21 CFR Part 812 · FDA EFS Guidance (2013)
Under FDA's 2013 Early Feasibility Study guidance, a novel cardiovascular implant can begin first in human testing on a small cohort before its design is finalized, generating clinical data nonclinical testing alone cannot provide. A non US sponsor must still resolve who holds US sponsor of record under 21 CFR 812.18(a) before the study can proceed.
The Guidance
FDA finalized its Early Feasibility Study guidance in October 2013, creating an IDE pathway built for a genuinely new device mechanism rather than a settled design awaiting confirmatory data. The guidance describes the purpose in plain terms.
"Allow for early clinical evaluation of devices to provide proof of principle and initial clinical safety data."
FDA, Investigational Device Exemptions (IDEs) for Early Feasibility Medical Device Clinical Studies, October 2013. Read the guidance →EFS fits transcatheter structural heart devices, cardiac rhythm management systems, and other implantable mechanisms where the clinical question is exploratory, not confirmatory. Cardiovascular implantables make up the largest share of EFS activity of any device area, since valve repair, valve replacement, and rhythm management programs routinely reach a point where nonclinical bench and animal testing cannot predict how the device performs in a beating heart.
Source: MDIC, A Decade of Innovation in Medical Device Testing →
Before You File
A Pre Submission meeting is where FDA's EFS determination actually gets made. Before drafting the IDE itself, a sponsor uses the Pre Submission to ask FDA directly whether the device mechanism and the available nonclinical package support an EFS entry point, what the agency expects to see in the clinical protocol, and how much design flexibility FDA will allow once the study starts. Coming to that meeting with a clear nonclinical data package and a candidate protocol produces a far more useful answer than asking in the abstract.
Building the Package
An EFS package still needs a full nonclinical evaluation; it just does not need to answer every question nonclinically before the first patient is treated. Bench testing establishes mechanical performance under simulated use. Animal testing, where a relevant model exists, demonstrates in vivo function and identifies failure modes before human exposure. Computational modeling, increasingly accepted alongside physical testing, predicts device behavior across a range of anatomies the bench and animal work cannot fully cover.
"Leveraging nonclinical data" in EFS language means FDA will accept a smaller, more targeted nonclinical package than a traditional IDE would require, on the understanding that the early clinical phase itself is where the remaining open questions get answered, under close monitoring and a protocol built to detect problems early.
The Protocol
Most EFS cohorts enroll 10 to 15 subjects, a number sized to generate an early safety and performance signal, not to prove effectiveness. FDA expects the protocol to carry three features regardless of cohort size.
Where It Runs
EFS site selection favors experience over convenience. MDIC runs the Clinical Site Alliance, a program, not a vendor, that brings together clinical sites with demonstrated EFS experience to share enrollment and operational best practices across the field. The Alliance's own working target, known as its "60/60/60" goal, is to execute the clinical trial agreement, obtain IRB approval, and enroll the first patient within 60 days each. A sponsor working with an experienced EFS site reaches first patient in far less time than one starting from a site with no EFS history.
Source: MDIC, Clinical Site Alliance →
Who Carries It
An EFS is still an IDE study, and every obligation in 21 CFR 812 Subpart C applies to it: investigator selection and agreements, monitoring, adverse event reporting, recordkeeping, and direct correspondence with FDA. Nothing about the EFS designation reduces the sponsor's accountability; it changes how much nonclinical certainty FDA requires before the study starts, not who answers for the study once it does.
What Comes Next
Moving from EFS to a pivotal study is a formal IDE supplement, not a new application. The supplement updates the protocol to reflect what the EFS cohort showed, sets a sample size powered to the pivotal effectiveness endpoint rather than the EFS safety signal, and locks the statistical analysis plan FDA will hold the pivotal study to. Any device design change made during or after the EFS is reviewed against what the original IDE authorized, so design changes made along the way are documented and, where they are significant, cleared with FDA before the pivotal study enrolls.
The Three Structures
A non US device company has three structures available to satisfy 21 CFR 812.18(a) for an EFS, the same three available for any IDE study:
See the full US Sponsor of Record tier structure → · Use the decision tool →
Common Questions
An Early Feasibility Study (EFS) is an FDA IDE pathway, defined in FDA's 2013 guidance, that allows a small first in human study of a novel device while the design may still change based on early clinical experience. It fits a genuinely new device mechanism, not a settled design awaiting confirmatory data.
Most EFS cohorts enroll 10 to 15 subjects, sized to generate an early safety and performance signal rather than statistical proof of effectiveness. The exact number is set with FDA during Pre Submission discussion, based on the specific device and clinical question.
Yes. Under 21 CFR 812.18(a), a company outside the United States cannot itself sponsor an FDA IDE study, including an EFS. A US person or entity must act as sponsor and carry every Part 812 obligation for the study.
Reviewed October 19, 2026 by ADB's regulatory team.
Related Reading
Planning an EFS?
A 30 minute call to assess your device, your nonclinical package, and whether EFS or a traditional IDE is the right entry point.