5

Nutrition and Health

The fourth session of the symposium featured three presentations and a discussion on the role of personalized nutrition, food composition, and equity considerations in nutrition and health. The session was moderated by Christina Khoo, Ocean Spray Cranberries, Inc.

THE PROMISE OF PERSONALIZED NUTRITION

Josh Anthony, Nlumn, discussed the role of personalized nutrition in creating an affordable, accessible, and acceptable food system that helps every individual make better choices and live a healthier life. Nlumn is a consulting company that assists businesses in participating more fully in the personalized nutrition and precision health marketplace. Personalized nutrition uses individual, specific information founded in evidence-based science to promote dietary behavior change that may result in measurable health benefits (Adams et al., 2020; Ordovas et al., 2018). Anthony described a cycle of four components—benefits, measurement, information, and behavior change—that create a workable, minimal, viable ecosystem. Specific benefits are related to the problem being addressed. Measurement involves tracking objective and credible metrics of health or function benefit, and Anthony explained that these measurements yield information that should empower users to improve their health and lifestyle, thereby affecting behavior change. In turn, he said, behavior change offers benefit, and the cycle continues.

Anthony outlined system-wide implications of health issues in the United States and technological advances that create a foundation for

personalized health approaches. He reported that the nation spends more than $4 trillion on annual health care costs, 90 percent of which are for chronic illnesses and mental health (Buttorff et al., 2023). Poor health has dramatic effects on worker productivity, resulting in more than $550 billion in annual costs to U.S. employers (Japsen, 2020), Anthony said, noting that this figure surpasses the annual revenue of Microsoft, Amazon, or Apple. Poor health also increases mortality rates, he said, and preventive care could save approximately 100,000 U.S. lives each year. Anthony asserted that personalization can drive preventive care and that technological advancements are making more personalized approaches possible. Over the past 30 years, he related, several technology innovations have enabled the rise of personalized approaches. Sequencing of the human genome from 1990 to 2003 led to a bold vision for “P4 medicine” (i.e., medicine that is predictive, preventive, personalized, and participatory) (Japsen, 2020; NHGRI, 2023). In 1992, IBM announced the first smartphone, and in 2007 Apple revolutionized the industry with the launch of the iPhone, enabling people to carry personal computers in their pockets. Smartphones are an unmatched source of social and behavioral data, Anthony remarked, and the challenge of translating these data into behavior change persists. In 1994, the rise of metabolomics began with experiments that combined liquid chromatography and mass spectroscopy. The development and commercialization of microarrays have furthered the field, he said, and the metabolome, proteome, and lipidome can now be examined in combination with genomics. Computational biology yields insights into the systems of biology, standing in sharp contrast to the reductionist approach to biology that was common in decades prior. The Massachusetts Institute of Technology developed smart clothes, augmented reality, and wearable devices in the late 1990s, and Seiko launched the first “smart watch” during the same period. Currently, Anthony continued, approximately 60 million people track health statistics with a smart watch or similar device. In 2019, he said, artificial intelligence (AI) advanced with the development of generative, pretrained transformer models capable of generating coherent and contextually relevant sentences and could search and analyze disparate datasets. Anthony stated that this new level of AI capability shifts approaches to biological data algorithms and holds promise in integrating behavioral data to propel efforts to change health outcomes.

Anthony acknowledged that, despite technical advancements, the ability to demonstrate efficacy in personalized nutrition is in the early stages. In this context, he said, efficacy refers to the ability to deliver health and functional benefits that can be maintained over time and are better than population-based guidelines (NHGRI, 2023). Anthony remarked that demonstrating efficacy will involve overcoming numerous challenges—for instance, assessing the ability of personalized nutrition to deliver health

and functional benefits requires (1) simple and reliable measures of nutrient status to establish baseline and progress; (2) a clear definition of the continuum of disease to health; and (3) identification of biomarker combinations that predict disease. Demonstrating that results are better than population-based guidelines, which are measured by the absence of disease, will require more representative datasets and experimental approaches, such as single-subject studies, he said. Showing that outcomes can be maintained over time involves longitudinal studies, which are often difficult and expensive, stated Anthony. Additionally, he explained, an individual’s health needs change over time, and achievement of one goal may coincide with identification of a different goal. Moreover, science must be translated to affordable, acceptable, and accessible solutions to achieve broad effects, he contended.

Barriers to Personalized Nutrition Plan Uptake

Anthony noted that understanding and addressing individual lifestyle choices and behaviors are equally as important as biology. An Nlumn consumer study found that 57 percent of consumers are interested in personalized nutrition and health, and some consumers are already participating in a variety of ways, he reported. Out of 3,000 individuals interested in personalized and precision nutrition, 84 percent experimented with a different eating plan or style and 15 percent participated in personalized nutrition in the last year. More than 40 percent of participants regularly use digital apps to track health, and many are willing to share their data with companies that manage those apps. Consumers voiced that plentiful data are available to them, but making sense of data is challenging. Anthony noted that the COVID-19 pandemic affected many people’s priorities; the most frequently reported consumer-desired benefits in the study included quality sleep, increased energy, general wellness, and mental and emotional health—participants ranked weight management and cardiovascular health as less important than these benefits. Anthony stated that research, anthropometrics, and biological markers are not yet robust enough for current consumer priorities; thus, he said, further efforts are needed to identify new markers and measure prioritized benefits.

Anthony highlighted that the value equation for personalized nutrition plans does not yet meet consumer needs, as the cost (measured in money and effort) is greater than the benefits. Indeed, the number one reason why people across all income groups cease using personalized nutrition plans is cost, he stated, adding that cost is the largest barrier to entry into using these plans, and many personalized nutrition plan users cease participation within 6 months. After cost, Anthony said, consumers most often cite lack of support as their reason for no longer participating in the plan. Although

two-thirds of consumers seek a high level of support, only about one-third indicate they receive such support, he noted. Health care professionals are the most trusted source of desired support, yet most people receive support through online search engines, which often leads to feelings of frustration and being overwhelmed. Sectors can work together to deliver more effective personalized nutrition models, said Anthony. Academia can increase cross-disciplinary research, build diversity in datasets, and develop meaningful markers of health and disease, he asserted. Anthony stated that industry can make consumer behavior data accessible, support translation of data to behavior change and associated benefits, and deliver innovative solutions that support end benefits. And, he said, government can increase funding for health research, expand claim opportunities, and provide a structure for incentives for the democratization of personalized nutrition.

A Future Personalized Nutrition Scenario

Anthony described how personalized nutrition could soon look for consumers. In this example, a fictitious person, “Sara,” recently lost 10 pounds and is pleased with her progress but struggles to reach her goal of losing 15 pounds before an upcoming reunion. Continuous sensors detect changes in satiety-related hormones, indicating a needed adjustment in energy and nutrients to support her goals and maintain body composition. Simultaneously, a hydration patch detects decreased hydration status during afternoons, which affects focus and contributes to hunger. In response to these data, Sara receives recommendations for afternoon beverages that meet her personal preferences as well as her biological needs. In addition, these data are communicated to her employer’s smart beverage machine. Because the employer understands the value of productivity associated with a healthy workforce, the employer covers the cost of the beverages. Later, Sara invites her sister over for dinner. In preparation, her smart refrigerator recommends tweaks to favorite family dishes that align with her weight loss goals. The refrigerator produces a shopping list of ingredients. Technology enables Sara to partner with people seeking the same benefits in order to capitalize on volume discounts. Sara’s weight loss reduces her risk calculation, and her health insurance company extends her a discount in response. She successfully reaches her goal weight and feels motivated to work toward a new goal of building the muscle needed to take up rock climbing. Anthony described this scenario as illustrating the cycle of benefits, measurement, information, and behavior change.

Anthony summarized that more tailored approaches to nutrition may be helpful in addressing chronic illness, reducing health care costs, and improving the health span. However, he said, demonstrating the efficacy of personalized nutrition plans is currently in the early stages. Consumers are

seeking more personalized nutrition solutions, he continued, but a value gap between the price paid and the results received leads many individuals to stop participating in these programs within 6 months. Collaboration between sectors is needed to address the lack of alignment between science, consumer expectations, and outcomes to deliver evidence-based models that are accessible and affordable, said Anthony. Biomarkers related to chronic lifestyle diseases are heavily studied, and these diseases most often affect low-income and minoritized communities. Thereby, he said, personalized solutions hold potential to change the health trajectories of disenfranchised populations.

FOOD COMPOSITION AND THE FUTURE OF NUTRITION AND HEALTH

Naomi K. Fukagawa, U.S. Department of Agriculture (USDA) Beltsville Human Nutrition Research Center, discussed USDA’s efforts to provide data on food composition. For more than a century, the department has been responsible for monitoring food composition and intake. Its history of research in food composition, nutrition, and health began in the late 1800s and featured the work of Wilbur O. Atwater, who published The Chemical Composition of American Food Materials in 1906 (Atwater and Chas, 1896). In addition to monitoring food supply, intake, and dietary patterns, Fukagawa related, USDA provides food and nutrition data to researchers, policymakers, industry, health care providers, and consumers. She said the agency extends examination of food composition beyond nutrients to the discovery of new components and the identification of bioactive compounds that affect health and wellness. Research on the metabolic use of food components continues, and Fukagawa contended that this research must evolve in response to rapid changes in the food supply that occur with advances in science and technology. Numerous factors affect food composition, she said, including agricultural practices, analytical approaches, communication and data sharing, understanding of diet–disease interactions, climate change, and growing populations. These factors ultimately affect nutrition education and policy, Fukagawa pointed out.

The varied users of food composition information require a “modernized” approach to using the data provided by USDA FoodData Central, which Fukagawa described as an integrated data system launched by USDA in 2019 to provide expanded nutrient profile data and links to related agricultural and experimental research (Agricultural Research Service, 2019). She explained that the system features five distinct data types that are related but not necessarily interchangeable, including (1) Foundation Foods; (2) National Nutrient Database for Standard Reference Legacy 2018; (3) Food and Nutrient Database for Dietary Studies; (4) USDA Global

Branded Food Products database; and (5) Experimental Foods. Information within each data type is collected with unique acquisition approaches and for specific applications. USDA has prioritized efforts to manage high volumes of data to enable data science and personalized or precision nutrition, Fukagawa noted. The department is working to embody FAIR data principles—i.e., data that are findable, accessible, interoperable, and reusable—within a flexible database network (Wilkinson et al., 2016). Findable data contain unique identifiers for data systems. Accessible data are open and free for public use. Interoperable data can be used with other data and other vocabularies. Reusable data enable automated meta-analyses and study replication from similar studies. Effectively using data to enable precision or personalized nutrition involves melding systems biology with data science, she explained.

Fukagawa outlined challenges USDA faces in meeting current data needs. She asserted that dietary intake assessment must be revisited, noting that the Beltsville Human Nutrition Research Center Food Surveys Research Group is undertaking this effort as part of the National Health and Nutrition Examination Survey. USDA must prioritize the analysis of certain food components because analyzing all components is not feasible, she said, and greater understanding is needed of the bioavailability of food components, food interactions, and individual metabolism. Moreover, Fukagawa stated, these mechanisms are affected by factors—such as genetics, the environment, management procedures, and processing—that require deeper exploration within the context of planetary health. A future vision of an integrated food and health ecosystem should include food composition, she maintained, noting that it constitutes a gap in some models. For instance, a model of precision nutrition includes multiple omics (multiomics), lifestyle changes, demographics, habits, and physical activity, but omits food composition (USDA, 2019). Algorithms that predict the healthiest foods and food quantities for individual consumption should include compositional analysis with high-quality data, Fukagawa contended. A sustainable food system provides adequate nutritious foods for the growing population while also protecting the environment. Creating such a system will require understanding how to integrate sustainable production with nutrient optimization (i.e., food composition) and developing new paradigms for realizing sustainable food systems, she said.

Numerous long-term challenges related to food composition remain to be addressed in fostering a strong future for nutrition and health, said Fukagawa. Ensuring data quality is foremost among these challenges, she remarked, as the data need to be transparent, easily accessible, usable, and linkable. Fukagawa stated that multiomics related to food and health outcomes should be incorporated, and that marketing and profit considerations for producers, manufacturers, and consumers are at play. Efforts are needed

to define wellness and establish biomarkers of food quality and health, she said. These challenges can be met through collaborative, cooperative dialogue and public–private partnerships, Fukagawa noted, highlighting the Food Forum as a vehicle for such collaboration. Furthermore, a balance in embracing tradition and new technologies can be achieved via workforce development, she said; a workforce capable of meeting current challenges will feature integrative physiologists, listeners, and communicators—given the general breakdown of clear communication in recent years—and transdisciplinary teams of scientists. Fukagawa stated that the diversity within collaborations should extend beyond disciplines to include different generations, bringing together forward thinking and expertise to solve the problems facing the food system. She emphasized that the focus should extend beyond single nutrients and single outcomes. Instead, she urged taking a systems approach to biology, food production, and ensuring the health of humans and the planet, in order to bridge the availability, accessibility, affordability, and acceptability of food.

FOOD AND NUTRITION SECURITY AND EQUITY

Angela Odoms-Young, Cornell University, described equity trends in the nutrition and health field and considerations for future security and equity. She said that, despite technological advances and the development of advanced data science techniques—including human genome sequencing—over the past 30 years, the high burden of chronic disease in the United States persists. In 2019, she reported, 53 percent of adults aged 18–34 years had at least one chronic condition, and 22 percent had multiple chronic conditions (Watson et al., 2019). She emphasized that the prevalence was higher in some population subgroups, with this high disease burden generating substantial economic and social costs. Odoms-Young said that issues of equity are evident in the higher rates of diet-related morbidity and mortality, maternal morbidity and mortality, and infant mortality experienced by specific subpopulations. Disparities related to race, socioeconomic status, health insurance coverage, geographical location, gender, sexuality, citizenship, disability, and age demonstrate that the disease burden disproportionately affects certain populations, she stated.

Odoms-Young discussed shifts within diet-related disease research and interventions over the past three decades. In the past, she explained, the primary focus was an individual’s choice between healthy and unhealthy foods. She highlighted that research on drivers of health disparities focused on the role of culture, and interventions in the 1980s and 1990s incorporated this focus in teaching culturally appropriate dietary practices and cooking techniques. However, she said, research on social and structural determinants of health has become far more robust and demonstrates that

individual, structural, and social levels all contribute to diet-related disease disparities. Upstream factors—including governance, socioeconomic position, education, social policies, and public policies—contribute to material, behavioral, biological, and psychosocial dynamics, Odoms-Young asserted, which in turn impact equity, health, and well-being, as well as social cohesion and social capital (Solar and Irwin, 2010). Health systems and other systems are drivers of health and affect individual health and dietary behaviors. Moreover, she said, structural oppression drives racial discrimination and disparities in housing and income, while the inequitable distribution of resources drives social conditions that create disparities in food and nutrition security. Odoms-Young cited a study that found the mean 10-year risk of cardiovascular disease, adjusted for age and sex, was significantly higher in non-Hispanic Black participants compared with White participants (He et al., 2015). Further adjustment for education, income, home ownership, health insurance and access to care, and employment attenuated the difference in health outcomes, she reported.

Odoms-Young outlined efforts needed to advance science to achieve improved dietary behaviors, nutrition outcomes, and health outcomes in terms of diet-related conditions. She explained that team science coalesces expertise from fields such as social science, AI, biological sciences, and genetics to consider the modification of various factors to improve health. Implementation science can increase the fidelity and effectiveness of interventions to change systems, she continued; it can also address drivers of poor diet and health outcomes at the programmatic, policy, and population levels. Furthermore, said Odoms-Young, the context of complex structural factors should be considered in downstream interventions such as precision nutrition. She stated that accelerated reduction of health disparities and improved diet and health outcomes require broader approaches that intervene in socioeconomic, environmental, and system-level factors. Odoms-Young asserted that structural, behavioral, and biomedical interventions are needed; that these should be rigorously evaluated and evidence based, she continued, and should address the social determinants of health that systematically lead to and perpetuate social and health inequities (Brown et al., 2019; Goldenberg et al., 2021).

Odoms-Young concluded by describing the concept of Sankofa. Originating with the Akan people of Ghana and embraced throughout the African Diaspora, Sankofa represents the idea that knowing the history and heritage of one’s self and culture enables people to better themselves and their world. Applied to nutritional sciences, Odoms-Young explained, Sankofa can represent understanding historical and cultural factors to build on what has worked well and learn from what has not, such as policies that have generated societal inequities. Illustrating the intersection of structural factors, community resources, food access, and health, she gave the

example of the Tops Friendly Markets supermarket in Buffalo, New York, where a mass shooting took place in 2022. This market offers increased access to food in the area and remains open for business. Odoms-Young pointed out that nutrition and food security exist in a broader context than individual choice alone, and social justice, food justice, food equity, and food sovereignty are components of potential solutions.

DISCUSSION

The discussion following the presentations summarized above focused on food insecurity among U.S. students; the evidence base for precision nutrition; and AI and diversity, equity, and inclusion (DEI) in the food system.

Food Insecurity Among U.S. Students

Ajay P. Malshe, Purdue University, asked about steps to address the high rates of food insecurity facing students. Odoms-Young replied that one of her undergraduate students completed a meta-synthesis on food insecurity among college students that revealed contributing factors and potential solutions. The review found that housing costs affect the income available for food, she said. College students who come from economically disadvantaged families are often in need of additional resources, continued Odoms-Young. She shared that several pilot projects have worked to address campus food needs, such as charitable food distribution via college-based food pantries and enrolling eligible students in food assistance programs. Potential structural solutions include colleges and universities subsidizing meal plans and housing costs for economically disadvantaged students, she remarked. Fukagawa noted that food insecurity affects communities throughout the United States and worldwide. Programs that reuse food waste in a safe manner are a potential solution for providing nutrition to all people in need. Malshe commented that packaged ramen noodles are a popular and inexpensive food choice among students, and the production of more nutritious ramen noodles could benefit this population.

Precision Nutrition Evidence Base

Peter Lurie, Center for Science in the Public Interest, remarked that precision nutrition proposals hold promise for reimagining nutrition service delivery in ways that may be complicated and expensive. He asked about evidence indicating that a precision nutrition approach can improve health outcomes. Anthony clarified that precision and personalized nutrition are different, with the former addressing specific dietary needs of a

group of people with a common condition, such as type 2 diabetes or cardiovascular disease. For example, he said, research has explored the risk of cardiovascular disease and type 2 diabetes in relation to higher-fat versus lower-fat diets. Established markers related to disease can be combined with methodologies for understanding the differences between individuals to target specific dietary recommendations, he said, noting that food companies can support precision nutrition without conducting single-subject studies. As an example, Anthony said that companies can improve the nutritional quality of their products consistent with their consumers’ dietary needs. He added that technology is not yet able to robustly deliver personalized nutrition at the individual level.

Fukagawa remarked that the goal of precision nutrition is understanding how the components of various foods may enhance wellness. Current basic understanding is that humans need adequate food—not too little or too much—and diversity in food intake, she said. People with specific medical conditions, such as metabolic disorders, may benefit from a truly specialized diet, she continued, but a specific prescribed diet for optimal health has not been established for the general population. Lurie replied that specific diets for people with metabolic disorders constitute treatment for a clinical condition. He clarified that Anthony is referring to a wholly different approach to public health nutrition that does away with broad recommendations and instead makes specific, individualized recommendations. Before making such a dramatic change, evidence that this new approach will improve health should be in place, Lurie maintained. In clinical medicine, practitioners prescribe medications that are likely to treat patients’ medical conditions effectively and avoid prescribing medications likely to cause adverse effects in certain individuals, he noted. He then asked about the status of the field of personalized nutrition in terms of being capable of similarly recommending intake or avoidance of specific foods. Anthony stated that a few studies have examined the benefits of individualized diets versus general dietary guidance, but the evidence base is still developing. He remarked that widespread guidance, such as reducing sodium intake, is helpful and is supported by data, but dietary guidance does not necessarily address behavior. Anthony contended that personalized approaches may influence behavior more effectively by connecting recommendations to specific needs or goals. Whereas general guidance tells people what to do, he explained, precision nutrition addresses individual needs and offers recommendations that are consistent with a person’s needs, culture, and preferences. Furthermore, precision nutrition is not limited to medical needs or food as medicine; it extends to individual needs and goals that may not be related to medical conditions.

Artificial Intelligence and Diversity, Equity, and Inclusion in the Food System

Christina Chauvenet, Newman’s Own Foundation, asked about the opportunities and challenges associated with AI in the food system in terms of health equity and nutrition, noting that AI sometimes perpetuates biases. Odoms-Young acknowledged issues of bias within data and algorithms that train AI across fields. Data science groups are working to address these issues, she said, but much work remains to be done in this area. Some companies are exploring the intersection of AI and DEI with a specific focus on using AI to address social inequities, she noted. And she pointed out that developing the capability to adjust inputs in terms of differences in housing, income, and other social and structural determinants of health holds promise in using AI to achieve the best outcomes possible. The nutrition space often focuses on a physiological or biological focus; broadening this to include biological inputs and omics with social and structural determinants of health could yield a better understanding of how to address inequities, said Odoms-Young. She noted that evidence on the effectiveness of DEI approaches indicates that DEI works by informing the types of questions scientists ask and the methods they apply. For instance, she said, researchers should consider social and structural determinants of health in community nutrition, and DEI efforts in the field of nutrition could also support diverse training and practitioners. She added that the registered dietitian workforce is not currently representative of the populations bearing higher disease burden, such as racial minority groups, disability communities, and LGBTQ+ communities.

Khoo asked about the use of technology to increase knowledge of food composition and address needs within population subgroups. Fukagawa stated that these efforts begin with high-quality data, which can be expensive and challenging to collect, especially given that new compounds appear in food that have claims of related positive physiological outcomes. Once sufficient quality data are achieved, she noted, AI can be used to help identify solutions. Multiple factors and nonlinear relationships are likely at play in terms of the effects of food composition on health outcomes, she said, adding that partnerships and data sharing between industry and academia are mechanisms for collecting adequate quality data. Fukagawa emphasized that food is not a pharmaceutical, food is a necessity. As people eat throughout their lifetimes, their bodies respond to different foods in various ways depending on environment, genetics, and various other factors, she said. Food cannot be regulated or prescribed in the same way that pharmaceuticals are, she explained, as the former is imperative and the latter optional. Fukagawa shared that she uses the acronym “DEIA” in reference to foods to recommend that one’s plate should feature diversity

of foods, equity across food groups with no one group overrepresented, inclusivity by using multiple foods and approaches to cooking, and being affordable and accessible.