In Japan, Nestlé’s Wellness Ambassador program provides personalized nutrition by allowing users to upload photos of the food they’re eating. Its app then uses AI to analyze the photos and combines that with personal data from DNA and blood tests to recommend nutrient-infused teas, smoothies and snacks. The program had picked up around 100,000 users.
Then there’s Care/Of, which invites its customers to complete a quiz that details key factors like their gender, lifestyle habits, age, diet and existing health problems. It can then use an algorithm to create a personalized pack of vitamins, saving users the time and effort of looking for suitable products themselves. Their vitamins are then delivered once per month through a subscription service.
A similar company called Rootine offers up microbeads that can provide a personalized dose of vitamins throughout the day. They say that this mimics the way that real food is absorbed. Rootine uses an algorithm that analyzes data from genetic testing, a lifestyle questionnaire, and bloodwork to create tailored microbead combinations for users.
One of the challenges of this approach is that online questionnaires have been criticized for being reductionist. The critics argue that any nutritional advice that’s based on a questionnaire is almost certainly less accurate and less beneficial than tests that are overseen by healthcare providers. However, it’s important to remember that there’s not a lot of evidence for the more extensive testing, either.
Wearables can provide us with hard data, but their use cases are limited and several nutrition-recommendation tools are still asking their users to self-report their activity levels and what they’re eating. Even if we assume that the nutrition platform is giving good advice, there’s a risk of its accuracy being undermined due to missing data or personal biases. This challenge is unlikely to be resolved any time soon, but wearables can at least help to provide us with a larger amount of data as the technology improves, taking some of the burden to report away from users.
AI can help by analyzing what people eat, understanding its nutritional content and providing feedback and guidance based on the patient’s goals. Some nutrition studies are asking participants to photograph their meals so that the images can be processed by deep learning tools to accurately determine what they’re eating. This avoids the hassle of manually logging the data and the use of unreliable food diaries. Other companies are playing with novel approaches that might pay off in the future, such as by analyzing the sound of chewing. Other companies like Vessyl track and analyze what you drink.
Many of the most common wearable devices, including those by Apple and Fitbit, can be used to log meals and track diets, usually through manual entry or by the user scanning barcodes. They’re then able to provide an estimated calorie intake along with a breakdown of micronutrients. The Georgia Institute of Technology was able to create a mouth wearable that monitors salt intake and uses sensors to wirelessly report back to an app to help its users to manage the sodium in their diet. Similarly, Tufts University researchers developed a tooth-mounted sensor that detects glucose, salt and alcohol intake.