Artificial intelligence could help develop the best diet for each individual

CNN
—

Is broccoli or beets the best option for your personal health? What foods can trigger the biggest spikes in glucose or blood cholesterol spikes?

Personalized diets are the next frontier in nutrition science, and artificial intelligence (AI) will play a key role in determining what each of us should and shouldn’t eat.

An ambitious program called Precision Health Nutrition (NPH) began in the US in January 2022 when the National Institutes of Health (NIH) awarded $170 million to fund institutions across the country to conduct a five-year study with 10,000 participants.

CNN spoke with Holly Nicastro, a program director in the NIH Office of Nutritional Research and coordinator of NPH, about the goals and scale of the project and how AI can benefit our health by helping to design optimal diets for each individual.

This interview has been edited and abridged for clarity.

CNN: There is already a lot of information about healthy eating. Why is your approach different and what is the goal?

Holly Nicastro: There is a lot of conflicting information and there is no single diet to stay healthy. NPH’s goal is to use AI to develop algorithms that predict individual responses to eating and dietary patterns.

Our approach is different because we are looking at a comprehensive set of factors, many of which are not typically examined in nutrition science. The NPH study will examine how genes, the microbiome, biology and physiology, environment, lifestyle, health history, psychology, and social determinants of health affect individual responses to diet. In addition, we will study one of the largest and most diverse participant cohorts for a precision nutrition study.

CNN: Who are your contestants?

nicastro: NPH participants will be recruited from the All of Us Research Program, which is run by the NIH. All of Us is inviting one million people across the US to help build one of the most diverse health databases in history. Most of the participants come from groups that have previously been underrepresented in the biomedical sciences. All of Us participants have provided information through surveys, electronic health records, biospecimens, and digital health technologies like Fitbits. The diversity of the All of Us Research Program cohort will allow NPH to examine factors such as age, gender, race and ethnicity.

CNN: What data do you collect and how do you analyze it?

nicastro: NPH incorporates three modules. In Module 1, information will be collected on the usual day-to-day diet of all participants. In Module 2, a subset of Module 1 participants will consume three different diets selected by the researchers. For Module 3, a smaller and separate subset of Module 1 participants will participate in a two-week study at research centers where researchers will carefully monitor their eating.

Each module will conclude with a food challenge test. Participants will fast overnight and then consume a standardized breakfast or drink so that we can examine their responses, such as blood glucose levels, over several hours.

We will use mobile photography apps and wearable devices that can passively capture information about what people eat. Participants will wear continuous glucose monitors and accelerometers that collect information on physical activity, sedentary time, and sleep. The researchers will also analyze various biomarkers, such as blood lipids and hormone levels, and the microbiome of the stool.

The NPH study will link data from the All of Us program and analyze it using AI. This approach creates an unprecedented opportunity to drive nutrition research toward personalized nutrition because, unlike human researchers, AI can quickly analyze and process large amounts of data and translate connections between data points into algorithms. These can predict an individual’s response to food and dietary patterns, taking into account the role of genes, proteins, the microbiome, metabolism, and environmental and lifestyle factors.

CNN: Who will benefit most from precision nutrition approaches?

nicastro: Some of the first direct benefits could be for people at risk of diabetes or who have trouble regulating their blood glucose levels. Blood glucose monitors that are attached to the skin let us know how a person’s blood sugar will change after eating certain foods, food groups, or meals, and then predict those responses based on the characteristics of the person’s blood sugar. person. This will help us develop personalized plans that prevent large swings in blood sugar.

We are also using precision nutrition approaches to see how well we can predict other responses to diet, including changes in blood pressure, blood cholesterol or triglyceride levels, mood, and cognition.

CNN: Are we really what we eat, and how much can we improve our health through diet?

nicastro: Poor diet is one of the leading causes of preventable disease and death worldwide, and is a big driver of healthcare costs. Our diet affects our growth, development, risk and severity of disease, and general well-being. Worldwide, nearly 40% of adults are overweight or obese, more than 30% have hypertension, and other diet-related chronic diseases are on the rise. Recent studies have indicated that about one in five deaths can be attributed to poor diet. Improving our diets has tremendous potential to change the health of the world.

If everyone followed population-wide guidance, like the Dietary Guidelines for Americans, we would expect to see improvements. These guidelines generally focus on including nutrient-dense foods and limiting added sugars, saturated fat, and sodium. However, this one-size-fits-all approach only gets us so far. We see some degree of individual variability in participant responses in virtually all dietary and nutritional intervention studies, and patients and consumers are increasingly seeking personalized approaches.

CNN: I’m convinced! When will personalized dietary advice be widely available to the public?

nicastro: Precision nutrition is already in practice. Clinicians will provide different recommendations based on a person’s disease profile, health history, or health and wellness goals. For example, the recommendations for someone looking to manage their diabetes will look very different from the recommendations for someone who needs to increase their muscle mass.

Nutritional advice will become increasingly accurate in the coming years. In the short term, more data points will be used to generate more personalized recommendations. In the longer term, I look forward to seeing the predictors identified by NPH used during standard intake by healthcare professionals. This may involve patients using new technologies, such as continuous glucose monitors or smart toilets that analyze the microbial composition of stool in real time, or it may involve a simple genetic signature test.

For us to realize the full benefits of precision nutrition approaches, it will be essential to study and address the barriers to following dietary recommendations. Precision approaches should focus on dietary recommendations that not only optimize a person’s health, but are also easy for a person to follow based on their own resources, lifestyle, preferences, and abilities.