Precision public health in the era of genomics and big data

Abstract

Precision public health (PPH) considers the interplay between genetics, lifestyle and the environment to improve disease prevention, diagnosis and treatment on a population level—thereby delivering the right interventions to the right populations at the right time. In this Review, we explore the concept of PPH as the next generation of public health. We discuss the historical context of using individual-level data in public health interventions and examine recent advancements in how data from human and pathogen genomics and social, behavioral and environmental research, as well as artificial intelligence, have transformed public health. Real-world examples of PPH are discussed, emphasizing how these approaches are becoming a mainstay in public health, as well as outstanding challenges in their development, implementation and sustainability. Data sciences, ethical, legal and social implications research, capacity building, equity research and implementation science will have a crucial role in realizing the potential for ‘precision’ to enhance traditional public health approaches.

Decades after the completion of the Human Genome Project, our understanding of the interplay between genetics, lifestyle and the environment continues to grow. This understanding has heralded the era of precision medicine, which aims to deliver the right intervention to the right patient at the right time—to improve disease prevention, diagnosis and treatment^1–5. More recently, the public health community has begun to consider how this interplay can be applied at the population level, through PPH. Complementary to precision medicine, PPH refers to the delivery of the right intervention to the right population at the right time, with the goal of improving health for all while considering key determinants of health^6–8. PPH acknowledges the limitations of a one-size-fits-all approach to public health and accounts for heterogeneity—both in and across populations—which holds promise to improve efficiency and effectiveness of public health approaches and interventions⁹.

Initially, the concept of ‘precision’ in a public health context might seem paradoxical, but the incorporation of individual-level data to refine the delivery of public health interventions dates back decades. Take, for example, newborn screening for health conditions, which is one of the ten great public-health achievements in the USA¹⁰. Today, public health surveillance continues to evolve and incorporate more precise data and advanced methods. For example, sequencing-based analyses of pathogens in wastewater are being used to inform infectious disease surveillance and response at the community level¹¹. As our ability to more readily capture genetic, lifestyle and environmental data increases, we can better assess population health and identify populations at higher risk of disease, allocate resources for more targeted interventions and understand the complexity of factors affecting health inequities.

Rapid and ongoing advances in genomics, other ‘omics’ fields, artificial intelligence (AI) and digital technologies have facilitated the collection of big data and led to a new era of PPH^6,12. In this Review, we explore the scientific foundation and practical applications of PPH, highlighting its potential as the future of public health, by outlining how human genetics, pathogen genomics, social and environmental data and AI are transforming the field (Fig. 1). We provide real-world examples from the USA and elsewhere, and explore key challenges and opportunities for advancing the field.

Fig. 1 |. Precision public health.

Human and pathogen genomics, social, behavioral, environmental and other individual-level data inform PPH; data sciences (including AI approaches), ethical, legal and socialimplications (ELSI) research, capacity building, equity research and implementation science offer methods to advance the impact of PPH. Together, these factors will play a critical part in realizing the potential for precision to enhance standard public health approaches.

Precision public health in action

Integrating genomic medicine across the care continuum for population impact

In 1997, the US Centers for Disease Control and Prevention (CDC) began addressing the public health implications of new human-genome discoveries, marking the beginning of public health genomics as a field in the USA and globally^8,13. A main goal was to identify, evaluate and integrate evidence-based genomics applications into public health programs, to prevent and control the leading causes of morbidity and mortality. In part owing to research using large-scale biobanks, such as the UK Biobank and All of Us from the US National Institutes of Health, our understanding of the association between genetic variants and health outcomes is better than ever before. Multiple applications based on human genomics are now used in healthcare, covering the entire lifespan¹⁴. These applications include, among others, expanded carrier testing, non-invasive prenatal testing, expanded newborn screening, genetic testing of children with birth defects and developmental disorders. In addition, genomic applications are emerging for diagnosis, screening and prediction of an increasing number of adult conditions, such as cancer and heart disease.

Most recently, population genetic screening after the newborn period has emerged as an important area of public-health genomics that is being explored^15,16. Individuals with common hereditary cancers and cardiovascular diseases are underidentified using current clinical guidelines. They often receive a diagnosis only after presenting with a substantial clinical or family history of disease—representing a missed opportunity for disease prevention and control. In response, population genetic screening programs are growing¹⁷. The CDC has identified three genomic applications with strong levels of evidence and clinical guidelines¹⁶ that have the potential to improve public health. Referred to as CDC tier 1 applications, these include Lynch syndrome, hereditary breast and ovarian cancer syndrome and familial hypercholesterolemia¹⁸. Indeed, one population health genetic screening program found that 90% of individuals with conditions associated with CDC tier 1 applications would have been missed in current clinical practice¹⁹. Most population genetic screening programs in the USA include screening for CDC tier 1 applications at minimum, but they vary in terms of the total number of genes screened (from only 11 to hundreds of genes, including pharmacogenes associated with drug metabolism and response), funding models (state, research, healthcare system), industry partnerships and size¹⁷.

Polygenic risk scores can also guide disease prevention and management, and their clinical use is growing²⁰. Polygenic risk scores quantify genetic predisposition to a disease or trait (on the basis of many genetic variants of small effect) and estimate disease risk or other clinical outcomes²¹. In some cases, integrated risk models incorporate both polygenic risk scores and non-genetic information to offer a more-precise risk estimation. In the clinical setting, these scores can be used to stratify groups according to prognosis or risk of disease, particularly in coronary heart disease and breast cancer. For instance, an integrated risk model named BOADICEA²² has been developed for breast cancer risk stratification, enabling more-precise screening and informing prevention strategies. A risk-based approach to breast cancer screening, compared with no screening or age-based screening, could improve the cost-effectiveness and benefit-to-harm ratio of breast cancer screening in the UK²³; currently, a risk-based screening approach is being tested in the WISDOM trial²⁴ in the USA. The Polygenic Score Catalog offers a curated, open resource of nearly 5,000 polygenic scores that meet accepted reporting standards (including those for key factors, such as ancestry), so that the polygenic scores can be disseminated, reproduced and further validated²⁵.

Certain direct-to-consumer testing companies also include polygenic risk scores for complex, multifactorial traits ranging from earwax type to likelihood of developing colorectal cancer. The public’s experience with genetic testing more commonly falls in the direct-to-consumer context than the clinical context²⁶. Although the delivery of direct-to-consumer testing might be outside the scope of PPH, the public’s experiences with these tests could shape their attitudes and knowledge about genomics more broadly. A review of healthcare professionals’ views about direct-to-consumer testing demonstrates that, although providers’ experiences with these tests have increased, gaps in knowledge and ethical and legal concerns remain, ranging from concerns about clinical utility to misinformation to low patient understanding about results²⁷.

In addition to disease prevention and diagnosis, public health genomics has changed how we approach population disease management and treatment. Learning health systems support the collection and analysis of data to continuously improve clinical practices and population health; some also integrate genomic data into electronic health records^28,29. In California³⁰ and New York³¹, health systems have merged genetic and electronic health record data to better understand associations between key risk factors (both genetic and environmental) and health outcomes in patient populations that might otherwise be under-represented or understudied in genomic research. This concept of fine-scale population health monitoring could offer a means to expand our PPH knowledge and redress inequities in the field. Efforts funded through the National Human Genome Research Institute are underway to develop a network of genomics-enabled learning health systems to better understand common challenges and strategies for the implementation and sustainability of these systems across the USA³².

Despite the growth of public health genomics, overall awareness of the field among the public remains low—a national survey has shown that just over half of the US population is aware of genetic testing for health risks³³. Awareness has been associated with sociodemographic characteristics, including factors such as race, ethnicity, education level and income³⁴. These gaps in genetic knowledge also affect providers, who have consistently reported feeling underprepared to incorporate genetic and genomic data into patient care³⁵. Published studies have consistently shown lower implementation rates for selected tier 1 genomic applications among minority racial and ethnic groups, rural communities, uninsured or underinsured people and those with lower education levels and income—which could in part be related to gaps and biases in both provider and public knowledge³⁶. Recent work evaluating the needs of primary care providers at federally qualified health centers found that, among this group, the perceived value of applying genomics in disease prevention and care was high, and respondents felt they needed more training on how to integrate genetics into their practice. Fortunately, there is a strong evidence base for interventions that can be used to build genomic capacity among the healthcare workforce and to increase public awareness and knowledge³⁷.

These gaps in basic awareness and knowledge, along with other well-documented multilevel barriers to the use of genetic information in clinical and public health settings, have affected the adoption, implementation and sustainability of genomics applications^38,39. These barriers, including structural racism, have had impacts on the delivery of healthcare generally and could have a particularly damaging impact on equity in the context of genetic services⁴⁰; for example, inequities in genetic screening have been reported among minoritized groups in the USA⁴¹. Inequities also occur upstream of implementation, in the development of public health genomics tools.

Currently, most of the data underpinning public health genomics come predominantly from European populations, which do not fully represent the racial and ethnic diversity of the USA⁴²—a problem that is repeated and often magnified across the globe⁴³. This undermines the potential impact of public health innovations. The very same biobanks and databases that have spurred genomic innovation have also contributed to their Eurocentric biases, calling into question the generalizability of innovations such as polygenic risk scores⁴⁴ to unstudied and understudied populations⁴². Biases in methodologies for genomic analysis further exacerbate inequities, particularly for admixed populations, which comprise more than one-third of the US population (and that number is growing). Historically, owing to a lack of statistical frameworks for analyzing genomic data from admixed populations, these groups have been excluded from genomic research^45,46. New and long overdue statistical approaches, such as Tractor, have been developed to facilitate genome-wide association studies in admixed populations⁴⁵, but more are needed. Addressing health equity in the development and implementation of genomic innovations for PPH is a public health imperative. An increasing focus on equity in implementation research has led to multilevel frameworks that consider equity at the earliest stages of translational research through implementation and evaluation^47–52.

Pathogen genomics: from outbreak investigation to global surveillance

Public health management of infectious disease threats is underpinned by the work of reference laboratories that provide routine surveillance services to support the design, deployment and assessment of interventions and local guidance, as well as investigative services to understand and respond to emerging threats and outbreaks. Molecular analysis of pathogens has been central to public health microbiology for decades, and the digital nature of molecular data facilitates timely sharing of information between laboratories—which is particularly important for identifying and responding to cross-border threats. For example, PulseNet, a network for surveillance of foodborne bacterial diseases, was established by the CDC in 1996 (ref. 53), and the World Health Organization (WHO) launched FluNet in 1997 (ref. 54). Since 2012, whole-genome pathogen sequencing has become integrated into standard public-health protocols in many countries, beginning with the GenomeTrackr⁵⁵ network (which includes the US Food and Drug Administration (FDA), the CDC, the US Department of Agriculture and the UK Health Security Agency⁵⁶) for identifying foodborne bacteria.

A major focus of pathogen genomics in public health is supporting outbreak investigation, because whole-genome sequencing provides maximal power to resolve relationships between clinical isolates and potential sources of infection and thus to support or discount potential transmission chains⁵⁷. Another key application is characterization and investigation of emerging pathogens or variants—for example, during the Ebola and Zika epidemics⁵⁸ and the COVID-19 pandemic⁵⁹—during which sequencing was used to characterize new threats, distinguish regional spread from local re-emergence and understand transmission dynamics at various levels to inform public-health guidance and restrictions. The newest application of genomics to PPH is wastewater-based pathogen surveillance, in which wastewater samples are sequenced (with or without target amplification) to detect the presence of pathogens and variants at the community level. Although this is not a new concept and has been used for some time to support polio eradication, it has resurged globally since the COVID-19 pandemic⁶⁰.

The COVID-19 pandemic provides the most salient and recognizable example of pathogen genomics informing PPH, when the names of specific genomic lineages of SARS-CoV-2—such as Alpha, Delta and Omicron—became a part of the general public discourse⁶¹. SARS-CoV-2 sequences are used to stratify the pathogen population into variants⁶²; this allows public health agencies to monitor whether important characteristics, such as transmissibility, virulence or immune escape, are changing, and to respond accordingly. An important part of this response is declaring a ‘variant of concern’, which facilitates detection of the emerging threat, communication and coordination of appropriate responses. For example, the Delta variant was declared a variant of concern in early 2021, following evidence suggesting that it was associated with enhanced transmissibility, higher risk of hospital admission and partial immune escape⁶³; the variant quickly spread and became dominant in at least 130 countries⁶⁴. Governments around the world responded to the arrival of Delta. For example, the UK government delayed its roadmap out of lockdown⁶⁵; the CDC updated its guidance to recommend mask wearing indoors in public settings, regardless of vaccination status⁶⁶; and in Australia, international and state borders were closed and lockdowns were imposed that became the longest in the world⁶⁷. Many governments cited Delta in explaining their pandemic-policy decisions. In addition to its impact on pathogen genomics, the COVID-19 pandemic fundamentally changed the study of host genetic susceptibility to emerging infectious diseases. Global collaboration increased at a rapid pace to collect information about people infected with COVID-19 and identify those at higher risk of severe outcomes^68,69. Advancement in host genetics studies continues to be an important area for PPH.

Since the COVID-19 pandemic, there has been an explosion in global capacity for pathogen sequencing in public-health laboratories; many countries have acquired sequencers for the first time, and those that already had sequencing capacity have expanded it, both in terms of infrastructure and expertise. In October 2020, the Africa CDC launched the Africa Pathogen Genomics Initiative, which in its first three years saw the number of African countries with genomics capacity in their public-health laboratories grow from 7 to more than 30; the goal is to extend this to all African countries⁷⁰. The global challenge now is to build on the current momentum to develop and sustain capacity for public health genomics; this will involve not just improving laboratory infrastructure, but also providing workforce training (covering competencies such as laboratory techniques and data analysis, interpretation and synthesis) and establishing equitable data sharing and governance⁷¹. Toward these broad goals, the WHO has recently released guiding principles for pathogen-genome data sharing⁷² and guidelines on developing national genomic surveillance for pathogens with pandemic and epidemic potential⁷³, and also launched the International Pathogen Surveillance Network. On the practical side, the Public Health Alliance for Genomic Epidemiology (launched in late 2019) aims to provide practical support for rapid, global, genomic-driven public health responses to disease outbreaks. This is achieved through collaborative working groups focused on improving openness and interoperability in public health bioinformatics. The alliance is open to members from any country, and recent outputs include developing data standards to promote and maximize the utility of SARS-CoV-2 data sharing⁷⁴, and a framework for promoting ethical benefit sharing in health research⁷⁵.

Expanding precision public health through the integration of social and environmental factors

PPH has expanded beyond genomics applications to bring together the complex interplay between genetic and non-genetic factors—including behaviors and environmental exposures—that affect complex diseases and health⁷⁶. For example, the interactions of environmental and social determinants of health with genomics (that is, gene–environment interactions), DNA methylation and the gut microbiome have all been shown to have a role in cardiovascular disease risk⁷⁶. Lifestyle factors, social stressors and air pollution can affect DNA methylation, individuals’ responses to environmental factors and disease risk and progression⁷⁶. Indeed, studies have shown that people with lower socioeconomic status are at an increased risk of cardiovascular disease, potentially owing to increased exposure to environmental stressors related to inflammation⁷⁶.

The ‘exposome’ (that is, the entirety of environmental exposures a person encounters) provides key insights into biomarkers of disease across the life course and potential targets to improve health⁷⁷. Indeed, geospatial analysis of where people live, key neighborhood and environmental characteristics—such as walkability, population density, climate-related variables, pollutants and vegetation index—and their associations with key health outcomes⁷⁸ can further inform place-based public health approaches. For example, exposure to higher levels of ‘greenness’ (that is, nature) has been associated with a lower incidence of high depressive symptoms, particularly in densely populated areas⁷⁹. In another study, higher walkability was associated with increased likelihood of depressive symptoms, but only in neighborhoods with low socioeconomic status—potentially owing to increased noise, air pollutants and social stressors⁸⁰. As methods to measure the processes underlying aging over the life course advance, future work might focus on understanding how environmental factors impact aging. Given the vast disparities in exposure to environmental stressors, the field of environmental PPH has great potential for understanding and intervening on public-health issues related to disease risk, disease progression and aging⁸¹.

The growth of interconnectivity in digital information has advanced our ability to not only measure environmental exposures, but also monitor social and behavioral data across multiple levels. Digital platforms have greatly expanded our ability to capture and analyze data about highly variable individual-level constructs (for example, self-efficacy) and more stable factors (for example, environmental exposure)⁸². Using these data, it is possible to develop and match interventions that are aligned with highly complex health challenges. The ability to ‘continuously tune’ interventions is a recent development, whereby interventions are adjusted and tweaked on the basis of a person’s own current data, using machine-learning techniques and control systems engineering⁸². This approach is distinct from adaptive interventions (based on prior participant data) or targeted interventions (based on a single variable), and can more accurately address the complexities of behavior change⁸².

Additional progress will require commitment to standardizing and enhancing data collection, protection, sharing and analysis across geographic jurisdictions. In particular, the standardization of relevant social and environmental metrics will be foundational to establishing guidelines for PPH research, particularly research involving machine learning techniques at scale. Addressing complex conditions that arise from the interplay of genetic, lifestyle and environmental factors will require complex interventions and thoughtful implementation⁸³. Indeed, the field of implementation science offers valuable methods and tools to improve the integration of evidence-based PPH interventions into clinical and public health settings. Calls for integration of implementation science and PPH are longstanding, and although work in this area has grown over recent years, it remains an area of need⁸⁴.

The role of artificial intelligence in precision public health

Rapid growth in the amount of health-related data, coupled with advances in storage and computational capacity, has led to new approaches to the analysis, interpretation and use of such information at the population level⁶. AI and machine learning have been increasingly applied in PPH to interpret big data through learning and problem solving to achieve public health goals⁸⁵, and to predict which interventions are likely to succeed for an individual or population.

AI is increasingly offering opportunities to identify emerging health threats and develop a more refined assessment of population disease and risk-factor distributions, and it can provide updated information in near-real-time. AI has been used to monitor community-health-related events, both geographically and temporally⁸⁵, by analyzing health data and web-accessible news and social media. For example, SENTINEL is a syndromic surveillance application that uses natural-language processing and neural-network algorithms to analyze millions of tweets per day, to predict disease trends and identify potential outbreaks⁸⁶. AI can also be deployed to summarize surveillance data from multiple unstructured sources: for example, analysis of free-text information in death certificates has been used to detect deaths potentially resulting from drug overdoses, often months before traditional data release⁸⁷.

The COVID-19 pandemic led to implementation of data modernization across public-health systems, the increasing use of AI and predictive analytic approaches and advances in healthcare delivery, such as telehealth and at-home testing. These advances raise unique opportunities for scalable innovations in population health management by using AI algorithms to identify and prioritize patient cohorts that could benefit from certain evidence-based interventions and then using digital health tools to deliver those interventions. For example, the BRIDGE trial used algorithms to screen data in electronic health records and identify patients that met criteria for genetic testing of hereditary syndromes; automated chatbots were then used to educate and offer genetic testing to eligible patients⁸⁸. During the roll out of the COVID-19 vaccine, the SCALE-UP trial (n = 107,000 patients) used a similar approach, in which algorithms used CDC criteria to prioritize patients receiving care at low-resource Utah Community Health Centers for vaccination; patients were contacted through text messages that offered them access to vaccination and connected them to a vaccination site⁸⁹. Similar approaches could be adapted to other contexts as a way to deliver evidence-based health interventions rapidly, equitably and at the population scale.

Nevertheless, for AI algorithms to fulfill their promise to improve health and healthcare, it is important to evaluate their methodological rigor and assess the risk of bias associated with these tools. A systematic review reported that 87% of articles on machine-learning models were at high risk of bias, most often owing to bias in statistical methods⁹⁰. The use of digital tools and AI for public health surveillance is subject to a multitude of potential biases⁸⁵, including biases related to differential implementation and measurement bias, which could widen health inequities in medically underserved populations⁹¹. Currently, most AI development is done using data from highly resourced academic medical centers, which have disproportionately lower representation of individuals with low socioeconomic status, historically marginalized groups and those living in rural or frontier areas. Lack of representative data can reflect systemic inequities in health and healthcare delivery that can unintentionally lead to unjust bias and further exacerbation of inequities^92,93. To address this issue, AI developers need to rigorously identify and mitigate biases from the earliest stages of model development (for example, by including data from low-resource health systems and actively oversampling data from under-represented groups); they should also conduct and transparently report evaluation of model performance across different population groups. Other concerns include the ‘black box’ nature of AI algorithms; issues around model development, interpretation and generalizability; and infrastructure challenges (such as data sharing, analytic capacity and an untrained public-health workforce); as well as a host of ethical, legal and social issues that highlight the urgent need for rigorous standards and validation of AI applications in PPH.

Looking ahead

As we consider areas of PPH advancement, we find that PPH is fundamentally rooted in the basic concepts of public health⁹⁴. The three essential functions of public health are assessment, policy development and assurance⁹⁵. Below (and summarized in Box 1), we discuss how these core functions can guide our thinking about how to advance PPH into the future⁹⁶.

BOX 1. Aligning core functions of public health with precision public health.

Core public-health functions and services are briefly explained, with case examples in the PPH context^6,19,96,116.

Assessment: using human and pathogen genomics, big data, modeling and analytics to improve population-health assessment across various dimensions, including people, place and time

• 1.Public-health surveillance: adding hereditary cancers and genomic markers to state-based cancer registries^107,108
• 2.Public-health investigations: integrating host genomics into public-health investigations of selected diseases⁶⁸

Policy development: using genomics, big data, modeling and analytics to develop the right policy, program and educational interventions for the right populations at the right time

• 3.Communication and education: educating the public about the importance of genomics and family health history in disease prevention¹⁰⁹
• 4.Community partnerships: engaging and supporting disease-specific support organizations to reduce population disease burden (for example, the global summit on familial hypercholesterolemia)¹¹⁰
• 5.Targeted policies: for example, The Evaluation of Genomic Applications in Practice and Prevention (EGAPP) initiative¹¹¹
• 6.Legal and regulatory actions: for example, Affordable Care Act coverage of the US Preventive Services Task Force recommendations on BRCA testing for hereditary breast and ovarian cancer¹¹²

Assurance: ethically and effectively integrating genomics and big data tools into individual and community health services, training the public-health workforce, developing necessary infrastructure and measuring impact through implementation science, evaluation and quality improvement

• 7.Health systems for equitable access: integrating cascade genetic testing into healthcare systems¹¹³
• 8.Development of public-health workforce: establishing genomic competencies in the public-health workforce¹¹⁴
• 9.Applied and implementation research: contributions of implementation science to the population health impact of genomic medicine¹¹⁵
• 10.Infrastructure development: for example, CDC five-year initiative in building capacity in public health genomics, 2021–2026 (ref. 116)

Assessment

Assessment, or the regular and systematic collection, analysis and dissemination of information about the health of a community, can be supported through precise methods of surveillance and investigation that leverage big data (including, but not limited to genomics), modeling and analytics. The field of data science will continue to play a major part in bringing to bear the promise of PPH⁹⁷. As PPH continues to evolve, these methods will become increasingly efficient and precise. Some posit that PPH can offer a means to reduce costs and improve quality by leveraging data from learning health systems, along with genetic, non-genetic and other digital health data, to assess the risk of, identify and manage chronic disease⁹⁷.

For such a vision to come to fruition, we must also carefully select and operationalize key metrics for assessment of PPH data on the basis of consensus across multiple perspectives, given the multidisciplinary nature of the field. Until these population-level metrics have been chosen, harmonized and operationalized, assessing the impact of PPH on population health outcomes more broadly remains difficult⁹⁸. Standardized data fields from across different sourcesallow for data to be pooled and shared, providing the sample sizes and statistical power needed to assess and monitor population health status as well as identify associations between genetic, environmental and lifestyle factors and key health outcomes. A crucial step in this direction is the expansion of state and national health registries to include genetic information. Population disease surveillance systems are benefitting increasingly from enhanced diagnostic classification of diseases that includes molecular markers of etiology and treatment response. In the USA, the eMERGE consortium provides a framework for accomplishing such work by combining DNA biorepositories and electronic medical record systems for large-scale research across multiple sites. Yet, published evidence demonstrating how PPH interventions have affected population health is lacking^84,99; therefore, generating an evidence base must be a priority in PPH research. Without demonstration of benefits to population health through assessment of PPH interventions, the implementation, scale-up and sustainability of these interventions remain uncertain.

Policy development

This function involves the development of comprehensive public health policies, partnerships and educational programs on the basis of scientific evidence to improve public health, as well as advocating for those policies and identifying resources to implement them. Here, research on ethical, legal and social implications (ELSI) will be imperative to mitigate unintended consequences of PPH interventions, particularly given that the rate of PPH discovery often outpaces the implementation of corresponding policy. This has been a challenge in AI research–for example, upon the release of the chatbot ChatGPT, a scramble ensued to develop policies addressing its unintended consequences across universities, funders and employers¹⁰⁰.

Policies that align with emerging technologies will be needed at all levels from local to national, to protect individuals’ privacy and, in the case of genetics, the privacy of their families. A systematic review¹⁰¹ highlighted public concerns about ELSI issues related to the use of genetic information, particularly use of this data by employers, insurers and the government. Indeed, gaps in protection from life insurance discrimination are a highly cited concern among providers, patients and other partners in clinical genomics¹⁰². Advances in prenatal testing have also spurred controversy around where to set ethical and social boundaries. For example, the use of polygenic scores for embryo selection has raised concerns about the ethical implications of selecting for particular physical and cognitive traits in future generations. These tests include screening for risk of certain cancers, diabetes and intellectual disability—leading to concerns about devaluing certain traits, changing population demographics and worsening inequities^103,104.

Ongoing societal analysis, bidirectional learning, collaboration with key community partners and timely responses to the changing ELSI landscapes will be crucial to ensure that new scientific knowledge can benefit all segments of the population and minimize unintended consequences.

Assurance

Public health assurance includes the role of public health agencies in ensuring that necessary services are provided to communities. Key areas of opportunity relate to assuring equitable implementation through engagement with health equity research and implementation science, as well as building workforce capacity in PPH. Training the next generation of PPH researchers and practitioners will be essential to realizing the benefits of new technologies. Efforts by professional associations, public health departments and national organizations (such as the National Human Genome Research Institute in the USA) aim to bridge these gaps through the development of educational resources in genetics and genomics applications for multiple partners. Yet, few PPH programs and curricula currently exist, and those that do primarily center on public health genomics—just one component of PPH. To address this gap, national and global efforts to offer training opportunities to PPH researchers have emerged. In 2021, the Precision Public Health Network was formed to convene early-career PPH researchers and practitioners annually to provide ongoing opportunities for research collaboration, professional networking and training around the world¹⁰⁵. In September 2023, the CDC Office of Genomics and Precision Public Health partnered with the Oak Ridge Institute for Science and Education to discuss current issues in genomics and precision public health. This year, the Training Institute for Dissemination and Implementation Research in Genomics and Precision Public Health, resulting from a collaboration across funders and organizations, will provide an opportunity for international early-stage investigators in PPH to gain additional training in the field of implementation science.

Another core function of assurance centers on equitable access and integration of public health. Concerns have been raised about the potential for PPH to exacerbate health disparities⁴⁷. Reliance on big data to inform research and intervention development underscore the importance of ensuring unbiased data inputs, adequate reporting and appropriate collection of data from minority or under-represented groups. The Precision Public Health Network recently described ways in which implementation science approaches, paired with an antiracism lens as prescribed by Shelton and colleagues¹⁰⁶, can serve as a framework for planning, implementing, assessing and sustaining equitable PPH⁴⁷. Developing infrastructure for implementation science and health equity will be critical to assuring the impact of PPH; in particular, consideration of inequities related to socioeconomic status, geography, sexual and gender identity and disability should be prioritized.

Conclusion

Moving forward, data sciences, ELSI research, capacity building, equity research and implementation science will all play a key part in successful expansion of PPH, as will transdisciplinary training and collaboration across the fields of human genetics, infectious diseases, social and environmental sciences, and AI. Through collaboration, the next generation of public health holds great promise for reducing inequities and providing the right interventions to the right populations at the right time.