LSU’s Pennington Biomedical Research Center has begun studying personalized diets created through predictions made using artificial intelligence.
The goal, researchers said, is to lower obesity rates in America. According to the Louisiana Department of Health, nearly one out of four adults in Louisiana is considered obese.
The program, titled Nutrition for Precision Health, will collaborate with the All of US research program, which is creating the largest precision medicine database in the world.
Eric Ravussin has been working at Pennington for 22 years and is one of the leading investigators for this research, funded by the National Institute of Health.
Ravussin explained that dietary habits are directly associated with health and disease, using the example of how vegetarians are healthier than people who consistently eat fast food.
According to Ravussin, nationwide rates of obesity were around 15% in the 1980s. Now, it’s closer to 35%.
“It’s an enormous increase, and it’s being triggered by our lifestyle,” he said. “Nutrition is an important part of that.”
The study will start with 12,000 people from across the United States, with six clinical centers responsible for conducting the research. In total, there are 14 principal investigators.
Sample group’s nutrition and dietary habits will be measured while they constantly wear a glucose monitor for 10 days.
According to Ravussin, the 12,000 people will represent everyone across the United States, including individuals with disabilities and pregnant individuals. The sample size will range in age from 18 to an upper age level of 95. They will not accept people with Type 1 diabetes to avoid changing their insulin injection regimen.
“We will know about their genetics, we will know about their socio-economic status, we know about where they live, we will know about their family history of disease and health,” Ravussin said. “This will be a cross-sectional study.”
Ravussin explained how out of the 12,000, 1,500 will be provided with three diets for two weeks. These diets will be referred to by colors and the participants will eat from all three diets without knowing which one they are on. They will start with one, continue with another and then finish with the last one with a washout period in between.
“There is a traditional American diet, which is the least healthy one,” Ravussin said. “There is a Mediterranean diet and there is another diet that is rather high in fat but in healthy fat.”
Around 500 to 600 people will continue into a study where they will stay at the inpatient facility at Pennington three times over two weeks. There, they will be fed those three diets and strictly monitored and ensured they follow the diet.
“People are people,” Ravussin said. “If you tell them just to stick to the diet and nothing else, we know that half of the people are not going to play the game and are going to have their ice cream or their three glasses of wine.”
The data from the participants will then be collected alongside information like whether they live in a food desert or a healthy area. The responses will then be put into an artificial intelligence processing system to build predictors of what diet works best for a specific person.
“Diet A might not be the best diet for this given person, but may be the best-given diet for another person,” Ravussin said.
Ravussin and other researchers have submitted a protocol to a Data Safety Monitoring Board, where they will work back and forth until the study begins early next year.
Lucio Miele, chair of genetics and assistant dean for translational science at the LSU Health Sciences Center, has been at LSU since 2014 and serves as a principal investigator in Nutrition for Precision Health.
“The goal of NPH is to determine how different individuals respond to their diets in terms of their metabolism, microbiome, physiological measures, metabolisms, etc.,” Miele said. “The final goal is to define and validate algorithms that predict how each individual will respond to diet.”
He said that artificial intelligence would be employed in the “analysis phase,” where diets will be appropriately matched to each subject.
“The ideal outcome is a series of predictive algorithms that help nutritionists and physicians customize the ‘best’ diet to maintain health and prevent diseases associated with dysmetabolism,” Miele said.
Miele is also the co-chair of the Community Engagement Committee for the NPH study, gathering input from communities and All of Us participants to design the most effective ways to disseminate what is found in this study to communities.
The All of Us research program is creating the largest medicine database in the world. Miele said that studies like this one will “work to improve the health of all Americans.”
Leanne Redman, professor of clinical sciences at Pennington Biomedical Research Center, has been there since 2005. She is the co-principal investigator for the Louisiana Nutrition for Precision Health Clinical Center, which will run the study in the state. She is also the co-chair of the Steering Committee, which includes over 450 individuals nationwide.
Redman said that scientific research and other decisions that regard the efficacy of things such as diet, medication or behavioral treatment are based on a difference in the treatment that a control group receives.
“Often overlooked is the wide range of benefits that are experienced among people in each group,” Redman said. “Some people might experience many health benefits when eating a low-fat diet, whereas other people might experience no benefits and some people might even experience worsened health effects.”
