5

Agency Perspectives: CMS, FDA, and NIH

Speakers from three federal agencies shared their perspectives on the roles of their organizations in improving diversity and inclusion in clinical trials. Shari Ling, deputy chief medical officer at the Centers for Medicare and Medicaid Services (CMS), shared her perspective on opportunities for CMS contributions under its current authority in a conversation moderated by Esther Krofah, executive vice president for health for the Milken Institute and member of the Forum on Drug Discovery, Development, and Translation. Monica Bertagnolli, director of the National Institutes of Health (NIH), Robert Califf, commissioner of food and drugs at the U. S. Food and Drug Administration (FDA), and Namandjé Bumpus, principal deputy commissioner at FDA, discussed the roles of their individual agencies as funder and regulator, respectively, and how they work together, in a conversation moderated by Freda Lewis-Hall, former executive vice president and chief medical officer at Pfizer.

CMS

CMS provides services for almost one in every three individuals in the United States. Ling explained that CMS can make coverage decisions; oversee compliance with health and safety standards by all participating facilities; implement quality measurement and quality improvement; and evaluate new care and payment models. These activities are all evidence based, and Ling noted that lack of data on the full spectrum of populations that CMS serves impacts coverage decisions. For example, data from trials of promising products sometimes do not reflect the full population impacted (and might even exclude those in whom the disease is more prevalent); that gap creates limitations for providing full coverage.

CMS, in its role as the largest U.S. public payer, collects claims data for beneficiaries who received services. However, these data do not always reflect the totality of populations impacted by the disease or conditions because there are still gaps in insurance. This is a limitation for determining full coverage and full delineations of care, health, and safety standards. Ling noted that community care can be a “force multiplier” and an opportunity to overcome difficulties CMS has observed regarding missing data and evidence.

Advancing equity and expanding access are two of the CMS strategic pillars, and the agency is prioritizing initiatives in these areas. Lack of evidence can result in inefficiencies in implementation and inequitable health outcomes.

CMS Opportunities to Stimulate and Transform Evidence Generation

CMS has several tools to stimulate generation of evidence to support coverage determinations that might also be leveraged to increase the diversity of clinical trial data.

Coverage Determinations

Most CMS coverage determinations are made at the local level by Medicare contractors, Ling explained. This allows for decisions that better meet the needs of the local beneficiary population. CMS also issues national coverage determinations.

CMS can issue guidance that outlines its expectations for data submitted to support coverage determinations. One example is the proposed coverage with evidence development (CED), released for public comment in 2023,¹ a tool that can be used when evidence is promising but incomplete,

___________________

¹ Coverage with Evidence Development Proposed Guidance Document (CMS, 2023). See https://www.cms.gov/files/document/proposed-ced-guidance-6–22–2023.pdf (accessed September 9, 2024).

Ling said. Per the proposed guidance, “the study protocol must explicitly discuss beneficiary subpopulations affected by the item or service under investigation, particularly underrepresented groups in clinical studies, and also describe how the inclusion and exclusion criteria and requirements affect enrollment and also retention,” Ling said. In essence, “the evidence has to be fit for purpose, but it also has to be fit for people,” she said.>

Another CMS tool is issuing a proposed procedural notice, such as the Transitional Coverage for Emerging Technologies pathway for breakthrough devices that was issued for public comment in 2023.² It conveys the agency’s expectations that data submitted for national coverage decisions will be applicable to the patients who will use them. It also provides opportunities for manufacturers to engage with CMS during development regarding their plans for evidence generation.

For Medicare coverage decisions, Ling said, “it is important not only for treatments and services to be safe and effective but also useful for the people who receive them… The outcomes should be meaningful for the beneficiaries.” Krofah noted that CMS coverage decisions can influence the entire commercial market with evidence and data for therapeutic interventions.

Value-Based Payment

The Affordable Care Act provided funding for CMS to test new care and payment models. Ling explained that CMS is moving away from payment based on volume and transitioning to a value-based model. One CMS tool for promoting quality and value is the merit-based incentive program for clinicians, which rewards providers for activities that improve care quality and outcomes. Clinicians are judged according to factors such as care quality, cost of care, promoting interoperability, and activities that improve clinical practice; these criteria were established through the CMS rulemaking process. Krofah noted that during the COVID-19 pandemic, CMS incentivized clinicians to refer patients to vaccine trials by allowing a practice improvement credit. Ling said that activity still qualifies for credit under practice improvement, with a focus on recruiting underrepresented populations. One other clinical trial–related practice improvement activity qualifies for merit-based incentive credit; it relates to trial leadership and designing community-focused trials.

Ling pointed out that practice improvement accounts for 15 percent of the clinician total score, and Krofah asked whether this is sufficient to incentivize primary care clinicians to participate in a clinical trial, given

___________________

² Medicare Program; Transitional Coverage for Emerging Technologies. Centers for Medicare & Medicaid Services (Federal Register, 2022). See https://www.govinfo.gov/content/pkg/FR-2023–06–27/pdf/2023–13544.pdf (accessed September 9, 2024).

their many competing priorities. Ling suggested that practice improvement credits are one way to address the issue but are not the full answer. One consideration is how to better enable clinicians to refer patients to clinical trials during the course of routine care. CMS is working on an activity that hopes to reduce administrative barriers and reporting burdens to give clinicians more time to focus on the patient. Ling noted that CMS annually considers suggestions from the public for merit-based incentive program activities.³

Ling and Krofah discussed the benefits of user-centered design and supportive infrastructure as it relates to clinical data collection and use at the point of care. For example, how can electronic health records be better leveraged as a tool for care, research, and policymaking. Ling noted that the Office of the National Coordinator for Health Information is working on standards for interoperability of health-related data to enable sharing and delineating data elements for health-related social needs, social determinants of health, race, geography, ethnicity, and distinction from race.⁴

Other Opportunities

Cost can be a barrier to clinical trial participation for some patients. This is exacerbated by a lack of clarity about insurance coverage for routine care elements of a trial and out-of-pocket cost sharing (e.g., copayments), Krofah said. She and Ling discussed that, per Section 210 of the Consolidated Appropriations Act,⁵ Medicaid now covers the costs of routine patient care and services for participation. Medicare also covers some routine and “reasonable and necessary” care costs for trial participation, Ling said. However, cost sharing is beyond CMS authority and will require legislation to address (as anti-kickback statutes are involved). Ling and Krofah noted the opportunities to provide patients with greater clarity about their coverage and potential costs associated with trial participation.

Butts shared her takeaway that sponsors should engage CMS before Phase 3 trials. Ling responded that the opportunity now is to converge on key, meaningful outcomes. This entails developing the infrastructure and processes that will facilitate throughput of the evidence to improve care, programmatic functions, and better outcomes for beneficiaries. Krofah and Ling both noted that CMS has a smaller staff than FDA, so those processes must be efficient and integrated into health care across the board.

___________________

³ See https://qpp.cms.gov/mips/mips-value-pathways/submit-candidate (accessed September 9, 2024).

⁴ Discussed by Konya in chapter 3. See the United States Core Data for Interoperability (USCDI) at https://www.healthit.gov/isp/united-states-core-data-interoperability-uscdi (accessed September 9, 2024).

⁵ Discussed further by Chalmers in Chapter 6. Pub. L. No. 116-260, § 210.

Ishwaria Subbiah of the U.S. Oncology Network observed that it is often process inefficiencies that prevent patients from accessing clinical trials. She shared an example in which a sponsor determined that a particular scan for the study was standard of care (and therefore did not pay for it), but insurance companies denied coverage for two patients who wanted to enroll. The clinical team had no means to address this, and the patients were unable to participate. Ling pointed out that CMS largely does not see such layers of coverage decisions that impact patients, providers, and researchers. “A claim that is never submitted will never be seen by CMS. It is a blank spot in our radar screen,” she said. Greater transparency across these layers is needed, so that the factors and decisions causing holdups can be identified and addressed.

FDA AND NIH IN CONVERSATION

NIH Perspective

NIH has invested in increasing access for people participating in clinical research for decades, including policies that require including women and racially and ethnically minoritized populations in NIH-funded research. Bertagnolli noted that, despite recent progress, representative inclusion of many other populations is still lacking, including Indigenous peoples, pregnant women, people who live in rural locations, people with disabilities, sexual and gender minorities, elderly people, and children.

Bertagnolli stated that because NIH primarily invests in basic or discovery science, the agency must be very strategic when funding clinical trials. She described several current NIH initiatives to improve the inclusion of diverse populations in clinical trials.

Community Engagement Alliance Against COVID-19 Disparities

The Community Engagement Alliance (CEAL) Against COVID-19 Disparities was established during the pandemic to directly engage communities in research. Examples of activities include recruiting 2,600 diverse participants for COVID-19 clinical trials and events at which more than 300,000 people were vaccinated for COVID-19, Bertagnolli said. Its activities are expected to continue and expand beyond COVID as a result of the program’s success in engaging diverse populations.⁶

___________________

⁶ See https://ceal.nih.gov (accessed September 9, 2024).

All of Us Research Program

All of Us⁷ has enrolled more than 800,000 individuals across the United States, 80 percent of whom identify with groups underrepresented in biomedical research and 48 percent of whom identify as a member of a racially and ethnically minoritized population. Bertagnolli noted the now numerous publications that use the genetic diversity data from the program.

Minority and Health Disparities Strategic Plan

Including diverse populations should become the norm in all NIH research, Bertagnolli said. A goal of the NIH Minority and Health Disparities Strategic Plan is to increase the overall proportion of patients from diverse populations in NIH-funded research by an additional 40 percent by 2030.⁸ As one example of how NIH will achieve this, she said the agency is entering into clinical research partnerships with American Indian and Alaska Native tribes, Native Hawaiians, and Pacific Islanders through the Native American Research Centers for Health program, which will support care and clinical trials in those communities.⁹

FDA Perspective

Califf agreed with Bertagnolli about progress toward increasing diversity in clinical trials and that areas for improvement remain, especially for the wider spectrum of underrepresented groups (e.g., older adults, rural populations). He mentioned the FDA Center for Drug Evaluation and Research (CDER) Drug Trials Snapshots Summary Report for 2023, which indicates that 48 percent of the participants across the studies included were women.¹⁰ The summary report also shows that for trials enrolling domestic and international participants, 76 percent (median) were not enrolled in the United States. The report listed two products approved by FDA in 2023 that had no U.S. trial participants and five products that had only U.S. participants. Califf pointed out the tension between striving for trial enrollment that is representative of the U.S. population while also developing products that might be used globally.

___________________

⁷ Also discussed by Watson in Chapter 2.

⁸ See https://nimhd.nih.gov/about/strategic-plan/ (accessed September 9, 2024).

⁹ See https://www.nigms.nih.gov/capacity-building/division-for-research-capacity-building/native-american-research-centers-for-health-(narch) (accessed September 9, 2024).

¹⁰ Percentages represent median enrollment excluding programs with sex-specific indications (e.g., prostate cancer, postpartum depression). Drug Trial Snapshots are discussed further by Thanh Hai in Chapter 6. Drug Trials Snapshots Summary Report 2023 (FDA, 2024a). See https://www.fda.gov/media/178602/download?attachment (accessed September 9, 2024).

Bumpus added that certain diseases have had more diverse clinical trials. For example, trials for certain neurodegenerative diseases do not have the same degree of representation as has been achieved for Type 2 diabetes or cardiovascular disease. Diversity in sexual orientation and gender identity has only been considered in a limited number of therapeutic areas. She also raised the importance of intersectionality. For example, a study might filter the data for people over age 65 but not additional subsets, such as what percentage of participants were older people of color. Other areas to improve representation are basic research, biobanking, and genome sequencing, all of which inform drug discovery.

Engaging with FDA

Bumpus reminded participants of the range of ways to engage with FDA. A draft guidance that is open for public comment is released before any final guidance is published; FDA also holds public workshops on various topics. Other opportunities exist through the FDA Office of Minority Health and Health Equity, which provides funding for community engagement, including building infrastructure in communities to enable clinical trials, Bumpus said. Feedback from community development emphasizes the importance of fostering longer-term, deeper relationships with communities, beyond simply building trial capacity and enrolling community members. Communities want to know the results of these trials, how the data are being used, and what impact their participation had.

Data Sharing to Enhance Diversity in Clinical Trials

Data sharing to enhance diversity in clinical trials was discussed. Bumpus noted several barriers to data sharing, including a lack of common formats for data collection, inability to find the data, and difficulty retrieving older data (e.g., on spreadsheets). She suggested that artificial intelligence could be used to help curate data and better enable data exchange. Bumpus also discussed the “richness of data” beyond demographics. Sharing those data can inform future experimental or study design to promote trial diversity.

Bertagnolli emphasized the need for collecting complete demographic data to understand which subpopulations are underrepresented or missing in a trial and enable cross-study analysis. She noted that NIH is working on ways to capture data on sexual orientation and gender identity for certain studies. She said trials should be fit for purpose when it comes to collecting data to address the needs of subpopulations (rather than relying on posthoc meta-analyses). She added that, as of January 2023, data from federally funded research must be made publicly available, and funding applicants

should describe how the results will be disseminated.¹¹ The challenge will be how to define and implement standards and infrastructure to enable such sharing.

Califf said that making data on trial diversity available to the public would help promote accountability and be useful for determining appropriate metrics for progress. He pointed out that many of the data from industry-sponsored clinical trials of products in development are considered to be confidential commercial information, and CDER Drug Trials Snapshots only include drugs that have been approved by FDA. Given that 90 percent of investigational drugs that enter Phase 1 clinical studies are not approved and data from failed or otherwise suspended programs are not publicly shared, this can make it difficult for the clinical research system as a whole to learn from what did not work. Califf pointed out that FDA’s ability to share data, including confidential commercial information, is limited by federal law.

Califf noted that a major difference between clinical trial data-sharing policies and capabilities between NIH and FDA is that NIH has more influence over investigators through funding criteria. FDA does not fund clinical trials. It also works with a global industry and does not have the authority to tell companies what they can or cannot do in other countries. The specific circumstances under which FDA can require trial participant data when it comes to age and pregnancy are also limited. He noted that industry is making good progress in meeting sex, race, and ethnicity criteria.

Interagency Collaboration

Bertagnolli discussed ways that NIH can work across government, including FDA, to support clinical studies and access. For example, NIH works with FDA to ensure that NIH-funded clinical trials of a new product or new indication are designed to support timely approval and uptake into clinical care. This might include pragmatic studies in diverse populations. A particular challenge is clinical trials for drugs to treat rare diseases or conditions. She said NIH, FDA, and other government agencies can work together to ensure access to these products, which can be very expensive. Califf noted that NIH is a leader in promoting discovery science that can be used by industry to develop new medical products.

Califf discussed that over half of medical product approvals use an accelerated pathway, which can initially allow for more uncertainties about benefit or risk to patients. He said that FDA, NIH, CMS, private insurers,

___________________

¹¹ Ensuring Free, Immediate, and Equitable Access to Federally Funded Research (OSTP, 2022). See https://whitehouse.gov/wp-content/uploads/2022/08/08–2022-OSTP-Public-access-Memo.pdf (accessed September 9, 2024).

and key entities can come together to address the barriers to completing postapproval studies because these data are needed to better understand who benefits from the treatment and how best to use the product in clinical practice.

Capacity for Community Engagement

Bumpus shared that the FDA Office of Minority Health and Health Equity funds community engagement work, including in rural communities. That work has shown that more effort is needed to build infrastructure that broadens opportunities to participate in clinical trials, including what locations can be a site, what health care professionals can participate, and awareness of opportunities to participate. Lessons learned include using networks to increase practitioner participation, building long-term relationships that last beyond trial initiation and recruitment, as well as explaining to people how their participation has an impact, including how their data are used. Bertagnolli noted that a major challenge to high-quality research is insufficient capacity for quality clinical care in rural locations and communities that are undergoing “particular crisis of need.”

Bertagnolli and Jesse Nodora, associate director for Community Outreach and Engagement at Moores Cancer Center at the University of California, San Diego, pointed to the Native BioData Consortium, a nonprofit research institute led by U.S. Indigenous scientists and tribal members to enable Indigenous sovereignty over biosamples and data ownership.¹² Bertagnolli said NIH is engaging with “Native American communities to develop ways that we can work effectively with these sovereign nations in order to advance health for everyone.” That includes her support for the concept “nothing about us without us when it comes to biomedical research.” She said that NIH is listening, engaging, and working with Tribal Nations, American Indian and Alaska Native populations, and Indigenous communities.

U.S. and Global Clinical Trial Populations

As sponsors develop diversity action plans, one consideration is how to determine the ideal trial composition for studies submitted for U.S. regulatory approval even when products are likely to be used globally. Califf said these discussions will evolve as diversity action plans are developed, and decisions will be based on both data and values (i.e., ethics). He expressed his view that much larger trials are needed to learn about heterogeneity of treatment effects. He noted that FDA publications now call for trials

___________________

¹² See https://nativebio.org/ (accessed September 9, 2024).

to be “representative of the people for whom the treatment is intended,” not the population of the United States. Bierer noted the challenge that epidemiological data regarding the demographics of the impacted population are often lacking, “particularly by subpopulation and even worse by intersectionality.”

This page intentionally left blank.