"Personalized nutrition” isn’t going to solve our diet problems

There's no end to Bay Area startups floundering with their promises to revolutionize health. Theranos tried (and failed) to democratize blood testing, the Apple Watch vowed to get people exercising (despite evidence to the contrary), and a company called Habit wants to fix our diets.

Habit is peddling a "personalized nutrition" service. In exchange for $299 and samples of your blood and saliva, you get reams of raw data about how your body responds to different foods, access to the Habit app, and a session with a certified dietitian. For more money, you can purchase meals shipped directly to your home, as well as customized eating plans.

At first glance, the concept is appealing — an individualized, paint-by-numbers food program for optimal health. Who wouldn’t want to know themselves at the "cellular level," to "eat better, feel better, be better," as the Habit introductory video promises? The food giant Campbell Soup was certainly sold, investing $32 million into the company.

When I started to poke around, however, my skepticism about Habit — and the other direct-to-consumer personalized nutrition programs out there — grew.

Researchers who work in this area told me there’s a huge gap between what we actually know about nutrition genomics and what companies like Habit claim to offer.

"We still don’t have the ability to accurately predict the most healthy diet for an individual … with or without the use of genomics," said Rasmus Nielsen, a geneticist at the University of California Berkeley. As Lund University genetic epidemiologist Paul Franks put it, "The concept is probably not quite ready for public consumption."

Habit thinks your genes can tell you if need you to eat more protein

Before diving into what Habit does, let’s back up for a moment. While many of the companies trying to commercialize personalized nutrition — or "nutrigenomics" — are new, the idea has been around for a while. Researchers have for years been interested in how different people respond to different foods — and why some people absorb certain essential nutrients better than others do. And leading nutrition researchers have argued that we will eventually move away from the "one-size-fits-all" approach to food toward more individualized eating plans. (More on this later.)

A number of entrepreneurs have seized on this dream and are trying to develop diets plans tailored especially to you. (Yes, they mean all of us.) The website for DNAFit, for example, opens with: "Let's talk about you. Our groundbreaking DNA test will change the way you think about fitness and nutrition forever." Another company, Nutrigenomix, offers to help you "Eat According to Your Genes." (The emphasis is mine.)

Habit, headquartered in the Bay Area, is just the latest variation on this theme. As many personal health gurus have done, CEO Neil Grimmer wrote at Medium about the health epiphany that led him to start the company. He gained a bunch of weight, had a personal crisis, and then underwent a transformation by, he says, hacking his biology — just as he’d like others to do with Habit.

"I want to democratize the process," he wrote, and help people "live a healthier, fuller life powered by foods that are just right for them."

The company he built is a kind of mashup of the at-home genetic testing company 23andMe and the personalized meal recommendation and delivery services Prepd and Blue Apron. (Hat tip to Ringer for pointing that out.)

After you submit blood and saliva samples, Habit gives you data on more than 60 biomarkers (for example, blood glucose, vitamin levels, and cholesterol) and feedback about how particular genes may determine responses to different foods — for example, sensitivity to caffeine or lactose intolerance.

All that information is then processed through Habit’s proprietary algorithms and uploaded to the Habit app, where users can log in for a half-hour session with a dietitian to learn about their results and which "Habit Type" and "Hero Foods" they best respond to. The "Protein Seeker," for example, is someone who should eat a diet high in protein, since his body has a harder time breaking down fat and carbohydrates. He also has the option of ordering tailored eating plans and meals — for extra charges.

One thing that sets the service apart, Habit science adviser Josh Anthony told me, is that it takes a "systems-based approach." He said, "We are using multiple measures that integrate users' preferences, behaviors, goals, in addition to [data about] where your body is right now — and bring in blood markers as well." In other words, while some personalized nutrition companies use only DNA or a person’s dietary habits to design a diet, Habit takes a bunch of different data and grinds it through algorithms in the hopes of returning advice that will motivate people to eat better.

Habit hasn’t tested its approach and won’t release its algorithms

Tom Merton

I asked Habit what scientific evidence informed this "systems-based approach" and for some detail about its algorithm. The company sent along a thoughtful reply, filled with analysis on some personalized nutrition studies I’ll mention below.

But its methods have never been tested in a clinical trial. (Habit is not regulated by the Food and Drug Administration, so it didn't have to go through preliminary studies to prove their product.)

"What we need to get to is the extent to which we can ultimately drive sustainable behavior change as well as quantifiably greater outcomes [than other personalized nutrition platforms]," he said. "That’s the type of research we’re in the process of doing."

So for now, the program is based on proprietary algorithms that haven't been made public, haven’t been studied (beyond pilot-testing the program for ease of use), and haven’t been published in scientific journals. (That’s another parallel with the super-secret and now-defunct Silicon Valley startup Theranos.)

I ran Habit’s responses to my questions by researcher Paul Franks to get his feedback. "There’s nothing in their response that provides anything useful to help evaluate their service," he said. He wasn’t convinced about the genes Habit was using to customize users' diet plans, "and the rest of the information they provide seems to hinge mainly on generic information about diet and health, which a good dietitian would recommend anyway."

Franks wants proof of the Habit algorithm to understand whether the approach has any substance, he added. "The field of nutrigenomics — or whatever one wants to call it — needs to aspire toward the standards for proof, disclosure, and interpretation the drug industry is made to live up to. Until that time, it’s hard not to be extremely skeptical."

John Mathers, the director of the Human Nutrition Research Center at Newcastle University, also noted that the lack of evidence behind nutrition companies makes them dubious. "I am skeptical about many of these products because of the slender, or nonexistent, scientific basis for them," he said.

Mathers would know. He’s been involved in the Food4Me study, a huge European research project aimed at understanding what impact personalized nutrition could have on people's health choices. The researchers recruited more than 1,200 adults from seven countries and tried to figure out how people responded to personalized nutrition treatments based on a few different targeted interventions.

So far, the evidence out of Food4Me isn’t very compelling. In one study, the researchers found study participants who got some form of personalized diet advice improved their eating habits more than the control group. But, they concluded, "There was no evidence that including phenotypic and phenotypic plus genotypic information enhanced the effectiveness of the [personalized nutrition] advice." The different forms of personalization made no difference to the outcome of the study.

In another Food4Me study, researchers found something a little more promising. They looked at whether a personalized nutrition intervention could help people adhere to the Mediterranean diet. Again, those who got personalized nutrition advice had an easier time sticking to a healthy pattern of eating than those who got generic diet advice. This time, the group that got the additional DNA-based guidance fared even better than the groups that didn’t. But, the researchers wrote, "Although differences were significant, their clinical relevance is modest." So they wouldn’t have much of an impact on health.

Those are just two studies, though. Let’s take a peek at the reviews of the literature in this area.

A systematic review of the research on the impact of giving people genetic risk information in the hopes of changing their health behaviors concluded there’s no evidence this method works.

A 2015 meta-analysis looked at the research behind commercially available "nutrigenomics" testing companies and found there was essentially no reliable data to support the claims these companies make. "As solid scientific evidence is currently lacking," the researchers wrote, "commercially available nutrigenomics tests cannot be presently recommended."

In one of the most rigorous tests of personalized diets to date, a 2018 randomized controlled trial published in JAMA, researchers once again found no difference in weight loss between groups of people who were matched to either low-fat or low-carb diets based on their genotypes.

This dearth of evidence is why the Academy of Nutrition and Dietetics advises against direct-to-consumer nutritional genomics programs.

"From a scientific perspective, [personalized nutrition] is problematic," said Tim Caulfield, a health policy researcher who has been tracking personalized medicine since the 1990s. "For most human beings, this information isn’t relevant. It’s not going to benefit us any more than the basic health advice."

The focus on individual nutrition also has troubling policy implications, Caulfield said. "It puts the lens on the individual as opposed to the society." At a time when we know many of our chronic health problems — diabetes, obesity, many cancers — are driven by our environments, he added, personalized nutrition programs lead us in the wrong direction.

Nutrition genomics holds real promise

So science suggests the direct-to-consumer personalized nutrition programs currently on the market are not going to revolutionize health and, like many fad diets, distract from what we know about health.

For the sake of science, though, we should keep an open mind. Nutrition genomics is a genuinely fascinating area — even if it’s not yet ready to be applied in our daily lives.

duncan1890

Consider this Science study on the Inuit people in Greenland published in 2015. Given their harsh Arctic environs, the Inuit traditionally relied on a diet that was spare in fruits and vegetables but rich in whale, seals, and fish. These animals are high in fat, and yet the Inuit have a low rate of heart attacks — a paradox that has long fascinated researchers.

According to the Science paper, it turns out that the Inuit evolved unique genetic adaptions that helped them respond to a diet rich in the proteins and omega-3 polyunsaturated fatty acids found in fish. So that would mean people with these genetic variants probably aren’t as sensitive to some of the health harms that come with high-fat diets as others may be.

There are other examples where one’s genotype may help guide personal food choices. People with the PKU genetic variant have trouble breaking down the amino acid phenylalanine and need to avoid protein-rich foods (milk, dairy, meat) that are high in phenylalanine. People with LCT genetic mutations have trouble digesting lactose and need to avoid dairy.

There’s also strong evidence that people have unique blood glucose responses to identical meals, which the researchers suggest can help inform people’s eating habits. (The algorithm used in this study formed the basis of one very targeted — and relatively compelling — personalized nutrition company focused only on delivering blood sugar information.)

So, as Franks put it, "There are well-known examples of where genotype might help guide diet choices." But they are still pretty narrow and specific. "Services [such as Habit] purport to go much further than these rare examples."

The nutrition advice most of us need to follow is pretty simple

While scientists and entrepreneurs parse nutrition at the individual level, let’s not forget that most Americans aren’t meeting basic recommendations of fruit and vegetable consumption or exercise. And there’s a risk that fine-tuning one’s diet according to DNA and blood biomarkers once again will overcomplicate the basics of a healthy lifestyle. It also may rip people off (pretty much all diet products do).

Eat a diet that’s rich fruits and vegetables, avoid junk food, don’t smoke, exercise regularly, and don’t consume too many calories. Any changes we make based on our personal health information would be minor compared with that general guidance.

"At the end of the day, if we really could get the majority of the population to eat wholesome real foods most of the time — regardless of macronutrient composition — most of the nutrition-related problems would be resolved," said Christopher Gardner, a professor of nutrition at Stanford who authored the new JAMA personalized nutrition trial.

Still, he’s optimistic about the future of personalized nutrition. "I really do believe there is not one diet for all, and that there are different diets that are more appropriate for some people than others," he said. We’re just nowhere near that yet.

We're here to shed some clarity

One of our core beliefs here at Vox is that everyone needs and deserves access to the information that helps them understand the world, regardless of whether they can pay for a subscription. With the 2024 election on the horizon, more people are turning to us for clear and balanced explanations of the issues and policies at stake. We’re so grateful that we’re on track to hit 85,000 contributions to the Vox Contributions program before the end of the year, which in turn helps us keep this work free. We need to add 2,500 contributions this month to hit that goal. Will you make a contribution today to help us hit this goal and support our policy coverage? Any amount helps.

One-Time Monthly Annual

$5/month

$10/month

$25/month

$50/month

Other

$

Yes, I'll give $5/month

Yes, I'll give $5/month

We accept credit card, Apple Pay, and Google Pay. You can also contribute via