REGULATORY SPONSORED ARTICLE
ment (e.g., an Apgar score calculator that calculates a score at the birth of a baby), FDA could determine to apply a lower level of regulatory oversight than for high-risk CDS, where the information is relied upon for final clinical decisions (e.g., determining radiation therapy, identifying anomalies in medical images). Other factors that might affect the risk calculation include the extent of reliance on CDS and whether there is time and opportunity for “competent human intervention,” the degree of acceptance of the methodology in clinical practice, the complexity of the clinical condition, and the ability to easily identify erroneous output (i.e., the transparency of the calculation or algorithm). Applying
this approach, FDA could decide, for example, that compliance with device registration/listing and labeling requirements (or possibly enforcement discretion) is appropriate for CDS that is lower on the risk spectrum; compliance with device quality systems and adverse event reporting requirements is appropriate for moderate risk CDS; and premarket review and clearance/approval are necessary for high risk CDS. *** Although FDA has been slow to develop a regulatory framework for health IT, there are indications that the Agency is now at-
tempting to adopt a flexible approach for this product category that will both foster innovation and protect the public health. By taking an iterative approach to regulation of health IT, selectively using the regulatory tools available to it (e.g., guidance documents, existing regulations, and rulemaking), and exercising its enforcement discretion, the FDA has the ability to address any risks that may arise, while allowing innovation to continue. The forthcoming final guidance on mobile medical apps and draft guidance on CDS will reveal further the extent to which the FDA will adopt a smart regulation approach.
14
LMG LIFE SCIENCES 2013