HITInfrastructure

Virtualization News

AHA: Speed FDA Approval of Clinical Decision Support Algorithms

AHA called for changes to the FDA approval process for clinical decision support algorithms, including exempting software that uses data generated by FDA-regulated devices.

Clinical decision support algorithms

Source: Getty Images

By Samantha McGrail

- The AHA strongly supports the FDA’s goals to implement Section 3060 (a) of the 21st Century Cures Act aimed at enhancing new treatments and driving innovative healthcare. But the association voiced concerns over the language governing FDA approval of clinical decision support algorithms.

Clinical decision support systems are key to delivering healthcare in the age of technology. Providers have brought the systems and their algorithms to patient bedsides in order to develop patient-specific recommendations and support patient care, the AHA explained in a letter to the FDA.

These systems have become so popular that Frost & Sullivan recently predicted clinical decision support systems to replace the EHR as the preferred user interface for health IT infrastructure due to their usability.

The FDA recognized the importance of clinical decision support tools in recent draft guidance to implement Section 3060(a) of the 21st Century Cures Act, which aims to deter over-regulation of clinical decision support software by establishing criteria to exempt low-risk software from FDA regulation, while also ensuring control over software that replaces decision-making of healthcare professionals.

The AHA strongly supported the FDA’s oversight of any software that automatically determines clinical treatment or action. But association expressed concerns that the FDA’s interpretation of various criteria could force existing clinical decision support algorithms to go through the FDA’s approval process, which would slow the development of new software.

READ MORE: FDA Proposes Risk-Based Clinical Decision Support Software Rules

The association recommended that the FDA clarify in official guidance that software that collects and analyzes data downstream from FDA-regulated devices to provide insights to healthcare professionals (HCPs) be exempt from FDA regulation.

Specifically, the association highlighted FDA’s first draft criterion that states FDA regulation would not be needed for software that is not intended to acquire, process, or analyze a medical image or signal from an in vitro (IVD) device or pattern or signal from a signal acquisition system. In regard to the criterion, the AHA said that the FDA should form a clear distinction between an algorithm that analyzes data from an EHR from one that generates the original data within the device.

The association also expressed concerns about the draft criterion that would exempt clinical decision support algorithms from FDA approval if they intend to support or provide recommendations to a healthcare professional about prevention, diagnosis, or treatment of a disease or condition.

While the AHA supported FDA regulation of any clinical decision support algorithm that “takes independent review and action out of the hands” of providers, the association commented that the draft criterion could apply an arbitrary distinction between two types of clinical decision support tools: those that inform clinical management and those that aim to drive it.

Applying this distinction does not reflect how clinical decision support algorithms support patient care, the AHA stated. For example, the International Medical Device Regulators Forum (IMDRF) Framework states that tools that inform clinical management do not trigger an “immediate or near-term action.” However, most input from clinical decision support tools does just that, the association explained. And that output is especially used to aid in prevention, diagnosis, and treatment of diseases and conditions.

READ MORE: Policies Needed for AI-Enabled Clinical Decision Support Systems

nder the proposed construct, CDS software used to ‘identify early signs of a disease or condition’ or ‘aid in diagnosis by analyzing relevant information to help predict risk of a disease or condition’ would fall under ‘driving clinical management’ and therefore not be exempt from FDA regulation under this criterion,” the letter stated. “Yet, these examples highlight some of the most compelling use cases for decision support as they create one of many inputs for the HCP to independently consider when determining diagnosis and treatment.”

AHA encouraged the FDA to reconsider applying the arbitrary distinction and instead propose a policy that is consistent with the statue’s focus on “supporting or providing recommendations about prevention, diagnosis or treatment.”

In the draft guidance, the FDA would also require that exempt software functions be transparent and describe the inputs used to generate the recommendation and the basis for rendering the recommendation to the healthcare professional.

The AHA stood behind this criterion but believes the FDA should clarify that a clinical decision support algorithm meet this requirement if the information is accessible regardless of if the healthcare professional chooses to access the information.

“We look forward to working with the FDA to ensure the agency’s regulatory approach to implementing Section 3060(a) is consistent with the language and intent of the Cures Act and prioritizes patient safety while at the same time allowing hospitals and health systems to continue to implement innovative decision support tools,” the AHA concluded.

X

Sign up for our free newsletter covering the latest IT technology for Hospitals:

  • Virtualization
  • Cloud Computing
  • Storage
  • And More

Our privacy policy


no, thanks