FDA De Novo Classification Pathway for Novel Devices
The De Novo classification pathway gives novel medical devices a route to legal US market entry when no substantially equivalent predicate exists and the device presents low-to-moderate risk. Established under 21 U.S.C. § 513(f)(2) and codified in 21 CFR Part 860, Subpart D, De Novo fills a structural gap in the device regulatory framework that previously forced manufacturers to seek full Premarket Approval (PMA) for products that did not warrant that level of scrutiny. This page covers the pathway's definition, procedural mechanics, the scenarios in which it applies, and the decision boundaries that distinguish it from adjacent pathways.
Definition and scope
The De Novo process is a risk-based classification mechanism administered by the FDA Center for Devices and Radiological Health (CDRH). It assigns a novel device to Class I or Class II — the two lower-risk classifications — rather than automatically defaulting it to Class III, which would require a PMA.
Before De Novo existed in its modern form, FDA's default rule under 21 CFR 860.3 assigned any device with no predicate to Class III regardless of its actual risk profile. The 1997 FDA Modernization Act (FDAMA) introduced the De Novo mechanism to correct that mismatch, and the 21st Century Cures Act (2016) streamlined it further by allowing manufacturers to submit De Novo requests directly, without first receiving a Not Substantially Equivalent (NSE) determination from a 510(k) submission.
The scope of De Novo includes physical devices, in vitro diagnostics, digital health tools regulated as devices, and combination hardware-software architectures — provided the device type has no legally marketed predicate and presents general controls or general-plus-special-controls–level risk. A successful De Novo decision simultaneously classifies the device type and creates a new regulatory framework that subsequent similar devices can use as a predicate for 510(k) clearance.
How it works
The De Novo process follows a defined procedural sequence. FDA's De Novo guidance document (updated 2021) describes the full submission requirements, but the core pathway proceeds in the following order:
- Pre-submission meeting (Q-Sub): The manufacturer meets with CDRH to discuss the proposed device type, intended use, classification rationale, and the special controls that would mitigate identified risks. This step is optional but strongly recommended by FDA because it materially shortens review time.
- De Novo request submission: The applicant submits Form FDA 3881 along with a device description, proposed classification, risk analysis, performance testing data, and a draft order that specifies proposed special controls. The request must include an explanation of why Class III controls are unnecessary.
- FDA substantive review: CDRH has a 150-calendar-day statutory review period under 21 U.S.C. § 513(f)(2)(B), beginning the day after the request is accepted as filed.
- Grant or decline decision: FDA either grants the request — issuing a De Novo order that establishes the device type's classification and special controls — or declines and provides written rationale.
- Predicate creation: Once granted, the approved device type becomes a legally recognized predicate. Third-party manufacturers can then pursue 510(k) clearance for substantially equivalent devices, a downstream effect that amplifies the pathway's market-shaping impact.
The user fee for a De Novo request under the Medical Device User Fee Amendments (MDUFA) for fiscal year 2024 is $23,445 for small businesses (fewer than 500 employees) and $117,224 for standard applicants (FDA MDUFA V Fee Schedule FY2024).
Common scenarios
De Novo is applied most frequently in three broad categories of novel device development:
First-in-class diagnostics. A point-of-care test for a biomarker with no FDA-cleared predecessor — such as an assay measuring a newly identified protein associated with a specific pathology — has no predicate. If clinical and analytical performance data support general-plus-special-controls risk management, De Novo is the appropriate pathway rather than PMA.
Software as a Medical Device (SaMD). FDA's Digital Health Center of Excellence has processed De Novo requests for AI/ML-based clinical decision support tools that cross the regulated device threshold. Because SaMD product types are new, predicates are often unavailable even when risk is moderate. Detailed guidance on FDA's digital health and software regulation framework addresses how these submissions are structured.
Novel therapeutic devices. A wearable neurostimulator designed to treat a specific condition for which no marketed device type exists may present Class II–level risk but cannot reference a predicate. De Novo provides a structured path to Class II classification, with special controls that define the parameters for subsequent 510(k)-eligible devices.
Decision boundaries
The choice between De Novo, 510(k), and PMA depends on two intersecting axes: predicate availability and risk level.
| Factor | 510(k) | De Novo | PMA |
|---|---|---|---|
| Predicate required? | Yes — substantially equivalent | No | No |
| Applicable risk tier | Class I or II | Class I or II (novel type) | Class III |
| Primary evidentiary standard | Substantial equivalence | Reasonable assurance via risk-based controls | Valid scientific evidence (clinical) |
| Outcome | Clearance letter | De Novo order + new predicate | Approval order |
| Review clock | 90 days (standard) | 150 days (statutory) | 180 days (standard) |
A device with an existing predicate should proceed through 510(k) clearance; De Novo is inappropriate if substantial equivalence can be demonstrated. Conversely, if a novel device presents high risk — meaning general or special controls alone cannot provide reasonable assurance of safety and effectiveness — it must pursue Premarket Approval (PMA), which requires valid scientific evidence including, in most cases, clinical trial data overseen through FDA's clinical trials oversight framework.
The boundary between Class II (De Novo-eligible) and Class III (PMA-required) is determined by whether FDA can identify special controls — such as performance testing standards, post-market surveillance requirements, or device-specific labeling specifications — that are sufficient to provide that reasonable assurance. If no feasible set of special controls exists, Class III assignment and PMA follow automatically.
Manufacturers seeking broader orientation to the device classification system, including how Class I, II, and III categories map to regulatory burden, can consult the FDA medical device classification framework and the main FDA authority reference index.