Proceedings of a Workshop—in Brief

Convened October 29, 2024

Exploring Clinical Guidelines for the Adoption of Genomic Testing
Proceedings of a Workshop—in Brief

Clinical practice guidelines, referred to here simply as guidelines, are condition- or disease-specific recommendations for the appropriate health care management of patients, and are based on the best available evidence.¹ Clinical guidelines are most often developed by professional societies and may be classified as “evidence-based” or “consensus-based,” which can lead to confusion (Djulbegovic and Guyatt, 2019). Guidelines produced by different organizations can be incongruent even when using the same evidence. On October 29, 2024, the National Academies of Sciences, Engineering, and Medicine’s Roundtable on Genomics and Precision Health and National Cancer Policy Forum jointly convened a workshop to explore how guidelines can impact adoption of genomic testing into routine medical care. Previous work of the roundtable illuminated how inconsistencies across guidelines may impact adoption of genomic testing and coverage decisions by payers,² said Victoria Pratt, representing the Association for Molecular Pathology (AMP) and director of Scientific Affairs for Pharmacogenetics at Agena Biosciences. Goals of this workshop included exploring how guidelines are developed by various organizations, how different relevant parties use guidelines, and how incompatibility of guidelines affects their work. This workshop will explore what could be done to advance compatibility and usability of guidelines, implement them into the clinic, and achieve coverage for recommended genomic testing for all, said Mylynda Massart, associate director of clinical services at the Institute for Precision Medicine, medical director of the UPMC Primary Care Precision Medicine Clinic, and associate professor at the University of Pittsburgh, at the start of the workshop.

WHY GUIDELINES MATTER FOR GENOMIC TESTING

Guidelines shape patient care, which is why it is important to write guidelines that will be implemented in a way that all those who require care, receive it, said Robyn Temple-Smolkin, senior director of clinical and scientific affairs and director of guideline development at AMP. Developers of genomic testing guidelines face a range of challenges, including the volume and complexity of genomic data across many clinical areas; lack of high-quality clinical evidence and implementation into routine care; concerns related to diversity, equity, ethics, and social implications; and ability of rapidly evolving technologies to be used for multiple applications. Despite these challenges, developers aim to create guidelines

__________________

that supplement the knowledge and skills of health care providers and that are tailored to reach relevant parties, including payers, patients, test developers, and clinicians, she said.

The guideline development process is grounded in considerations of conflict-of-interest and transparency, Temple-Smolkin said. Professional and medical societies set priorities and convene panels of experts to develop guidelines. When developing or reading a guideline, it is important to understand the patient, population, or problem; intervention; comparison; and outcome (PICO) that the guideline is intended to address. The PICO framework³ guides a systematic review of the literature to assemble relevant evidence that informs the development of clinical recommendations. Some developers choose to use the living systematic review methodology, which involves a literature search more frequently than typical systematic reviews (e.g., every 3 months), to analyze new evidence and determine whether recommendation updates are warranted.

One example of an evidence-based guideline development process is Grading of Recommendations Assessment, Development, and Evaluation (GRADE),⁴ Temple-Smolkin said. The GRADE approach is used to review the certainty of evidence which will inform the development of clinical recommendations. Evidence is “upgraded” if there is more certainty (e.g., a large magnitude of an effect), or “downgraded” if there is less certainty (e.g., there is risk of bias, or data are inconsistent or indirect). In the GRADE approach, once recommendations have been developed, their strength is determined based on the certainty of the evidence along with considerations of benefits, harms, burdens, patient preferences, resources, and cost. It is also important to understand the value of interventions and outcomes to patients, she said. Recommendations have been changed because of input from patients. An intervention will not be helpful, regardless of expert recommendation, if patients do not want it, due to inaccessibility or unwanted side effects, for example. To effectively communicate guidelines to interest holders, regardless of development approach used, developers can use language and structure that identifies whom the recommendation is for, what is recommended for or against, strength of the recommendation, certainty in supporting evidence, complementary guidance, and who made the recommendation. To promote transparency, GRADE requires the development of a “Summary of Findings” table to provide information about how the working group came to their final recommendations. Not all guidelines are developed using the GRADE process but there are key components to developing trustworthy guidelines⁵ that should be followed, Temple-Smolkin said. Tools are available to help individuals determine if a guideline has followed these standards. There can be differences in recommendations across guidelines on the same topic, even when the same literature has been evaluated. To understand the differences, PICO and other key attributes can be compared,⁶ Temple-Smolkin said.

Guidelines may have long development times due to rigorous methodology, multidisciplinary input, complexity of topics, volume of evidence, extensive review process, resource constraints, and conflict resolution, Temple-Smolkin said. Guidelines must be updated, and updates can be provided more quickly by employing living systematic reviews and living guidelines and using emerging technologies, such as artificial intelligence (AI) to assist in literature searches, she noted. Practice guidance for clinicians can also be found in guidelines and position papers. These tools have different levels of evidence assessment stringency but can be useful in genomics where practice guidance is needed but the literature and evidence are often emergent, inconsistent, or limited, she said.

Perspectives of Interest Holders in Guideline Development

Patient and Family Perspective

To achieve the best patient care, it may be necessary for guidelines to account for diversity of populations and to

__________________

follow systematic processes and strict methodologies, said Vimal Scott Kapoor, chief occupational physician for Bruce Nuclear Power, aviation physician for the U.S. Federal Aviation Administration and Transport Canada, emergency physician at Markham Stouffville Hospital, and clinical assistant professor at the University of Toronto. Kapoor shared a personal story of the death of his brother, Anil Kapoor, a urologist, as a vivid example of the human cost of lacking, flawed, or poorly implemented guidelines. After being diagnosed with colon cancer, Anil Kapoor began treatment with fluorouracil (5-FU), and within 45 days he fell ill and passed away from a reaction to the treatment. Doctors can adjust the dosage of 5-FU if a blood test for a genetic variant in the DPYD gene, which leads to deficiencies in the enzyme activity that breaks down 5-FU, is performed. Dosage adjustments can be made for patients who will likely be poor metabolizers to avoid potentially fatal toxicities. Pretesting for these variants is feasible but not yet incorporated consistently into guidelines, Kapoor added. Such testing is routinely done in many European countries, several Canadian provinces, one U.S. state, and at a growing number of institutions. However, he stated that most physicians test for only a handful of potential variants, and genetic diversity needs to be considered in these tests to increase equity. Kapoor noted that his brother was pre-screened before treatment, but the DPYD variant he had was not included in the panel, even though it had been previously identified.

Payer Perspective

It is important to include interest holder perspectives when developing guidelines, especially those of clinicians, patients, and payers, said Lindsay Zetzsche, owner of Science Geek Games, consultant at Integrity Genetics Consulting, LLC, and former genetics director at a large payer. Different payers evaluating the same evidence base and practice guidelines can come to different coverage decisions, she said, adding that biomarker laws present an additional layer of consideration in these decisions. Incongruent guidelines present a particular challenge where payers decide which guideline to follow by examining the studies included, how the evidence was rated, objectivity, panel member conflicts of interest, and other factors. Payers are interested in opportunities to both reduce costs and increase quality of care, she noted. People who have their own biases are behind the creation of guidelines, medical policies, and coverage determinations; therefore, being objective about levels of evidence and comparisons between conflicting guidelines is important because patients are impacted by the decisions made.

Industry Perspective

Mismatches between genetic ancestry, which can indicate risk, and self-reported ethnicity, which can influence the testing a patient receives, can lead to discrepancies in patient care, said Aishwarya Arjunan, genetic counselor and senior medical science liaison at GRAIL, Inc., speaking about previous work done at Myriad Genetics. Additionally, the conflation of race, ethnicity, and ancestry by some guidelines can exacerbate racial and ethnic health disparities. Guidelines could be re-examined to identify where race, ethnicity, and ancestry are used to dictate the care provided and to question why that is and whether it is causing harm, she suggested. Genetic testing is ordered by a range of professionals, each with their own professional society issuing guidelines. Collaboration on guideline development helps all interest holders, including guideline developers, who can save resources and time. It is important to build trusted partnerships with patients and communities, developing guidelines with them rather than for them, Arjunan emphasized.

Non-Genetics Clinician Perspective

Given the lack of genomics providers and barriers to accessing them in the United States, enabling test ordering through non-genomics clinicians could enhance patient access to and implementation of genomic testing, said Brianne Phillips, nurse practitioner at the University of Pittsburgh Medical Center. However, there are barriers that non-genomics providers face, including a lack of knowledge or confidence in ordering genomic tests, limited guidance in their society’s resources, and payer restrictions. Clinics that order genetic testing daily can still encounter challenges of inequitable reimbursement, where different payers may cover different tests and criteria for coverage may be unclear or inconsistent. The development of genomic testing guidelines in primary care by a relevant society might be needed, she suggested.

Collaboration, Implementation, and Education for Advancing Guidelines

Many challenges faced in the guideline development space could be reduced through collaboration. Making a commitment as an organization to facilitate collaboration is important and can produce a better guideline product, Temple-Smolkin said. Collaboration across organizations can also reduce redundancy of efforts and lower costs, which can present a challenge, for each organization, Arjunan suggested. Improved communication about ongoing and planned guidelines can help identify opportunities for collaboration, she said. For payers, developers can do some of the legwork by figuring out what evidence payers need, how to approach them, and how to help them consider cost-saving and medically beneficial opportunities, Zetzsche said.

Planning for implementation during the guideline development process is key for recommendations to reach those who need them. Providers could be engaged through dissemination beyond journal publications, including through infographics, collaboration with patient advocacy groups, and targeted education, Temple-Smolkin suggested. Making reports more accessible and educating physicians about how to read, understand, and implement guidelines may also help providers. Educating patients, families, and communities could empower them to have meaningful conversations about genetic testing with their clinicians, Arjunan said. Speakers discussed elements of the development process that they would like to change, and those ideas are captured in the “Looking Ahead” section.

CURRENT GUIDELINE DEVELOPMENT FOR GENOMIC TESTING

Incorporating Health Equity into the Guideline Development Process

Health equity describes care that is proportionate to the need, which means equality (e.g., equal access) is not sufficient. Health equity considerations could be intentionally built into the entire guideline process, said Jennifer Lin, director of the Kaiser Permanente Evidence-Based Practice Center. Health inequities are avoidable and reparable differences in health across economic, social, geographic, or demographic groups. Addressing health equity in guidelines, from topic nomination to dissemination and implementation of recommendations, could result in elimination of differences in clinical risk factors based on social determinants of health, in delivery of health care, and in disease incidence and related patient outcomes. An example health equity framework for guideline development created for the U.S. Preventive Service Task Force includes action items for each phase of development and a list of extra resources needed to implement these actions (Lin et al., 2024).

Throughout the guideline development process, individuals with health equity expertise or training and those with lived experiences could be engaged, Lin suggested. Health equity can be incorporated as a criterion in topic selection and prioritization. During the planning phase, broader questions around upstream drivers of health, implementation, and downstream care can be considered. It is important to understand the applicability of the available data to various populations, considering the representation of those populations and the coded biases within the data, she said. Guidelines also have opportunities to affect equitable implementation of recommendations; to discuss gaps in evidence, policy, and practice; and to identify areas for research. Guideline panels should ensure “that their recommendations do not perpetuate or exacerbate inequities,” she said, and they should strive to develop recommendations that mitigate inequities. Incorporating equity will require additional resources, therefore progress can be incremental as resources allow.

Guideline Development for Pharmacogenomics

Getting evidence into practice is challenging in medicine, particularly in pharmacogenomics, said Kelly Caudle, multi-principal investigator and director of Clinical Pharmacogenetics Implementation Consortium (CPIC) and associate member of St. Jude Children’s Research Hospital. CPIC creates guidelines that help clinicians know what to do with genetic results and provide recommendations to guide prescribing based on test results. The CPIC guideline development process is rigorous, has a system to grade evidence and recommendations, and incorporates equity throughout (Caudle et al., 2014). However, establishing multidisciplinary writing committees in pharmacogenomics without conflicts of interest is challenging because of the increased number of experts working at or consulting for pharmacogenomic testing companies, she added.

More resources are needed to support a timely and efficient guideline development process that can keep up with rapidly evolving fields and engage with patients, Caudle said. Due to study design issues (e.g. sample sizes), cost, and the ethical issue of control groups not receiving testing, evidence from randomized clinical trials (RCTs) is lacking in pharmacogenomics. Phar-macogenomic guideline developers could benefit from solutions that address heterogeneity across studies which can create hurdles when evaluating outcomes from tests and treatments across studies. Guideline developers can be transparent about how data are extrapolated from specific populations studied to the broader population. Standardization of clinical laboratory genetic testing and results reporting could lead to more widespread adoption of pharmacogenomics, she suggested.

Guideline Development for Genomic Tumor Testing in Oncology

Many of the previously mentioned challenges for genomic testing guidelines also affect guideline development for genomically informed targeted cancer therapy, including a lack of RCTs and experts free from conflicts of interest, said Funda Meric-Bernstam, chair of the Department of Investigational Cancer Therapeutics and medical director of the Institute of Personalized Cancer Therapy at MD Anderson Cancer Center. RCTs are lacking because they are either not feasible or present ethical concerns related to a control group in cases where if a test is available to inform treatment, then all participants should be tested, she said. To overcome feasibility challenges, genomically informed clinical trials of targeted therapies are using novel trial designs, such as basket trials and umbrella trials.⁷ Clinical utility is a necessary metric in guidelines, and for genomic testing, it can be affected by the odds of identifying actionable genomic alterations and the availability of an approved, genomically matched therapy. Disparities in access to genomically matched therapy and clinical trials present an additional challenge in guideline development. New approaches are needed for establishing data that are of sufficient quality to support clinical recommendations and for streamlining guideline development and translation, she added.

ACMG Practice Resources

Expert statements can provide clinicians with timely information on genomic testing before resources and systematic evidence review permit the development of comprehensive guidelines, said Heidi Rehm, director of the Genomic Medicine Unit at Massachusetts General Hospital, institute member and clinical laboratory director at the Broad Institute, and professor of pathology at Harvard Medical School. In addition to evidence-based guidelines, the American College of Medical Genetics and Genomics (ACMG) also releases professional practice resources, consensus technical standards, and points-to-consider statements to give guidance to clinicians and laboratories on clinical and laboratory practice where there is a lack of systematic evidence-based review. During ACMG’s guideline development, patients are represented on the workgroup, topic ideas are solicited from the community, and collaborations with other professional societies occur. Expanding partnerships of ACMG members with non-genomics professional societies on the genomic testing aspects of their recommendations can ensure the necessary expertise is represented, Rehm suggested.

Rehm cited the American Society of Clinical Oncology’s (ASCO’s) work with the Clinical Genome Resource (ClinGen) as an example of an effective collaboration. ASCO used resources produced by ClinGen to inform its guidelines on which genes should be included on a cancer panel, Rehm said. The ClinGen resource is an evidence-based framework for evaluating gene–disease relationships to support clinical utility. This resource has informed some but not all professional society guidelines and served as a technical standard for laboratories on which genes to include in testing panels. To enhance the evidence base for resources and guidelines, data sharing is essential, Rehm suggested. Finally, she reflected that, “implementation is important to ensure that clinical utility is effectively achieved once [recommendations are] put into practice.”

Facilitators and Barriers to Guideline Development

The increased interest in genomic testing has been a facilitator of guideline development despite the perceived lack of high-quality data, Meric-Bernstam noted. Lin highlighted the importance of stable funding and commitments to developing infrastructure within guide-

__________________

line developing organizations. Guideline developers can struggle to prioritize health equity, but lessons can be learned from processes used by other guideline groups, Lin said. Developing standing agreements across societies can quickly facilitate collaboration, Rehm said. Additionally, having pharmacogenomics or genomic representation on disease-specific guideline panels could help facilitate inclusion of genomic testing in discussions and dissemination of information about genetic testing, Caudle suggested. The informed consent process could be streamlined to overcome the barrier of requirements for consent when ordering genetic and genomic testing and return of results which can dissuade some physicians, Rehm added.

IMPACT OF GENOMIC TESTING GUIDELINES ON PAYER DECISIONS

Using Guidelines to Drive Payer Coverage Policy

Payers use guidelines to inform their coverage decisions which can contribute to a feedback loop where the more payers use guidelines, the more societies will be encouraged to invest resources into them, hypothesized Gillian Hooker, chief scientific officer at Concert. Although payers use guidelines frequently, they are not always developed with the payer in mind, therefore, more test ordering software could integrate payer policy to help clinicians and patients understand the coverage of a test and anticipated cost to the patient, Hooker noted. Concert’s reference policy documents interpret current evidence-based clinical guidelines to provide clarity and to support the automation of coverage decisions.

Additional challenges for payers in applying guidelines to the development of coverage policy include inconsistent, discrepant, and vague language in guidelines; a lack of clarity about which audience the recommendations were written for (e.g., clinicians); the developer’s choice of clinical outcomes expected from the audiences following the recommendations; and timeliness. Translating inconsistent language, both within and between guidelines into a standard format could help policymakers, Hooker suggested. Discrepancies even occur across guidelines from the same organization and can result in contradictory recommendations. Standard language, such as saying that a guideline organization “recommends,” “recommends consideration of,” or “does not recommend,” could be used across all guidelines. To address contradictory recommendations, policy makers can review the evidence directly and contact the organization for clarification, or organizations could ensure their own guidelines align with each other. Vague language, even though used to protect clinician decision making in the practice of medicine, can be a problem for policy makers’ decisions, so clinical observations could be defined into discrete fields, she said.

If it is not clear for whom a set of guidelines has been written, errors can arise in the interpretation of guidelines since assuming one audience or another can lead to different conclusions being reached. Hooker suggested, guideline developers can clearly define how different interest holders should interpret and read the guideline recommendations. Diagnostic yield,⁸ a common outcome chosen in guidelines for genomic testing, may not exhibit enough clinical impact to drive coverage of testing. Guideline developers could chain evidence together to link diagnosis to clinical impact, such as avoidance of unnecessary treatments, or could identify more clinically impactful outcomes to drive coverage. Payers struggle to keep up with and implement coverage policy that is based on frequently updated guidelines; updating twice per year is ideal, she suggested. Old guidelines need to be updated, retired, or reaffirmed, Hooker said, and guideline deserts, conditions for which there are no guidelines being developed, need to be identified. Organizations involved in diagnostics (e.g., AMP, ACMG) can partner to identify and prioritize which guidelines will be most valuable in addressing guideline deserts.

The Social Science of Evidence Development and Adoption for Payers

To help payers better use guidelines, guideline developers could follow the Transparency, Redundancy, Utility, and Efficiency (TRUE) approach, suggested Trent Haywood, founder of Knowality, LLC. TRUE is Haywood’s social science-based approach to evidence development and adoption which promotes habit formation that leads to building community trust. Transparency involves guideline developers being transparent about the timing and availability of new guidelines to align with when payers

__________________

examine evidence and make coverage decisions, which Haywood said affect the health care coverage market and adoption of recommendations. Being transparent also provides a level of consistency and accountability. One practical challenge to transparency is accessibility of guidelines that may be behind journal paywalls, he noted. Redundancy occurs when interest holders are aligned on their standards of what the evidence says and what patients and communities need. When test developers and manufacturers, other payers, clinicians, and patients disagree about the state of the evidence, it is more difficult for the payer to come to a clear coverage decision. Payers also need to differentiate between what current clinical practice is and what the research is believed to be showing, he added.

Utility refers to the clinical utility of a test; guidelines need to clearly convey the extent of the evidence supporting clinical utility, Haywood said. Utility is associated with efficacy and value—specifically, anticipated benefit relative to cost. In addition to using the PICO framework, he continued, it is important to consider the “number needed to treat and number needed to harm,” which allows the payer to evaluate whether a test is an improvement or not relative to other interventions. Efficiency in guidelines is essential for payers. To support adoption of the recommendations by payers, developers can consider how guidelines can be efficient from procedural and resolution standpoints. Following TRUE when developing guidelines can help payers make coverage decisions and adoption of recommendations more efficient, he suggested.

Lessons Learned from Coverage of Germline and Somatic Testing

The Mercy health system has consistently obtained coverage for germline and somatic genomic testing for cancer care when the testing guidelines were met, said Gautum Agarwal, director of Precision Medicine at Mercy, sharing experiences from his provider group. To provide this care, Mercy relies on a primary care network, system-wide electronic health record (EHR) platform, and cloud-based intelligent data platform. Using National Comprehensive Cancer Network guidelines has led to coverage of germline genomic testing. Partnering with genomic testing and AI technology companies has helped bring in new tools for patient care and has advanced the availability of genomic testing, Agarwal said. Cascade testing for the family members of those who receive positive germline results is offered at no cost to first-degree relatives. Partnering with one large next-generation sequencing company enables Mercy to standardize congruence with guidelines and output formats, he added.

Germline screening provides patients and their clinicians the opportunity to mitigate cancer risk, Agarwal said. An opt-in questionnaire can standardize and automate the identification of patients who meet guideline criteria for testing (specifically for breast, ovarian, pancreatic, colorectal, endometrial, and gastric cancers). Testing is integrated into the EHR system to facilitate ordering and viewing of results by clinicians, provide clinical decision support tools, and link to a central precision medicine team for pre- and post-test counseling, he explained. Somatic testing of cancerous tissue informs treatment choice, and by following clear and actionable guideline criteria, Mercy has been successful in getting testing and targeted therapies covered by insurance. Beyond oncology, genomic testing spans the entire practice of medicine and can be incorporated into each discipline’s practice guidelines, he said. Patients can be empowered to ask about genomic testing through outreach and education, he added. Screening for diseases early has a measurable economic impact due to fewer work hours lost, less disability, and less cost in general, he said.

Coverage Considerations

Test developers could have discussions about coverage during the study design phase, so that the collection of evidence sufficient to support coverage can be incorporated into the study and guidelines up front, Haywood said. Reimbursement is not just about policy, Hooker added, explaining that there can be many reasons coverage is denied, such as prior authorization. Early genetic testing can affect outcomes and costs over the long term, Agarwal noted, while Haywood said that hospitals can take opportunities to demonstrate to payers how the use of a test early can lead to costs savings overall. To address variability across payer policies, he continued, the roundtable could convene interest holders to discuss the practical issues of promoting consistency among payers including tradeoffs between predictability and

flexibility and whether consistency across payers might violate antitrust laws.

CLINICAL CARE IMPLEMENTATION OF GUIDELINES FOR GENOMIC TESTING

Implementation Science and Cancer Genomics

The goal of implementation science is to make the application of a guideline or tool within care settings successful in reaching as many eligible people as possible, said David Chambers, deputy director for implementation science in the Division of Cancer Control and Population Sciences at the National Cancer Institute. Implementation science involves understanding the intervention, parties affected, implementation challenges, and desired outcomes; developing strategies to improve implementation; and establishing measures of success for those strategies. For an evidence-based genomic intervention to benefit patients, not only must it be implemented in a way that it reaches all those who are eligible to receive it, but there must also be trained clinicians who are able to deliver it within their practices, Chambers suggested. Implementation ultimately succeeds when improvement is seen at population levels, people are living their best lives, there is equity in care, and health benefits are seen, he added.

It is important to think of implementation of services throughout the guideline-based genomics care pathway, from risk assessment to education and consent, testing, delivery of results, and risk management and care, Chambers said. Considering implementation of services beyond the guideline and where patients may be lost to care at each step of the process could lead to the development of recommendations to address these issues. Guidelines are generally written to be broadly applicable, or one-size-fits-all. It may be necessary to adapt or tailor the application of the evidence and the intervention according to certain factors, including the target population, service setting, mode of delivery, and culture, he said, adding that it may also be necessary to balance evidence that is generally useful versus evidence that meets case-specific needs. Those implementing guidelines must also grapple with an ever-evolving and dynamic world of genomic research. This is where a learning health system approach could work synergistically with implementation science and precision medicine to help the system navigate changes over time, he said.

Lessons Learned from Sickle Cell Disease Screening

Although there are clinical care guidelines for sickle cell disease (SCD), there is minimal evidence or guidance regarding sickle cell trait,⁹ said Charles Jonassaint, associate professor at the University of Pittsburgh. The historical discriminatory experience that those in African American communities have had with sickle cell screening offers lessons for improving genomic testing in populations experiencing health disparities and for developing and implementing inclusive clinical guidelines. When newborn screening for SCD was implemented, there were no guidelines for screening and some states mandated screening for adults. These programs targeted only African American communities even though sickle cell is widespread within more populations in and descendants from global regions with higher malaria concentration. Exacerbated by a lack of education of the public about what sickle cell trait means, African American communities have been concerned about discrimination, he explained, which highlights the importance of cultural competencies in guidelines. Screening in adults is now voluntary and supported by SCD education, counseling, and research programs. Nearly all newborns in the United States are screened, but that does not translate to parents being screened, meaning most African Americans—the population most affected by SCD in the United States, do not know their sickle cell status, and many are hesitant to participate in education, counseling, or screening due to continued concerns about discrimination. Public understanding remains limited, and “proper counseling and education to mitigate the potential negative social impacts of having sickle cell trait” is needed, he suggested.

Jonassaint shared some ideas for improving the development and implementation of guidelines:

• Considering risks of negative social implications linked to genomic testing results.
• Developing innovative approaches for raising awareness and educating the public, including use of popular media.

__________________

• Engaging the community in all phases of guideline development and implementation.
• Addressing the lack of diversity, including socioeconomic diversity, in clinical research and reporting.

It can be hard for communities to engage with guidelines based on poor-quality evidence or on evidence that is based on a narrow scope of the population and that excludes other factors such as structural racism or social determinates of health. Evidence from trials that include historically underrepresented populations is essential for developing more broadly relevant clinical guidelines, Jonassaint said.

Cardiology Guidelines in Practice

While there are other resources to inform practice, such as scientific statements, that do not necessarily require RCTs, clinicians follow professional society guidelines that are dependent on evidence from RCTs, said Naveen Pereira, professor of medicine at the Mayo Clinic College of Medicine. A lack of RCTs that demonstrate clinical utility, where a genomic test result alters care strategy or therapy and subsequent clinical outcomes, leads to a lack of coverage, which presents a challenge in implementing genomic testing in cardiology, Pereira said. A clinical example where this challenge has arisen is the case of clopidogrel, the most commonly used anti-platelet drug. Approximately one in three people carry a genetic variant that leads to poor metabolism of this drug. The U.S. Food and Drug Administration (FDA)-approved label for clopidogrel contains a black box warning highlighting the availability of genetic tests for CYP2C19 to identify poor metabolizers and stating that clinicians should consider alternative therapy for these individuals.

There are challenges in conducting RCTs to demonstrate clinical utility of a genotype-based therapy to support guideline inclusion. For example, allele frequency drives the required sample size, and the number of participants needed for a sufficiently powered study can be unfeasible. Randomizing participants to conventional therapy versus genotype-guided therapy can be unethical and require complex trial designs. More funding may be necessary for these types of studies, systematic reviews, and meta-analyses, Pereira suggested. Additional opportunities that could bolster clinical implementation of genomic testing guidelines include rapid turnaround genotype testing, incorporation of results to the EHR, point-of-care alerts to the clinician, coverage, and education of clinicians and patients, he said. Implementing genomic testing at the point of care may necessitate governance at the institutional level by a multidisciplinary committee, buy-in from clinicians, education, laboratory capability, and informatics and information technology resources to enable EHR-embedded clinical decision support.

Data for Guideline Development

Several participants and speakers identified the lack of data sharing across systems as a barrier to realizing the benefits of preemptive genomic testing at scale. Incremental change could be used to achieve crosstalk in the U.S. health systems, Pereira said. In the interim, mobile apps can allow patients to see their clinical test results on their phones, giving them the opportunity to refer to their results when seeing a doctor in another system, he added. Social determinants of health data could be considered in guideline recommendation development and clinical care to provide information on how the environment plays a role in health, Jonassaint said. If standards for evidence move beyond RCTs, professional societies could leverage the real-world evidence in large datasets, such as from the National Institutes of Health’s All of Us Research Program, to support guideline development, Pereira added. This database includes evidence, from a variety of populations, including whole-genome sequencing and EHR data with information about socioeconomic status and other social determinants of health.

GUIDELINE DEVELOPMENT IN A RAPIDLY EVOLVING FIELD

Genomic Testing for Diagnosis and Treatment of Rare Diseases

Before 2020, genetic studies lacked outcome reporting, leading to little evidence related to outcomes, said Murugu Manickam, a clinical geneticist at Nationwide Children’s Hospital and associate professor of clinical pediatrics at Ohio State University College of Medicine. During the development of the 2021 ACMG evidence-based guideline on pediatric exome and genome sequencing, the evidence base focused mainly on describing disorders and their links to genes. Due to the gaps in evidence related to outcomes, he said, the guideline recommends reporting outcomes, such as changes

in clinical management, stemming from actions taken based on sequencing results. Increased reporting related to outcomes is starting to be seen; however, recording outcomes beyond diagnosis is still a challenge for rare genetic diseases.

Only 5 percent of people with rare diseases have FDA-approved therapies available (Fermaglich and Miller, 2023). There is a need to move from gene discovery to treatment more quickly and one possible solution would be repurposing existing approved therapies for genetic disorders that affect the same pathway as the rare disease, suggested Manickam. Additionally, bone marrow treatment, gene therapies, and RNA-based therapies may offer additional treatment options. For those still without treatments, outcomes can also be measured in terms of palliation or avoiding unnecessary procedures. Proper follow-up care after genomic testing may require expertise in managing a rare disease, he added; without access to individuals with such expertise, patients may not get the help they need. If the goal of genomic testing is diagnosis, then studies can provide evidence focused on defining and improving diagnostic yield with innovative technologies, improving interpretation of variants to increase equity, and increasing access to testing, he said; however, if the goal is treatment, improvements to the percentage of treatable rare disorders must be addressed through testing to identify eligible patients for clinical trials and outcomes.

Systematic Review Considerations

Evidence frameworks for evaluating genetic tests have existed for some time, said Karli Kondo, director of evidence synthesis at the American Cancer Society. For example, the analytic validity, clinical validity, clinical utility, and ethical, legal, and social issues (ACCE) evidence framework for evaluating genetic tests was developed 20 years ago. Similarly, nearly 15 years ago, the Agency for Healthcare Research and Quality’s (AHRQ’s) evidence-based practice center methods workgroups were grappling with and developing analytic frameworks for conducting systematic reviews evaluating genetic tests, taking into consideration ethical, legal, and social issues (ELSI) as well as financial and operational impacts. Current systematic reviews don’t formally include studies examining ELSI; however, they are relevant for guideline development as they relate to incorporating equity and patient preferences and values.

In fields that may evolve rapidly, such as genomic testing, guideline developers and other interest holders may only have insufficient evidence to work with, Kondo said. Evidence might be insufficient for a variety of reasons: there is only a single study, two conflicting studies, available studies have small sample sizes, or studies available are early phase diagnostic accuracy, observational, or case studies. In cases such as these, it is especially important when determining the strength of the evidence to look beyond the published research literature. Other sources of evidence include grey literature such as article preprints, media, protocols, clinical-trials.gov, and 510k premarket submissions to FDA. Other approaches to evidence gathering include reaching out to experts at professional societies, test developers, and other interest holders for more information that may not have been published. Living reviews may be a solution in the rapidly evolving field of genomics that facilitate guideline updates, Kondo suggested. A medical librarian can help iteratively refine the evidence search process to keep search parameters current. While living reviews are resource heavy, they can provide efficiency as evidence is continuously surveilled, keeping conclusions up to date.

A Vision for the Future of Guidelines for Genomics

Throughout the day, participants heard about the state of guidelines and the many challenges facing the field, said Sandra Zelman Lewis, past president and founder of EBQ Consulting, LLC, sharing her vision for the future while reflecting on the ideas that others had presented (Box 1). The Trustworthy Consensus-Based Statement (TCBS) process supports guideline development, in rapidly evolving spaces such as genomic testing, when the evidence is still growing and will not support meta-analyses. TCBS is evidence-informed and consensus-based and is based on five pillars, she said: panel structure and membership criteria, systematic literature searches, Delphi-based consensus process, transparency of data and methods, and rigorous review procedures. In this approach a qualitative review of existing evidence is shared with panel members as they are drafting recommendations, the methodology is rigorous, and true guideline recommendations that are GRADE-compliant are produced.

AI tools will likely help guideline developers in monitoring and screening new evidence and other time-intensive tasks, Lewis said, noting that using AI will require objective review, replicability, and oversight for accuracy and reliability checks. Ultimately, the essential element for achieving Lewis’ vision, outlined in Box 1, would be funding mechanisms to enable guideline developers to conduct and maintain living systematic reviews and guidelines. While it is a conflict of interest for industry to fund guideline development, they can fund research, particularly on populations with varying genetic propensities, and publish all results, both positive and negative, to inform guidelines, she suggested.

LOOKING AHEAD

Although speakers did not directly compare case examples of how and why guidelines may be inconsistent, speakers did address potential solutions to enhance guideline development for and adoption of genomic testing (Box 2).

During each panel discussion, speakers were asked how, from their perspective, the needle can be moved in guideline development to create more equitable and efficient care in genomics. To incorporate equity into guidelines, developers can learn from, and incrementally add, frameworks and exemplar processes where equity has been included, Lin said. Sharing data and evidence can

also support equity in guideline development, Manickam and Rehm said. To better share data and evidence, data abstracted in a systematic review for the development of one guideline could provide a starting point for other guideline developers, Kondo noted while highlighting the AHRQ Systematic Data Review Repository and Prospero as examples of systematic review databases. Lewis and Kondo noted that funding is needed not only to support the resource heavy processes of guideline development today, but also to fund the potential future of living systematic reviews and living guidelines.

Educating both doctors and patients on genomic testing and pharmacogenomics may help in shared decision making and implementation, said Agarwal, Caudle, and Temple-Smolkin. Engaging patients and communities throughout the guideline development process, including decision making and implementation, could lead to better adoption of genomic testing, said Arjunan, Jonassaint, and Phillips. Reaching patients with recommended care will be a challenge if guideline development panels do not involve those patients from the start, Jonassaint added, which Pereira said could be done through strong patient advocacy groups. Another way to move guidelines forward would be to create standardized approaches to developing guidelines, Kapoor said—and standardized approaches for implementing them, Caudle added. Implementation outcomes and considerations incorporated into guideline deliberations could help ensure that recommendations are truly feasible and acceptable while also taking into account the costs of implementation, Chambers said. Providing support for interpreting the genomic testing results could aid in implementation, Meric-Bernstam said, and Jonassaint added that EHR and other technologies could be used in decision support and delivery of care (e.g., telehealth).

To address payers’ concerns about coverage, Hooker suggested convening interest holders to explore factors that could drive policy changes related to standard processes that work for all relevant parties. Addressing barriers that would allow payers to be consistent with one another could be beneficial, Haywood added. Moving away from RCTs for the guideline evidence base may be necessary for genomic testing, Meric-Bernstam and Pereira suggested. Zetzsche and Lewis said, keeping high standards of evidence, having high-quality frameworks, and minimizing conflicts of interest can help push guideline development forward. Jonassaint cautioned that moving away from RCTs could lead to practice-based variation. Finally, Rehm said that to move the needle, new models for how genomic testing is delivered could be considered. Collaboration is a potential solution to many of the challenges that are currently being faced, such as inconsistencies across guidelines, lack of resources, cost burden, and getting genomic testing into guidelines for many specialties, which several speakers addressed throughout the day.

REFERENCES

DISCLAIMER This Proceedings of a Workshop—in Brief has been prepared by Kathryn Asalone Shively, Theresa M. Wizemann, and Sarah H. Beachy as a factual summary of what occurred at the meeting. The statements made are those of the rapporteurs or individual workshop participants and do not necessarily represent the views of all workshop participants; the planning committee; or the National Academies of Sciences, Engineering, and Medicine.

*The National Academies of Sciences, Engineering, and Medicine’s planning committees are solely responsible for organizing the workshop, identifying topics, and choosing speakers. The responsibility for the published Proceedings of a Workshop—in Brief rests with the institution. The planning committee comprised Mylynda Massart, University of Pittsburgh; Victoria Pratt, representing Association for Molecular Pathology; Trish Brown, CVS Health; Pranil Chandra, PathGroup; W. Gregory Feero, representing JAMA; Gabriel Lazarin, Myriad Genetics; Funda Meric-Bernstam, MD Anderson; Rebecca Morgan, Evidence Foundation; Wanda Nicholson, George Washington University Milken Institute; Mary Nix, Agency for Healthcare Research and Quality; Pim Suwannarat, Mid-Atlantic Permanente Medical Group.

REVIEWERS To ensure that it meets institutional standards for quality and objectivity, this Proceedings of a Workshop—in Brief was reviewed by Shalini Selvarajah Jackson, ACMG; Kenneth Offit, Memorial Sloan Kettering Cancer Center; and Julie Wiedower, Guardant Health. Leslie Sim, National Academies of Sciences, Engineering, and Medicine served as the review coordinator.

SPONSORS This workshop was partially supported by contracts between the National Academies of Sciences and 23andMe, Inc; American Academy of Nursing; American Association for Cancer Research; American Cancer Society; American College of Medical Genetics and Genomics; American College of Radiology; American Medical Association; American Society of Clinical Oncology; American Society of Human Genetics; Association for Molecular Pathology; Association of American Cancer Institutes; Association of Community Cancer Centers; Biogen; Blue Cross Blue Shield Association; Cancer Center at Illinois; Centers for Disease Control and Prevention (Contract No. 75D30121D11240, Task Order No.75D30124F00042); College of American Pathologists; Flatiron Health; Geisinger Health; Genome Medical, Inc.; Health Resources and Services Administration (Contract No. HHSH250201500001I, Task Order No. 75R60220F34021); Illumina, Inc; The Jackson Laboratory (JAX); Kaiser Foundation Health Plan, Inc.; Merck; Myriad Genetics; National Comprehensive Cancer Network; National Institutes of Health (Contract No. HHSN263201800029I, Task Order No. 75N98023F00019 and 75N98023F00022; Purchase Order No. 75N92024P00357), including All of Us Research Program, National Cancer Institute, National Human Genome Research Institute, National Institute of Mental Health, and National Institute on Aging; National Society of Genetic Counselors; Novartis Oncology; Oncology Nursing Society; Partners In Health; Patient Advocate Foundation; Pfizer Inc.; the University of California, San Francisco; and the University of Vermont Health Network Medical Group.

STAFF Sarah H. Beachy, Senior Program Officer; Kathryn Asalone Shively, Associate Program Officer; Michelle Drewry, Associate Program Officer; and Ashley Pitt, Senior Program Assistant from the Board on Health Sciences Policy, and Jennifer Zhu, Associate Program Officer from the Board on Health Care Services supported this work.

For additional information regarding the workshop, visit https://www.nationalacademies.org/our-work/examining-clinical-guidelines-for-the-adoption-of-genomic-testing-a-workshop.

Suggested citation: National Academies of Sciences, Engineering, and Medicine. 2025. Exploring clinical guidelines for the adoption of genomic testing: Proceedings of a workshop—in brief. Washington, DC: The National Academies Press. https://doi.org/10.17226/28572.

Health and Medicine Division Copyright 2025 by the National Academy of Sciences. All rights reserved.