- The American Medical Informatics Association (AMIA) recently requested the Food & Drug Administration (FDA) improve transparency and performance standards for clinical decision support systems in submitted comments.
AMIA’s comments come in response to FDA guidance released for public comment in December 2017. The guidance specified which kinds of clinical decision support system functionality do not meet the definition of a medical device as defined by the 21st Century Cures Act. Additionally, the guidance indicated which kinds of clinical decision support functionality do qualify as medical devices, but will not be regulated by FDA.
In the letter, AMIA President & CEO, Douglas B. Fridsma, MD, PhD, commended the federal agency’s decision to draft new guidance about clinical decision support and patient decision support systems. The association stated the guidance could help to initiate necessary discussions about the future of clinical decision support systems as the healthcare industry moves toward cloud-based computing, machine learning, and improved patient EHR data access.
“This area of innovation is developing at a tremendous rate, and we believe several trends are converging to transform the delivery of computable biomedical knowledge at-the-point of care,” wrote Fridsma.
However, AMIA also advised FDA that the guidance lacked discussion about how the administration plans to regulate clinical decision support and patient decision support that are considered medical devices.
“As we understand it, this draft guidance seeks to distinguish between decision support software that is, and is not, a device, consistent with the amendments to the Food Drug & Cosmetic Act made through the Cures Act,” Fridsma explained. “In this regard, we anticipate lingering confusion among developers and clinicians trying to determine whether specific decision support software is, or is not, considered a device.”
The association’s letter emphasized that the criteria FDA is using to determine if a functionality falls under the device designation is ambiguous.
In order to avoid causing confusion among health IT developers and clinicians, AMIA recommended FDA make the following improvements to draft guidance:
Implementing these recommendations may give health IT developers and clinicians more clarity about FDA regulations surrounding clinical decision support systems.
AMIA also requested FDA host a public meeting to discuss standards for transparency and performance of clinical decision support systems in machine learning-based environments. Advancements in cloud computing and machine learning in particular have opened the door to new patient safety risks and may require new kinds of regulation.
“While we do not contend that such functionality should be free from FDA oversight, there is a need to explore the range and types of potential applications related to decision support software, and a need to discuss their varying risks of harm to patient safety and public health,” Fridsma added.
In December 2017, AMIA also advocated for the development of the EHR reporting program in accordance with provisions of the Cures Act.
AMIA teamed up with Pew Charitable Trusts and requested congressional appropriators ensure ONC has adequate funding to implement the EHR reporting program, which mandates that EHR developers submit data about the functions of their products.
Implementing the EHR reporting program could contribute to the establishment of new requirements related to EHR certification, support care coordination, reduce clinician burden, and improve patient safety.