FDA Releases Draft Guidance on Evaluation and Reporting of Age, Race, and Ethnicity Data in Medical Device Clinical StudiesJuly 4, 2016
McKenzie E. Cato* and Allyson B. Mullen
Many device companies struggle to ensure that the patients enrolled in their clinical studies are representative of the sex, age, race, and ethnic make-up of the U.S. population. In part, this task is difficult because device clinical studies are often small and occur at only a few study sites. FDA has previously attempted to provide clarification regarding sex, age, race, and ethnicity in earlier guidances, Evaluation of Sex-Specific Data in Medical Device Clinical Studies (Dec. 2011) and Collection of Race and Ethnicity Data in Clinical Trials (Sept. 2005).
In 2012, Congress directed FDA publish a report “addressing the extent to which clinical trial participation and the inclusion of safety and effectiveness data by demographic subgroups including sex, age, race, and ethnicity, is included in applications submitted to the Food and Drug Administration.” The report was required to include information about clinical trials for all products, not only medical devices. See Food and Drug Administration Safety and Innovation Act, Pub. L. No. 112-144, § 907, 126 Stat. 993, 1092-93 (2012).
FDA submitted the required report to Congress in August 2013. The report reviewed original premarket approval (PMA) applications for medical devices approved in 2011. Of the approved PMAs evaluated for the report, only 40% publicly reported an age-based analysis, only 27% contained a race or ethnicity subgroup analysis, and only 16% had public statements regarding race or ethnicity analyses. In August 2014, FDA published an Action Plan on enhancing the collection and availability of demographic subgroup data across all premarket submission types, not just PMAs. On April 9, 2015, the Institute of Medicine held a public workshop to discuss “strategies for ensuring diversity, inclusion, and meaningful participation in clinical trials.” FDA released a draft guidance on June 20, 2016, titled Evaluation and Reporting of Age, Race, and Ethnicity Data in Medical Device Clinical Studies, incorporating the recommendations from the IOM workshop. Once finalized, the draft guidance will be an extension of the earlier guidances.
FDA’s stated goal of the draft guidance is to “improve the quality, consistency and transparency of data regarding the performance of medical devices within specific age, race and ethnic groups.” The three specific objectives of the draft guidance are to:
- Encourage the collection and consideration of relevant age, race, ethnicity, and associated covariates during study design for devices for which safety, effectiveness, or benefit-risk profile is expected to vary across these groups;
- Outline recommended analyses of study subgroup data; and
- Specify FDA’s expectations for reporting age, race, and ethnicity-specific information in summaries and labeling for approved or cleared medical devices.
FDA’s recommendations for each of these three objectives are discussed in turn below. FDA emphasizes throughout the draft guidance that when there are clinically relevant differences in treatment effects across age, race, or ethnic groups, the effects should be considered in the study design and reported in the device labeling.
Achieving Appropriate Enrollment
The draft guidance explains that it is important for medical device clinical trials to include representative proportions of age, race, and ethnicity subgroups to reflect the intended use population, particularly when there are clinically meaningful differences in safety, effectiveness, or benefit-risk profile across demographic subgroups. The draft guidance discusses potential barriers to enrollment and cites a 2009 FDA report to Congress on how to address low enrollment of certain populations of clinical trials. The draft guidance also includes a list of suggestions for enhancing enrollment of relevant age, race, and ethnicity subgroups. Sponsors are advised to justify in the investigational plan how the enrollment criteria will provide reasonable representation of the intended population.
While the enrollment suggestions may be helpful to some sponsors, they may be of limited utility in certain situations. Perhaps most importantly, the draft guidance does not directly address the unique challenges faced by many device companies in planning a clinical trial, such as having only a few study sites and small total numbers of patients enrolled in many studies, particularly for 510(k)s.
If there are age, race, or ethnicity differences in terms of (1) prevalence, (2) diagnosis and treatment patterns, (3) proportions of age, race, and ethnicity subgroups included in past studies for the target indication, or (4) outcomes related to safety or effectiveness, FDA advises sponsors to provide this information in the protocol and submission documents. Specifically, FDA recommends that sponsors describe this information as part of the risk analysis section of the investigational plan, the study protocol, investigator training materials, sections of the marketing application containing results of clinical investigations, and in any interim or final reports for any mandated postmarket studies.
It appears that FDA intends to continually assess whether there are any clinically meaningful age, race, or ethnicity differences in both the premarket and postmarket setting. If the Agency determines that additional data are needed from a particular demographic subgroup before the device is approved or cleared, it may recommend that sponsors include provisions to encourage enrollment of certain subgroups. It is unclear how a sponsor would incorporate a “recommendation” from FDA for “further enrollment” during the premarket review stage, since presumably, by that time, the study will be completed and data collected and analyzed. FDA may also determine that postmarket evaluation of age, race, or ethnicity may be warranted if there are premarket “signals” indicating that there may be clinically meaningful outcome differences in age, race, or ethnic subgroups.
The draft guidance provides recommendations for sponsors and investigators for avoiding loss to follow-up of subjects. FDA explains that disproportionate loss to follow-up among minorities and older patients may be a barrier to diverse study representation.
Study Design, Analysis, and Interpretation of Study Results
FDA explains that biological differences across age, race, and ethnic groups, such as gonad development, skin texture and color, hormone levels, metabolism, and bone density, can influence the safety and effectiveness of a device. Thus, FDA believes it is important for any differences in data across age, race, and ethnic groups to be accounted for in study design and analysis.
At the IDE study design and early enrollment stage, sponsors are advised to discuss in the Statistical Analysis Plan their strategy for achieving diverse enrollment, any key covariates that could explain differences across subgroups, and whether clinical outcome measurements may differ across subgroups.
For both premarket and postmarket studies, the draft guidance recommends that sponsors submit descriptive statistics to the Agency for outcomes of interest by subgroup, and address the issue of confounding variables by using multivariable analyses adjusted for patient characteristics that could confound the relationship between the subgroup and study outcomes. The recommended statistical analyses are discussed in detail in the draft guidance.
In interpreting study results, if sponsors find that there are any clinically meaningful differences, they are advised to discuss with FDA whether any additional data are needed. If the results of study analyses demonstrate that there are insufficient data to assess whether there are clinically meaningful differences, FDA may determine that clinical data from additional subjects in one or more subgroup is necessary. Additionally, FDA may request additional confirmatory studies, though the draft guidance notes that this would be rare. In cases where sample sizes may not be large enough to detect clinically meaningful differences, the draft guidance recommends that sponsors consult with FDA.
Submitting Age, Race, and Ethnicity Data in Submissions to FDA and Reporting in Public Documents
FDA recommends that sponsors submit and publicly report, in 510(k), PMA, and de novo documents, the study demographics (i.e., proportion enrolled and completed by subgroup). If known, sponsors are advised to discuss the generalizability of study findings to the subgroups. FDA recommends that sponsors analyze and submit any co-morbidities and/or other baseline characteristics by subgroup. FDA also recommends that sponsors report whether loss to follow-up disproportionately affected a particular subgroup. Sponsors are advised to submit or publicly report this information as follows:
- IDE Study Design and Early Enrollment Stage: As part of IDE annual progress reports
- Premarket Submission Stage: As part of the marketing application (including in the labeling) in sections discussing results of clinical investigations and in the 510(k) Summary
- Postmarket Submission Stage: In interim and final reports for any required postmarket studies
FDA staff will also include this information in the PMA Summary of Safety and Effectiveness, HDE Summary of Safety and Probable Benefit, de novo decision summaries, and mandated postmarket study summaries, all of which are publicly available on FDA’s website. Interestingly, FDA does not say that it will include this information in its 510(k) decision summaries published by the Office of In Vitro Diagnostics and Radiological Health (OIR).
FDA recommends that outcome analyses, including any covariates that may explain outcome differences, also be included in labeling and review summaries. In particular, FDA recommends that sponsors discuss how differences across subgroups affect the benefit-risk profile of the device.
The draft guidance includes several flowcharts for various study designs: (1) recommendations for demographic subgroup-specific statistical study design; (2) recommendations for demographic subgroup-specific statistical analysis for one-arm studies; (3) recommendations for demographic subgroup-specific statistical analysis for comparative studies; and (4) recommendations for submitting and reporting subgroup-specific participation and outcome information.
The recommendations in this draft guidance have the potential to impose a substantial burden on sponsors not already performing these analyses. Information provided in an FDA guidance document is, of course, not binding on sponsors, but it does represent FDA’s current expectation for manufacturers and applicants. While FDA can require additional premarket or postmarket studies to address questions regarding device safety and effectiveness across demographic subgroups, it remains to be seen the extent to which sponsors will proactively approach FDA with these questions or report demographic subgroup differences in their submissions. We agree that it is important to understand the safety and effectiveness of devices in age, racial, and ethnic subgroups. This draft guidance, however, does not appear to consider the practical limitations to understanding these key points based on the current study design of most device clinical studies. This draft guidance would benefit from taking a more practical approach to performing these analyses.
* Summer Associate