It is common for uncertainty to arise during the premarket review of novel medical devices. How CDRH has handled that uncertainty in the context of its benefit-risk assessment has generally been somewhat of a black box. While FDA has issued numerous benefit-risk guidances, the Agency has not addressed uncertainty head on. CDRH is now providing industry with insight into its handling of uncertainty in this context through its recently issued draft guidance, “Consideration of Uncertainty in Making Benefit-Risk Determinations in Medical Device Premarket Approvals, De Novo Classifications, and Humanitarian Device Exemptions” (draft available here).
The draft guidance is intended to describe factors that FDA will consider when assessing uncertainty as part of a benefit-risk assessment in PMA, De Novo, and HDE submissions. CDRH explains that whether an “acceptable” level of uncertainty exists will depend on the totality of information in a premarket submission, and FDA will consider the following factors when determining acceptability of uncertainty:
- Probable benefits;
- Probable risks;
- Uncertainty regarding the benefit-risk profile of alternative treatments or diagnostics;
- Patient perspective regarding uncertainty of the device’s benefits and risks;
- Public health need;
- Feasibility of generating extensive premarket data;
- Ability to reduce or resolve uncertainty;
- Likely effectiveness of postmarket mitigations (e.g., labeling);
- Type of decision being made (e.g., more uncertainty is acceptable for HDEs than PMAs); and
- Probable benefits of earlier patient access.
With regard to De Novos, specifically, the draft guidance notes that the probable risks will play a large role in analyzing uncertainty because the uncertainty of a De Novo device’s probable benefit can be mitigated if the risks are minimal or through imposition of special controls.
While uncertainty can arise in a number of different contexts, the draft guidance gives two general types of examples – breakthrough devices subject to PMA and devices with small patient populations. The draft guidance provides three examples, in these two general categories, and stratifies the amount of uncertainty presented in each case: conventional; modest; and high. The draft guidance correlates the level of uncertainty with the statistical confidence from a submission’s clinical study, with clinical studies of: conventional uncertainty having a statistical confidence of 97.5% – 95%; modest uncertainty having a statistical confidence of 95% – 90%; and high uncertainty having a statistical confidence of 80%.
In all of these cases, the draft guidances only answer to uncertainty is whether and how much post-market data will be required. Interestingly, the guidance does not address alternative data sources or information that could be provided to supplement a prospective clinical study with the confidence level described above. Instead, the draft guidance states that the statistical confidence of a submission’s clinical study will drive the amount of post-market data required. Conventional uncertainty will require no post-market data, whereas modest and high uncertainty will require modest and substantial post-market studies, respectively. The draft guidance does not provide additional information regarding what would constitute a modest or substantial post-market study. In our view, it would be helpful for FDA to provide additional clarity as to the post-market considerations, as well as additional data sources that can contribute to or mitigate uncertainty, other than post-market data.