On March 27, 2012, the U.S. Food and Drug Administration (FDA) released a final guidance document describing the principal factors the agency will consider when making benefit-risk determinations for medical devices subject to the premarket review and de novo classification processes, respectively. The final guidance, entitled “Factors to Consider When Making Benefit-Risk Determinations in Medical Device Premarket Approval and De Novo Classifications,” describes the types of medical devices to which the guidance applies, identifies the types of scientific evidence FDA will consider when making benefit-risk determinations, enumerates a series of factors FDA will consider in making such determinations and provides a series of examples illustrating the manner in which FDA will use such factors to make benefit-risk determinations.
Scope of the Final Guidance
The final guidance applies to both diagnostic and therapeutic devices that are subject to premarket approval (PMA) or de novo classification petitions. This represents a shift from the scope of the draft guidance, which addressed the factors for PMA devices and, “in limited cases, [for] devices subject to premarket notification (510(k)) requirements,” but did not address the factors for devices that are the subject of a de novo classification petition.
FDA states the contents of the final guidance should be considered during the design, non-clinical testing, pre-Investigational Device Exemption (IDE) and IDE phases, as well as in assembling and assessing PMAs or de novo petitions. Under the Federal Food, Drug, and Cosmetic Act, sponsors of low- to moderate-risk devices that have been determined by FDA to be not substantially equivalent to a predicate device may submit a petition to the FDA requesting de novo classification and marketing authorization for the device.
Types of Scientific Evidence that Will Be Considered During Risk-Benefit Analysis
Unsurprisingly, FDA states it puts a “great deal of emphasis” on results obtained using clinical testing methods (e.g., randomized controlled trials, well-controlled investigations, testing on clinically-derived human specimens). However, FDA acknowledges that, in certain situations, non-clinical data (e.g., data generated by performance testing for product safety, reliability and/or characterization, computer simulations) may play an important role in demonstrating the safety and effectiveness of a medical device, as medical devices often have attributes that cannot be tested using clinical methods alone. As such, both clinical and non-clinical data may be useful in demonstrating the safety and/or effectiveness of a device.
Factors FDA Considers in Making Benefit-Risk Determinations
In the final guidance, FDA groups into three categories the factors the agency will use when making benefit-risk determinations: the factors to be used in assessing the benefits of devices, the factors to be used in assessing the risks of devices, and additional factors to be used in the assessment of the probable benefits and risks of devices.
In assessing the benefits associated with the use of a device, FDA will consider:
- Types of benefits—What primary endpoints, secondary endpoints or surrogate endpoints were considered? What value do patients place on the benefit?
- Magnitude of benefits—What was the magnitude of treatment effect for each endpoint? What scale is used to measure the benefit? How did the benefit rank on that scale?
- Probability of a patient experiencing one or more benefits—Was the study able to predict which patients will experience a benefit? What is the probability that a patient for whom the device is intended will experience a benefit? How did benefits evaluated vary across subpopulations?
- Duration of effects—Could the duration of each treatment effect (including primary and secondary endpoints) be determined (if relevant)? Is the duration of benefit achieved of value to patients?
With respect to the assessment of risks, FDA will take the following factors into account:
- Severity, type(s), number and rate(s) of harmful events associated with the use of the device—What are the device-related serious adverse events, device-related non-serious adverse events and other procedure-related complications that a patient may be subject to?
- Probability of a harmful event—What percentage of the intended patient population would expect to experience a harmful event? What is the incidence of each harmful event? Is there uncertainty in that estimate? How does the incidence of harmful events vary by subpopulation? Are patients willing to accept the probable risk of the harmful event given the probable benefits of the device?
- Duration of harmful events—How long does the harmful event last? Is it reversible? What type of intervention is required to address the harmful event?
- Risk of false-positive or false-negative results from diagnostics—What are the consequences of a false positive? What are the consequences of a false negative? Does the device provide the only means of diagnosing the problem, or is it part of an overall diagnostic plan?
- Aggregate effect of harmful events—Do multiple harmful events occur at once?
- Uncertainty—How robust were the data? How was the trial designed, conducted and analyzed? Are the results repeatable? Is the study the first of a kind, or have other studies achieved similar results? Can the results of the study be applied to the population generally, or are they intended for discrete, specific groups?
- Characterization of the disease—How does the disease affect the patients that have it? Is the condition treatable? How does the condition progress?
- Patient tolerance for risk and perspective on benefit—Did the sponsor present data regarding how patients tolerate the risks posed by the device? Are the risks identifiable and definable? Is the disease so severe that patients will tolerate a higher amount of risk for a smaller benefit? Is the disease chronic? How long do patients with the disease live? If the disease is chronic, is illness easily managed with less invasive or difficult therapies? Are patients willing to take the risk of this treatment to achieve the benefit? How well are patients able to understand the risks and benefits?
- Availability of alternative treatments or diagnostics—What other therapies are available for this condition? How effective are the alternative treatments? How well-tolerated are the alternative treatments? What risks are presented by the alternative treatments?
- Risk mitigation—Are there ways to mitigate risks (e.g., product labeling, establishing educational programs, providing add-on therapy, etc.)? What type of mitigation strategy is proposed?
- Postmarket data—Are the probabilities of effectiveness and the rates of harmful events for similar devices similar to what is expected for the device under review? Is there reason to believe that long-term device performance, effectiveness of provider training programs, safety and/or efficacy of the device in subgroups or rare adverse events should be evaluated in the postmarket setting? Is there a reason to expect a significant difference between “real world” performance of the device and the performance found in premarket trials? Are there any data that would be provided to support approval that could be deferred to the postmarket setting?
- Novel technology addressing unmet medical need—How well is the medical need this device addresses being met by currently available therapies? How desirable is this device to patients?
Appendix B of the final guidance includes a worksheet that clearly summarizes the above-listed factors that reviewers will use in making benefit-risk determinations as part of the PMA or de novo classification process. Appendix C provides useful examples of how reviewers might utilize the worksheet in the review of products. As such, these appendices provide a clear roadmap for device manufacturers to follow and may facilitate smoother approval of innovative devices.
In issuing this helpful final guidance, FDA indicates a commendable willingness to undertake a more flexible and transparent approach to benefit-risk determinations, at least as to devices subject to PMA and de novo review. The multi-factor approach to benefit-risk assessment described in the final guidance, if appropriately implemented, will enable FDA to conduct an individualized assessment of each medical device subject to PMA or de novo classification. Moreover, by explicitly stating the criteria FDA will use to assess the above-referenced submissions, device manufacturers should be better able to anticipate the types of data they will need to submit to support a PMA or de novo petition.
In particular, the final guidance is notable in that it indicates FDA is willing to consider a patient’s informed request to utilize a device as evidence that such device should be made publicly available—even in situations where the agency might otherwise consider such device to be too risky. In recent years, FDA has been the subject of some criticism for declining to approve products intended to treat serious or chronic conditions when it deems them to raise issues relative to their risk. Although it remains to be seen whether FDA will consider patient risk tolerance data to be persuasive, the final guidance represents, at minimum, a nominal step toward embracing broader benefit-risk factors and potentially increasing patient access to innovative medical devices.