered “Non-Device CDS”). A software function must meet all of the following four criteria to be considered Non-Device CDS:

  1. Not intended to acquire, process, or analyze a medical image or a signal from an in vitro diagnostic device (“IVD”) or a pattern or signal from a signal acquisition system
  2. Intended for the purpose of displaying, analyzing, or printing medical information about a patient or other medical information (such as peer-reviewed clinical studies and clinical practice guidelines);
  3. Intended for the purpose of supporting or providing recommendations to a health care professional (“HCP”) about prevention, diagnosis, or treatment of a disease or condition
  4. Intended for the purpose of enabling such HCP to independently review the basis for the recommendations that such software presents so that it is not the intent that the HCP rely primarily on any of such recommendations to make a clinical diagnosis or treatment decision regarding an individual patient

Software functions that do not meet all four criteria are considered device functions subject to FDA oversight. Notable updates to FDA’s interpretation of the four criteria include the following. 

Criterion 1: Do Not Acquire, Process, or Analyze Medical Images, Signals, and Patterns

Under the first statutory criterion, Non-Device CDS do not “acquire, process, or analyze a medical image, signal from an in vitro diagnostic device or pattern or signal from a signal acquisition system.” In the CDS Guidance, FDA elaborates on its interpretation of what it considers to be medical images, signals, and patterns for purposes of Criterion 1.

  • Medical Images: The CDS Guidance clarifies that FDA considers not only images generated by use of “medical imaging systems (e.g., computed tomography (CT), x-ray, ultrasound, and magnetic resonance imaging (MRI)),” but also images that were not “originally acquired for a medical purpose but are being processed or analyzed for a medical purpose,” to be medical images.
  • Signals: The CDS Guidance’s interpretation of signals is consistent with that of the Draft Guidance, although the CDS Guidance no longer uses the Draft Guidance’s term “physiological signals.” Specifically, both the Draft Guidance and CDS Guidance explain that software functions that use as an input signals from an IVD or “signal acquisition system that measures a parameter from within, attached to, or external to the body for a medical purpose” are devices
  • Patterns: The CDS Guidance expands on the Agency’s interpretation of the term “pattern” and explains that the term as used in Criterion 1 refers to “multiple, sequential, or repeated measurements of a signal or from a signal acquisition system” 

FDA also clarifies that software functions that “assess or interpret the clinical implications or clinical relevance” of medical images, signals, or patterns are software functions that do not meet Criterion 1 because they “acquire, process, or analyze.”

Criterion 2: Display, Analyze, or Print Medical Information

Under the second statutory criterion, Non-Device CDS are intended for the purpose of “displaying, analyzing, or printing medical information about a patient or other medical information (such as peer-reviewed clinical studies and clinical practice guidelines).” Thus, Non-Device CDS utilize as an input the type of data described in Criterion 2 (i.e., medical information) as opposed to the types of medical images, signals, or patterns as described in Criterion 1. 

Both the Draft Guidance and CDS Guidance set forth the Agency’s interpretation that Non-Device CDS “display, analyze, or print patient-specific information, such as demographic information, symptoms, certain test results, patient discharge summaries, and/or other medical information (such as clinical practice guidelines, peer-reviewed clinical studies, textbooks, approved drug or medical device labeling, and government agency recommendations).”  

However, the CDS Guidance provides additional considerations for whether information qualifies as eligible “medical information” for purposes of meeting Criterion 2.

  • The CDS Guidance clarifies that “medical information about a patient to be the type of information that normally is, and generally can be, communicated between HCPs in a clinical conversation or between HCPs and patients in the context of a clinical decision, meaning that the relevance of the information to the clinical decision being made is well understood and accepted.” Accordingly, FDA explains that “medical information about a patient” includes “data/results from devices (including IVDs), when used in a manner consistent with the FDA-required labeling” and when Criterion 1 is met
  • The CDS Guidance explains that “other medical information” includes “information such as peer-reviewed clinical studies, clinical practice guidelines, and information that is similarly independently verified and validated as accurate, reliable, not omitting material information, and supported by evidence”
  • The CDS Guidance introduces the concept of “sampling frequency” as a consideration when determining whether information is considered “medical information” under Criterion 2 or a signal/pattern under Criterion 1. FDA explains that a “single, discrete test or measurement result that is clinically meaningful” is medical information under Criterion 2, while “a more continuous sampling of the same information” is a pattern/signal under Criterion 1. To illustrate the point, FDA highlights that a blood glucose lab test result is medical information while continuous glucose monitor readings are patterns/signals. FDA acknowledges that there is a “continuum between a single sample and a continuous sample”

Criterion 3: Supporting or Providing Recommendations to HCPs

Under the third statutory criterion, Non-Device CDS are those software functions that are intended to support or provide “recommendations to an HCP about prevention, diagnosis, or treatment of a disease or condition.”

FDA significantly expands its interpretation of Criterion 3 in the CDS Guidance by introducing the concepts of software automation bias and time-critical decision-making in determining whether a software function is intended for the purpose of supporting or providing recommendations to an HCP. The CDS Guidance explains that automation bias (“the propensity of humans to over-rely on a suggestion from an automated system”) can ultimately lead to “errors of commission (following incorrect advice) or omission (failing to act because of not being prompted to do so),” particularly if a software provides a single output. Time-critical decision-making further exacerbates the risk of automation bias because an HCP may not have enough time to consider information outside of the software’s output. 

To that end, the CDS Guidance explains that FDA interprets Criterion 3 to describe software that:

  • Provides condition-, disease-, and/or patient-specific information and options to an HCP to enhance, inform and/or influence a health care decision
  • Does not provide a specific preventive, diagnostic, or treatment output or directive;
  • Is not intended to support time-critical decision-making
  • Is not intended to replace or direct the HCP’s judgment

The CDS Guidance also identifies three outputs that could be considered “supporting or providing recommendations to an HCP” as long as they are not intended to support time-critical decision-making and/or to “replace or direct the HCP’s judgment”:

  • List of preventive, diagnostic or treatment options
  • Prioritized list of preventive, diagnostic or treatment options
  • List of follow-up or next-step options for consideration (e.g., after a physician office visit, hospitalization, procedure)

By contrast, the CDS Guidance identifies software intended to provide “a specific preventive, diagnostic, or treatment output or directive that addresses a time-critical decision” to not meet Criterion 3’s requirements. 

Notably, the CDS Guidance also explains that software that provides an assessment regarding whether a patient “may exhibit signs of a disease or condition” or identifies a patient’s “risk probability or risk score for a specific disease or condition” to be “providing a specific preventive, diagnostic, or treatment output” that would not meet Criterion 3. The Agency did not specifically address risk probabilities or risk scores in the Draft Guidance. 

Criterion 4: Enabling Independent Review

The final statutory criterion provides that Non-Device CDS must be intended to enable an HCP to independently review the basis for the recommendations presented by the software so that the HCP does not rely primarily on such recommendations to make clinical decisions for an individual patient. The CDS Guidance provides an updated and more granular explanation of the Agency’s expectations for certain disclosures to enable HCPs to independently review the basis of a software’s recommendations.  

Specifically, the CDS Guidance identifies specific software and labeling recommendations to meet Criterion 4 based on FDA’s experience evaluating CDS; however, the CDS Guidance provides that “sponsors may use alternative approaches so long as their approach enables an HCP ‘to independently review the basis’” for a software’s recommendations. The specific software and labeling recommendations include identification of the product’s intended use, the intended HCP user, the intended patient population, the required input medical information, and “a plain language description of the underlying algorithm development and validation that forms the basis for the CDS implementation.” The CDS Guidance also recommends that the software output provide “any relevant patient-specific information and other knowns/unknowns for consideration.” 

As was the case with Criterion 3, FDA introduces the concept of time sensitivity, explaining that software functions intended for a “critical, time-sensitive task or decision” will not meet Criterion 4 because “an HCP is unlikely to have enough time to independently review the basis” of the software function’s recommendations. The Agency further identifies that, in some cases, developers may need to perform usability testing to determine whether their software functionality satisfies Criterion 4.

Additional Notable Changes

In a departure from the Draft Guidance, the CDS Guidance eliminates discussion of two enforcement discretion policies that were proposed to apply to certain types of decision support software that did not meet all four of the statutory criteria for Non-Device CDS.  

Specifically, in the Draft Guidance, FDA proposed adopting a risk-based approach to regulating decision support software that did not meet all four of the statutory criteria for Non-Device CDS based on two factors from the International Medical Device Regulators Forum framework for risk categorization and corresponding considerations relating to software as a medical device (“SaMD”): the significance of information provided by a SaMD to the health care decision and the state of the health care situation or condition. Under the Draft Guidance, FDA planned to exercise enforcement discretion for the following:

  • Decision support devices that meet only Criteria 1-3 (i.e., CDS devices for HCPs where the HCP is not able to independently review the basis for the software’s recommendations) but that “inform clinical management” for “non-serious situations or conditions”
  • Decision support devices that meet only Criteria 1, 2, and 4 (so-called “patient decision support software”) but that “inform clinical management” for “non-serious situations or conditions”

With respect to patient decision support software, FDA now states that the Agency “intends to be consistent with existing policies,” including those described in the FDA’s other software-related guidance documents.

The Agency concludes the CDS Guidance by providing numerous, specific examples of Non-Device CDS and software functions that are a device, including some examples that have potentially far-reaching implications. Software developers should carefully consider the Agency’s CDS Guidance, including the cited examples, to assess whether any contemplated functionalities could subject their software to FDA device regulation. 

Contact Goodwin FDA team members for any questions related to the impact of the CDS Guidance on your software product(s).