What evidence FDA requires for medical devices

What has to be proven depends on the marketing route. A premarket notification must show that the device is substantially equivalent to a legally marketed predicate device, while a premarket approval application must show reasonable assurance that the device is safe and effective for its conditions of use.¹² A De Novo request must show that general controls, or general controls together with special controls, provide that reasonable assurance for a device type that has no predicate.³ A humanitarian device exemption application rests on safety and on a showing that probable benefit outweighs risk of injury or illness.² Section 860.7 sets the rules on what counts as evidence for classification, for establishment of special controls for class II devices and for premarket approval of class III devices.³ FDA also applies those criteria when deciding a De Novo request under § 860.260(e) and a PMA under § 814.45(c).²³

How the class of the device sets the standard of proof

Part 860 defines three classes by the level of control needed to give reasonable assurance of safety and effectiveness. A device is in class I where general controls are sufficient, or where there is insufficient information to determine sufficiency but the device is not life-supporting or life-sustaining and presents no potential unreasonable risk.³ A device is in class II where general controls alone are insufficient and there is sufficient information to establish special controls, which may include performance standards, postmarket surveillance, patient registries and guidelines for the submission of clinical data in 510(k) submissions.³ A device is in class III where insufficient information exists to determine that general or special controls would give that assurance and any of the following applies.³

• the device is life-supporting
• the device is life-sustaining
• the device is of substantial importance in preventing impairment of human health
• the device presents a potential unreasonable risk

Section 860.10 sets a default for implants and life-supporting or life-sustaining devices. A classification panel recommends class III for such a device unless it determines that class III is not necessary for reasonable assurance of safety and effectiveness.³ Where the panel recommends a lower class, it must state its reasons and identify the risks to health.³ The Commissioner classifies such a device into class III on the same terms, and a decision to place it elsewhere must be accompanied by a full statement of reasons with supporting documentation and data satisfying § 860.7.³

Valid scientific evidence under § 860.7

Section 860.7 governs every safety and effectiveness determination made for classification, for special controls and for premarket approval.³ A manufacturer may submit any form of evidence, and the agency relies only on valid scientific evidence in deciding whether there is reasonable assurance that the device is safe and effective.³

Valid scientific evidence means evidence from well-controlled investigations, partially controlled studies, studies and objective trials without matched controls, well-documented case histories conducted by qualified experts, and reports of significant human experience with a marketed device.³ The test is whether qualified experts can fairly and responsibly conclude from it that there is reasonable assurance of safety and effectiveness under the conditions of use.³ The amount of evidence required varies with the characteristics of the device, its conditions of use, the adequacy of warnings and restrictions, and the extent of experience with its use.³ Isolated case reports, random experience, reports lacking sufficient detail for scientific evaluation and unsubstantiated opinions are not regarded as valid scientific evidence of safety or effectiveness. The agency may still consider them in identifying a device of questionable safety or effectiveness.³

Safety and effectiveness carry separate definitions. There is reasonable assurance of safety where valid scientific evidence shows that probable benefits outweigh probable risks when the device is accompanied by adequate directions and warnings.³ The types of evidence that may be required for safety include investigations using laboratory animals, investigations involving human subjects, nonclinical investigations and analytical studies for in vitro diagnostic devices.³ There is reasonable assurance of effectiveness where valid scientific evidence shows that use of the device will give clinically significant results in a significant portion of the target population.³ Effectiveness evidence must consist principally of well-controlled investigations, unless the Commissioner authorizes reliance on other valid scientific evidence determined to be sufficient.³

Evidence in a premarket notification

A 510(k) submission is comparative. Section 807.87(f) requires a statement indicating how the device is similar to or different from other products of comparable type in commercial distribution, accompanied by data to support the statement.¹ That information may identify similar products, materials, design considerations, energy expected to be used or delivered, and the operational principles of the device.¹ Where the submission covers a significant change that could significantly affect safety or effectiveness, or a new or different indication for use, § 807.87(g) requires appropriate supporting data showing that the manufacturer has considered the consequences of the change for safety and effectiveness.¹

Section 807.100(b) sets the criteria FDA applies. The device must have the same intended use as the predicate.¹ Where technological characteristics are the same, that comparison carries the finding.¹ Where they differ, the submitted data must establish substantial equivalence. That data must include information demonstrating that the device is as safe and as effective as a legally marketed device, including clinical data if the Commissioner deems it necessary. The differences must also raise no different questions of safety and effectiveness.¹ The predicate must not have been removed from the market at the Commissioner’s initiative or found misbranded or adulterated by judicial order.¹

The 510(k) summary reflects whatever evidence was relied on. Where the determination rests in part on performance data, § 807.92(b) requires a brief discussion of the nonclinical tests and of the clinical tests relied on, a description of the subjects tested, the safety or effectiveness data obtained with specific reference to adverse effects and complications, and the conclusions drawn.¹ Section 807.97 provides that a representation creating an impression of official approval because of compliance with the premarket notification regulations is misleading and constitutes misbranding.¹

Additional evidence applies to comparisons against preamendments class III devices. Where the submission claims equivalence to a device classified into class III under section 513(b) that was in commercial distribution before 1 December 1990 and for which no final regulation requiring premarket approval has issued, § 807.87(k) requires a class III summary of the types of safety and effectiveness problems associated with that device type with citations, plus a signed class III certification in the form at § 807.94.¹ FDA may require submission of the adverse safety and effectiveness data described in that summary.¹

Evidence in a De Novo request

A De Novo request applies where no legally marketed device exists on which to base a substantial equivalence review, either after a not substantially equivalent order or where the requester determines no predicate exists.³ Section 860.220(a)(8) requires a classification summary containing the following.³

• a description of the searches used to establish that no legally marketed device of the same type exists
• a list of classification regulations, PMAs or HDEs for potentially similar devices
• a list of 510(k)s, EUAs or product codes for potentially similar devices
• a rationale for how the subject device differs

Where the device was previously found not substantially equivalent, the requester gives the 510(k) number and a summary of the search confirming no intervening classification.³

The substantive showing is built on risks and mitigations. Section 860.220(a)(9) requires a summary of probable risks to health known or reasonably knowable to the requester. The proposed mitigations for each risk must include general controls and, where class II is sought, special controls.³ Each mitigation involving performance testing or labeling must be cross-referenced to the supporting information in the request.³ FDA applies § 860.7 in deciding whether to grant or decline, and may use information other than that submitted.³

Section 860.260(c) lists grounds for declining that bear directly on evidence. These include a nonclinical study subject to part 58 that is essential to show reasonable assurance of safety and was not conducted in compliance with part 58, a clinical investigation not conducted in compliance with parts 50 and 56 or with good clinical practice such that subject rights or safety were not protected or the data are unreliable, and a study necessary to the demonstration that has not been completed per the study protocol.³ FDA may inspect relevant facilities before deciding, to confirm that clinical or nonclinical data accurately represent the benefits and risks of the device.³

Evidence in a premarket approval application

A PMA must contain separate sections on nonclinical laboratory studies and on clinical investigations involving human subjects.² The summary required by § 814.20(b)(3) covers indications for use with the patient population, a device description explaining how the device functions and the scientific concepts behind it, existing alternative practices and procedures, and the foreign and United States marketing history including any withdrawal for reasons of safety or effectiveness.² An applicant that omits a required item must justify the omission.²

FDA has 45 days after receipt to notify the applicant whether the application has been filed, and the 180-day review period runs from the date of filing.² Approval follows where none of the grounds in § 814.45 applies.² Among those grounds are a nonclinical laboratory study essential to show safety that was not conducted under the good laboratory practice regulations in part 58, and a clinical investigation not conducted in compliance with parts 50 and 56 or good clinical practice such that subject rights or safety were not protected or the data are unreliable.² Section 814.45(c) directs FDA to use the criteria in § 860.7 in deciding approval and permits reliance on information beyond the application.²

A PMA may rest solely on foreign clinical data where three conditions are met. The data must be applicable to the United States population and United States medical practice, the studies must have been performed by clinical investigators of recognized competence, and the data must be valid without an on-site inspection or capable of validation by FDA through inspection or other appropriate means.² Applicants are encouraged to meet FDA in a presubmission meeting where approval on foreign data alone will be sought, which is a recommendation rather than a requirement.² Where data from investigations conducted outside the United States support a PMA, the applicant complies with § 812.28.² A class III device with a product development protocol declared completed by FDA is considered to have an approved PMA.²

Changes after approval carry their own evidentiary burden. A PMA supplement is required before a change affecting safety or effectiveness, including the following.²

• new indications for use
• labeling changes
• a different manufacturing facility
• changes in sterilization procedures or packaging
• changes in performance or design specifications, components or principle of operation

The information in a supplement is limited to that needed to support the change.²

Evidence in a humanitarian device exemption application

An HDE seeks exemption from the effectiveness requirements of sections 514 and 515 of the act.² The applicant first obtains humanitarian use device designation from FDA’s Office of Orphan Products Development, supported by documentation with authoritative references showing that the device is designed to treat or diagnose a disease or condition affecting or manifested in not more than 8,000 people in the United States per year.² Authoritative references include literature citations in specialized medical journals, textbooks, specialized medical society proceedings and governmental statistics publications.² FDA acts on the designation request within 45 days.²

The application itself carries all the information required under § 814.20(b), with two substitutions. In place of the summaries, conclusions and results from clinical investigations required for a PMA, the applicant supplies the summaries, conclusions and results of all clinical experience or investigations reasonably obtainable that are relevant to an assessment of the risks and probable benefits of the device, whether adverse or supportive.² The labeling must bear the statement that the device is authorized by Federal law for the stated use and that its effectiveness for that use has not been demonstrated.² The application must also explain why probable benefit to health outweighs the risk of injury or illness, taking account of currently available devices and alternative treatments, with a description of the underlying disease process and the known or postulated mechanism of action.² FDA completes its review within 75 days of a filed HDE.²

Questions people also ask

Does a 510(k) always require clinical data?

No. Section 807.100(b) requires clinical data only where the new device has different technological characteristics from the predicate and the Commissioner deems clinical data necessary to demonstrate that the device is as safe and as effective as a legally marketed device.¹ Section 807.87(f) requires data supporting the statement of similarity and difference in every submission, and § 807.87(g) requires supporting data where a significant change or a new indication for use is involved.¹

What is the difference between a De Novo request and a reclassification petition?

A De Novo request under section 513(f)(2) seeks classification into class I or class II for a device type with no legally marketed predicate, and FDA acts within 120 days of acceptance.³ A reclassification petition under § 860.123 seeks a change to the classification status of a device type already classified, and must include a full statement of reasons with supporting data satisfying § 860.7 and representative data unfavourable to the petitioner’s position.³

Can foreign clinical data alone support approval of a class III device?

Yes, where the criteria in § 814.15(b) are met. The data must be applicable to the United States population and medical practice, the investigators must be of recognized competence, and the data must be valid without an on-site inspection or capable of being validated by FDA through inspection or other appropriate means.²

What happens if a nonclinical study was not run under good laboratory practice?

Section 814.45(a)(4) permits FDA to deny a PMA where a nonclinical laboratory study essential to showing safety was not conducted under part 58 and no reason for the noncompliance is provided, or where the stated differences from part 58 do not support the validity of the study.² Section 860.260(c)(8) applies the equivalent ground to a De Novo request.³

Does an approved HDE mean the device has been shown to work?

No. An HDE is an exemption from the effectiveness requirements of sections 514 and 515 of the act, granted on a showing that probable benefit outweighs risk.² The labeling must state that the effectiveness of the device for the authorized use has not been demonstrated.² A humanitarian use device may be administered only in facilities with oversight by an institutional review board acting under part 56, and only where that board has approved the use.²

Footnotes

1. 21 CFR Part 807, Establishment Registration and Device Listing for Manufacturers and Initial Importers of Devices. https://www.ecfr.gov/current/title-21/chapter-I/subchapter-H/part-807 ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹ ↩¹⁰ ↩¹¹ ↩¹² ↩¹³ ↩¹⁴

2. 21 CFR Part 814, Premarket Approval of Medical Devices. https://www.ecfr.gov/current/title-21/chapter-I/subchapter-H/part-814 ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹ ↩¹⁰ ↩¹¹ ↩¹² ↩¹³ ↩¹⁴ ↩¹⁵ ↩¹⁶ ↩¹⁷ ↩¹⁸ ↩¹⁹ ↩²⁰ ↩²¹ ↩²² ↩²³ ↩²⁴ ↩²⁵ ↩²⁶ ↩²⁷ ↩²⁸ ↩²⁹

3. 21 CFR Part 860, Medical Device Classification Procedures. https://www.ecfr.gov/current/title-21/chapter-I/subchapter-H/part-860 ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹ ↩¹⁰ ↩¹¹ ↩¹² ↩¹³ ↩¹⁴ ↩¹⁵ ↩¹⁶ ↩¹⁷ ↩¹⁸ ↩¹⁹ ↩²⁰ ↩²¹ ↩²² ↩²³ ↩²⁴ ↩²⁵ ↩²⁶ ↩²⁷ ↩²⁸ ↩²⁹ ↩³⁰