A New Agenda for Implementing Population Genomic Screening

Population genomic screening has been associated with multiple positive clinical and system-level outcomes [8]. It is more effective than indication-based genetic testing at identifying people with actionable, collectively common [1], and penetrant disease risk [2, 9] and has been shown to facilitate earlier diagnoses and care [9‒12]. Additionally, possible negative psychological responses to receiving genomic screening results appear to be transient and manageable [13]. Furthermore, cost-effectiveness modeling of population genomic screening is favorable, particularly when programs screen for multiple conditions [14, 15].

Available evidence also points to key lessons regarding clinical implementation of genomic screening programs. First, systems can manage the scale of reviewing large datasets for actionable risk by excluding variants of uncertain significance and leveraging expert curations to identify pathogenic variants [1]. Second, sub-optimal uptake of recommended risk management [9, 11, 16, 17] underscores the need to incorporate strategies to facilitate adherence to management recommendations as part of long-term population health management. Promising strategies include fitting the program into existing clinical workflows via clinical decision support [18, 19] and care coordination and collaborating closely with co-managing clinicians [10, 20], including in primary care [8]. Third, achieving broad population health impact will require robust uptake of family testing when appropriate, with early evidence highlighting the need to develop interventions that improve family communication and testing for the familial variant [21].

Still, there is much left to learn regarding the clinical utility of population genomic screening (see Table 1). For example, penetrance of disease for certain genetic conditions has been lower than predicted when individuals are identified via population genomic screening [17, 20, 22]. This underscores the importance of research to identify the basis for reduced penetrance in these conditions so that genomic screening programs can minimize over-diagnosis and unnecessary healthcare utilization. Additionally, limited representation of historically underserved groups in some genomic screening programs [9] highlights the importance of designing for equitable program implementation [3], both in ascertainment of at-risk individuals and in longitudinal, accessible management of their risk. Furthermore, data on program sustainability within routine healthcare are scarce, as existing programs in health systems, industry partnerships, and national initiatives typically rely on internal or research support rather than offering genomic screening as a clinically reimbursable test. Additionally, evaluation of integrated risk assessment approaches that incorporate monogenic disease risk, clinical risk factors and, when relevant, polygenic risk, is needed. Finally, new genomic screening programs may need guidance on how to adapt and continuously improve their programs within their local context as new evidence and tests become available [23].

We cannot answer these unanswered questions in the USA – or the overarching question of whether population genomic screening can equitably improve long-term health outcomes – by continuing to generate evidence in silos within single systems or research programs. Because of the constantly evolving nature of genomics, ongoing learning through synchronizing research and practice – research while implementing – is critical [5]. To understand and equitably maximize the impact of genomic screening at the system, clinician, and individual levels, we must concurrently implement genomic screening programs while evaluating them using methodologies designed for multi-level and real-world data collection and evaluation. Without the rigorous study of evidence while implementing within systems and clinical contexts, we risk moving too slowly, leading to shortcomings in efforts to achieve equitable access to and use of genomic information and failing to generate lessons necessary to guide adaptation of genomic screening programs to diverse settings and populations.

Implementing population genomic screening equitably at scale will require deliberate and ongoing engagement with those who provide and receive genomic screening. Routine implementation of genomic screening across diverse populations will likely involve integration into primary care [24], where patients typically receive longitudinal care and shared decision-making about screening occurs. This integration could help address concerns regarding access to genomic screening that arise from limitations in capacity of the genetics workforce [25] and lack of racial and ethnic diversity in this workforce [26]. Careful collaboration among clinicians in primary care, genetics, and other specialties will be needed to enable this integration [8, 24, 27]. Engagement with patients and advocates in communities where genomic screening may be offered will also be necessary to enhance the acceptability of screening programs, inform strategies for addressing community members’ concerns regarding privacy and insurability, and encourage representative participation in population genomic screening [27‒29].

With these considerations in mind, we propose the following agenda to more rapidly determine whether population genomic screening is effective at equitably impacting long-term health outcomes and, if so, to determine how to implement screening in different settings.

• Convene a multi-institution consortium of population genomic screening programs that are collectively diverse in race, ethnicity, geography, rural/urban setting, and venue of delivery, and embrace a common data model for longitudinal collection of effectiveness data and implementation outcomes.

• Work closely with payers and regulatory bodies to generate data that will inform their decisions regarding coverage and reimbursement of population genomic screening tests, with consideration of models such as coverage with evidence development that cover a medical procedure while effectiveness data on it are gathered [30].

• Explore service delivery models that mainstream population genomic screening into primary care through implementation, evaluation, and iterative adaptation.

• Create a framework for defining and updating which genetic findings are clinically actionable – and have broadly accessible risk management procedures – in the population screening context.

• Design and evaluate multi-level interventions that support individuals and families with actionable risks in adhering to treatment and risk management guidelines and facilitating family testing.

• Determine what ongoing clinical support is needed to effectively integrate genomic screening results into clinical care, evaluate program effectiveness, and continually improve processes to maximize health outcomes.

• Develop, validate, and share integrated risk models that incorporate monogenic disease risk, polygenic risk, non-genetic risk factors, and social determinants of health.

• Use implementation science tools and pragmatic study designs [5] to gather evidence while implementing in systems, identifying the core program elements that apply across diverse settings and influence patient- and population-level impacts.

• Support ongoing and meaningful multidirectional engagement among patient, family, community, and clinical partners throughout every step of the proposed agenda.

Nearly a decade into the proliferation of population genomic screening programs [31], it is clear that such programs provide benefit in certain contexts. Determining whether these programs provide net positive outcomes across diverse populations, however, will require a significant shift in approach and scale. This shift can be modeled on a precision health research agenda that capitalizes on pragmatic and implementation science methods, learning health system principles and stakeholder engagement [5]. Acknowledging that research and clinical implementation must coexist [5] will help us realize the promise of population genomic screening.

We thank Christa Martin and Huntington Willard for their thoughtful feedback as this manuscript was developed.

Adam H. Buchanan holds equity stake in MeTree and You, Inc. He has received grant funding from the NIH, Exact Sciences, and Freenome Holdings, Inc. Amy C. Sturm is an employee of and stockholder in 23andMe and a consultant for Nest Genomics. She has received grant funding from the NIH. Alanna Kulchak Rahm has received grant funding from the NIH and is now an employee of the NIH. The views presented in this perspective reflect her views and not those of the federal government.

This manuscript was not supported by any sponsor or funder.

A.H.B., A.K.R,. and A.C.S. jointly conceived this manuscript. A.H.B. wrote the initial draft and finalized the manuscript. A.K.R. and A.C.S. reviewed and edited drafts and approved the final version of the manuscript.