Noosheen Hashemi, Founder & CEO, January AI
America is in metabolic disarray; diabetes is the most expensive chronic disease in the US according to the American Diabetes Association (ADA). One in every four healthcare dollars treats someone with diabetes, and the number of people with diabetes in the US has grown 3-fold since 1999, from 10.9 Million to over 34 Million. The numbers grow to over 120M, once you add in the 88M with prediabetes. Twenty two percent of people with diabetes and ninety percent of people with prediabetes remain undiagnosed; this epidemic is a ticking time bomb that will greatly burden the U.S. population in dollars and human suffering. One company hoping to help people on the diabetes spectrum is January AI, a precision health start-up, founded in 2017 by Silicon Valley veteran and CEO Noosheen Hashemi and Mike Snyder, Chair of Genomics and the Director of Personalized Medicine at Stanford University. They are harnessing science, medicine, and machine learning to transform how individuals control their health.
While multiple technology solutions have appeared in the last decade to address the increasing prevalence of diabetes, innovation has been marginal, making the sector ripe for disruption. One set of “modern” Diabetes Prevention Program offerings simply digitized the dinosaur by providing access to curriculum and coaching via the web and mobile app. Another set uses painful finger pricks and connected blood glucose meters to take frequent readings and get people to manage their HbA1c. Many rely on coaches as their primary intervention, but coaches are too expensive to interact with 100 plus million people who would need them. The advice dispensed is generic and ultimately not effective for everyone. The fundamental problem with any solution that relies on blanket recommendations is that we now know that glycemic responses differ, not only between people but also for a given person depending on what they eat, when they eat it, their activity and sleep patterns, and so on. All of these solutions put together are reaching fewer than one million people. One reason is the lack of personalized data needed to understand the effects of food, activity, stress, and sleep on blood glucose and programs that integrate this highly personalized information into people’s daily lives. Machine learning (ML) can combine food and activity data with novel markers like Time-in-Range (TIR) – a valuable output produced by continuous glucose monitors (CGMs) – the corpus of science and standard of care to gain insights that are otherwise invisible and imperceptible to everyday human observation. Importantly, ML algorithms “get smarter” over time as they learn individual user behavior.
Different from the prevalent reactive healthcare status quo, January AI is taking a preventative approach to lifestyle-driven chronic conditions, complications, and comorbidities. In the presence of comorbid conditions, healthcare costs grow exponentially, up to 10x the cost of treating diabetes alone. Since 98% of people with type 2 diabetes have at least one comorbid condition, and over 90% have two or more, preventative measures are crucial in managing the disease over the long term.
COVID-19 unmasked the severity of chronic conditions in the U.S. A high percentage of people who didn’t survive the illness have underlying chronic conditions. 2 Hospitalizations were six times higher and deaths 12 times higher for people with chronic conditions like diabetes, cardiovascular disease, and chronic lung disease. Telemedicine has emerged to address these gaps in care. According to McKinsey research, COVID-19 has increased the use of telehealth, from 11 percent of US consumers using telehealth in 2019 to 46 percent. The US telehealth market is expected to reach $10 billion by the end of the year, and with the acceleration of consumer and provider adoption, up to $250 billion of current US healthcare spend could potentially be virtualized.
January AI’s telemedicine is able to help people get the tools and expert analysis that they need to guide their own health journey. “We believe that people own their health data and should have the opportunity to combine and integrate it to benefit from more comprehensive and actionable insights. They are investing time and financial resources in logging their food, tracking their heart rate and glucose, undertaking elimination and other diets, ordering food sensitivity, genetic and microbiome tests. Wouldn’t it be wonderful to be able to combine these someday?” asks Hashemi. “We are investing in helping physicians and coaches to deliver the best care that they can while empowering users to know the most about their bodies, self-advocate and become their best self,” she continues.
Using Machine Learning to Accurately Predict Glycemic Response
January AI recently presented findings from an in-house study of over 1000 participants at the American Diabetes Association Scientific Sessions. January AI’s new AI algorithm effectively predicted individualized glycemic responses to specific meals in healthy users, individuals with type 2 diabetes, and individuals with pre-diabetes. Each group also improved their TIR. This study is the first in a series designed to establish a body of evidence for the January AI program and its ability to motivate behavior change and help participants achieve better clinical outcomes. The program uses technology to unify the corpus of science with multiple data sources to produce actionable insights. January AI built its own mobile application to capture and unify various data points into one AI platform, collecting nearly 25 million data points for the study. On average, 64% of participants significantly improved their TIR over the course of the challenge, and 28% of participants improved TIR by 10% or more. January AI is planning for more future clinical studies of longer duration. “While general guidance has existed for some time regarding the effect of dietary composition on blood glucose of people with diabetes, the reality is that food, exercise and medicine impact each of our bodies differently,” said January AI Co-founder, Dr. Michael Snyder. “Using technology to provide individualized predictions about the glucose response to specific meals may provide benefits beyond general dietary guidelines, enabling people with diabetes to make healthier ‘personalized’ food choices.”
An Evidence-Based and Integrated Approach to Dialing Lifestyle
January AI unleashes the power of machine learning to broaden the scope for successful outcomes – it can be hard for people to drop 25 pounds at once but easier to reduce consumption of high glycemic load foods which over time can lead to weight loss. 3 January AI has made a significant investment in understanding how people choose their food and the intrinsic glycemic response a person has to the foods they consume. Food logging is notoriously erroneous within the industry; crowdsourced food databases lack integrity and individuals generally under-report what they eat. January AI has attempted to address each of these problems and has aggregated the most complete food database available. January AI’s advanced technologies infer glycemic index and glycemic load for all foods in their database, over 16 Million! For example, a personalized recommendation informed by machine learning may encourage a user to swap out regular pancakes and consume avocado buckwheat pancakes instead. This recommendation provides a healthier alternative based on an individual’s predicted glycemic response to pancakes, but also takes into account that the user enjoys eating pancakes based on past behavior. January AI realizes that individuals will be most satisfied if they are able to consume enjoyable, healthy, and familiar foods, and so it recommends personalized alternatives that are predicted to lead to a favorable glucose response. The recommender is optimizing for the healthiness of foods at the personal level based on how much glycemic load, fiber, saturated fat, and total calories they have, answering the everlasting question of “what should I eat?”
We know that activity has a direct impact on blood glucose but how much activity do we need to do and when? January AI’s qualitative research shows that people are more likely to take action when the recommendation is specific. Those with well-controlled diabetes know exactly what to do while those with uncontrolled diabetes know generally what to do. January AI’s predictive technology can let a user know how much activity they need to bring the glycemic response to a particular food within a healthy range. Eating an indulgent food for one person may require 25 minutes of walking but the very same food for another person may require 35 minutes. People can use the glycemic prediction for healthy living in their everyday lives. They can look up their potential glycemic response to a food before they eat it to decide whether to have it or not. They can use it to compare two foods at the grocery store, or two recipes or two menu items, etc. This technology is unprecedented.
Broadly recognized health interventions like intermittent fasting accompany these recommendations to help users achieve results quickly. Part of intermittent fasting is common sense, as the body uses fat and other energy reserves as fuel instead of food, but researchers have found more significant benefits as well. The University of Alabama conducted a study on a small group of obese men with prediabetes. They compared a form of intermittent fasting called “time-restricted eating,” where all meals were fit into an early eight-hour period of the day (7 am to 3 pm) or spread out over 12 hours (7 am to 7 pm). In both groups, the participants’ weights did not change during the study, but after five weeks, the eight-hours group had dramatically lower insulin levels and significantly improved insulin sensitivity, as well as significantly lower blood pressure. The best part? The eight-hour group also had significantly decreased appetites. They weren’t starving and simply changing the timing of meals, by eating earlier in the day and extending the overnight fast, significantly benefited metabolism even in people who didn’t lose a single pound. January AI’s technology can tell when a person has eaten and can help users seamlessly monitor and manage their eating and fasting schedules. Other fasting apps in 4 the market depend on user reported information which, when it comes to eating, can be faulty.
For people with diabetes, it’s important to keep blood glucose levels within a specified range. Whenever blood glucose gets too high or too low, it can cause damage to the body and internal organs. The longer that blood glucose remains high or low, or the more extreme these levels become, the more damage done to the body. To help users maintain a healthy range, January AI reports how lifestyle choices impact metrics like TIR, Area Under the Curve (AUC), and the Glucose Management Indicator (GMI). AUC builds on TIR, incorporating both how frequently blood glucose levels are out of range, along with the severity of these glucose excursions. In contrast, GMI reports an average glucose value similar to gold-standard industry metrics like HbA1c that were developed in the 1970s. GMI can be quickly calculated from modern medical devices without the need for finger-pricking. GMI, combined with TIR and AUC, provide a much more holistic, timely and actionable set of benchmarks for both users and their physicians to optimize glucose patterns, greatly improving on HbA1c that has been known to ignore important glucose fluctuations and be misleading based on genetic variability and certain medical conditions.
The Fastest Path to January 1
Dubbed “The Season of Me,” January AI’s program consists of a mobile app, a heart rate tracker, and a CGM that users wear during the initial phase of the program. Starting the program is easy, as January AI has created a telehealth platform providing seamless access to the required devices, ensuring as many individuals as possible can benefit. For the initial two weeks, January AI’s proprietary algorithms quickly learn the user’s health status and responses to different lifestyle choices as users learn how blood sugar reacts to different foods and activity through clinically designed experiments. This is where users begin receiving food and activity suggestions from January’s recommendation engine, and are introduced to time-restricted eating, sometimes referred to as intermittent fasting. In the middle of the program, users receive a midpoint evaluation to reflect on learnings and deepen their understanding of their personalized glycemic response patterns. For the following two weeks, users integrate learnings into a daily routine as part of a process to improve their behaviors, the beginning of their journey towards a healthier lifestyle. In contrast to similar services on the market, January AI produces intuitive and approachable personal insights that can be integrated into the user’s holistic healthcare plan, promoting healthier lifestyle routines in as little as 30 days to make it one of the fastest ways for consumers to improve their glucose habits.
We Can’t Change the Past, but We have the Power to Define the Future
The Ancient Romans imagined Janus — the guardian of time, transitions, beginnings, and ends — as a two-faced god, one facing backward and one facing forward. In his honor, the month of January became a time of symbolic renewal, during which Romans could both accept the past and plan for the future. If the cards that are dealt to us like genetics and the foods we grew up with represent the past, deciding how we live, as a people of reason, is our future. 5 “At January AI, we believe that every day is January 1st,” says Hashemi. “That every day is a fresh start. We believe that health isn’t a place you reach, but the simple process of doing just little better than yesterday. We believe that self-improvement is a team sport.”
That’s why January AI is creating a community of everyday people building their own healthy habits, together.
Hashemi concludes: “We make radical change radically easy. January AI’s cutting-edge technology learns from your habits and your body to give you one-of-a-kind recommendations that are healthy for you, not just healthy in general. Then, we break down everything into tiny steps you can easily achieve every day. Finally, we connect you to others on the same path, making sure you have support for every setback, and celebration with every win.”
For people with diabetes, their care teams, or anyone who simply wants to optimize their blood sugar, January AI has the only solution of its kind. Be on the lookout this summer for the availability of the January app and the Season of Me Program, available at www.january.ai.