Web Content Editor Ian Bolland spoke to Dana Chanan, CEO and co-founder of Sweetch, an artificial intelligence-based digital health coach that covers metabolic syndrome diseases. Topics discussed included Sweetch’s origins, what next for digital health and wellness, and how the likes of Amazon and Apple will affect the sector.
First of all, where did the idea for Sweetch come from?
I have always been fascinated by human behaviour – and specifically technology’s potential to impact it. My background is in creating positive user experiences and generating long-term engagement through the gamification of banking services and other industries. I joined forces with Dr. Yossi Bahagon, Sweetch’s co-founder and chief medical officer, to create Sweetch out of a shared belief that it is possible for technology – specifically, the integration of behavioural change science with best-of-breed user experience – and advanced analytics to bring about lasting, highly personalized behavioural change and help people adapt to a healthier lifestyle at the population level.
According to the Centers for Disease Control and Prevention (CDC), more than 100 million U.S. adults are now living with diabetes or prediabetes. These diseases, however, are preventable; 150 minutes of activity a week and a 5-7% weight loss can reduce the risk of developing Type 2 Diabetes by as much as 58%. But even though people know the advantages of diet and exercise, they have difficulty staying motivated, even when their health is at stake. According to the CDC, only 22.9% of U.S. adults comply with these recommendations. Our aim is to encourage patients with chronic diseases to reach their health goals in a way that seamlessly embeds healthier life habits with their schedule and way of life. This required a solution centred around AI, user experience, patient education, and social and gamification elements. To truly change human behaviour, it has to communicate the most hyper-personalised recommendations and goals possible, ensuring the time between doctor appointments is spent combating the epidemic before it starts.
How personalised does the app become? Does it rely on the user to inform the app about their everyday activities from exercise to work?
Sweetch is recommended to patients by a partnering health system or insurer. The platform combines patient’s medical history with users’ input and information that is automatically collected from the patient’s smartphone, wearables, and a Bluetooth-connected scale which is part of the Sweetch suite.
We wanted to build a product that would fit seamlessly into the lives of our target users, and we understood that building trust in the app would be crucial to user retention and disease prevention success rates. Therefore, we included the AI to go above and beyond the one-size-fits-all model found in most disease prevention platforms. Sweetch’s Behavioural Change Engine uses advanced machine learning algorithms to learn a user’s daily habits, generating continuously optimised goals and recommendations that reflect the context, timing, location, and tone most likely to lead to action. For example, a user will never be prompted to take a 20-minute walk if it is raining outside, or if they are in a scheduled meeting at work. The solution generates messages that are most likely to trigger action based on the user’s past behaviour.
Is this the kind of facility that would be wise to use in cooperation with healthcare professionals?
Definitely. The Sweetch platform is “prescribed” by a healthcare provider or the patient’s insurer (or employer, in cases of self-insured companies). Currently one in seven US healthcare dollars is spent treating diabetes and its complications, so health systems and insurers are in urgent need of relieving this financial strain. Sweetch addresses this by predicting who is most likely to develop diabetes and then preventing it. As a result, Sweetch enhances the health provider’s ability to detect patients with the highest risk of developing diabetes and allocate intervention resources more effectively. This results in significantly reduced diabetes and metabolic syndrome-related expenditures. The solution is already in use by healthcare professionals in the US, Asia-Pacific and the UK.
Give us an idea of what goes into the Sweetch prediction platform?
Sweetch’s proprietary machine learning prediction platform utilises patient history and medical data (i.e. blood tests, medications, claims, etc.) to quantify patients’ personal risk of developing diabetes or other metabolic syndrome-related diseases in a defined time frame. Our proprietary machine learning-based diabetes prediction platform calculates patient risk, while the AI-powered Behavioural Change Engine monitors and analyses personal, environmental and behavioural digital biomarkers through a user’s smartphone, automatically learning a user’s routine and life habits, as mentioned previously. Unlike existing risk calculators, which present a vague risk estimation, i.e., low/medium/high in 7-10 years, Sweetch’s prediction platform output is quantified for the short term. For example, Sweetch might predict that a certain patient will have a 79% chance of developing diabetes in the next year, and another patient will have an 84% chance of developing diabetes in the next two years. Sweetch’s diabetes prediction algorithm yields 5-7x more accurate risk predictions than current gold standard prediction methodologies. As a fully automated digital platform, our risk prediction tool is also scalable – meaning that it can assess large patient groups to quantify patients’ risk, enabling health systems to evaluate and monitor their entire patient population.
How much of Sweetch’s development was a result of consultation with healthcare professionals?
As a practicing family physician, Dr. Yossi Bahagon has decades of experience treating patients. He also led the nationwide digitisation of Israel’s largest HMO (the second largest in the world) and has over 15 years of leadership experience in biomedical informatics and health information sciences. His extensive and in-depth knowledge of medicine, digital health, and patient-centred care significantly informed the development of Sweetch.
What kind of results have you seen from the users of Sweetch/when developing Sweetch?
Sweetch is clinically proven to help prediabetes and other chronic disease patients achieve a healthier and more active life through data-driven recommendations. In a clinical trial led by Johns Hopkins’ Endocrinology, Diabetes and Metabolism Division, Sweetch was found to have a retention rate of 86% and to achieve clinically significant outcomes on all measures evaluated: Sweetch reduced Haemoglobin A1C (HbA1C) levels – a diabetes biomarker; generated weight loss; and increased physical activity. Real-world data has shown a retention rate of 82% at six months of usage (and counting).
What do you think Amazon and Apple’s entry into healthcare/fitness/wearable markets means for the way we keep track of our health going forward?
Powerful giants like Amazon and Apple entering into these markets signals that digital health is here to stay.
According to a StartupHealth report, 2018 was the biggest funding quarter for digital health ever, with $14.6 billion invested in 765 companies, surpassing 2017 numbers by almost $3 billion, with patient empowerment the most active domain. But as people across the world use smartphones to keep track of their wellbeing, their mounting data is often left unutilised and abandonment rates are high.
At Sweetch, we believe that as we use technology to keep track of our health moving forward, health system involvement is crucial. We need more involvement from doctors and medical teams for higher chances of long-term success and improved consumer health apps, wearables, and digital coaches. When doctors utilise data collected by these digital tools, they can provide more optimally informed treatment and improved prevention capabilities. This will have a positive effect on patient satisfaction, accountability, motivation, and success.
What are the future plans for Sweetch? Are there any plans for it to cover people with other conditions aside from metabolic syndrome diseases?
Definitely. We built Sweetch to be hyper-personalised, scalable, cost-effective, and customisable to different use cases. Sweetch is seamlessly applicable to other disease verticals where exercise, weight loss, diet, and nutrition can have a positive impact on disease outcomes and prevention.
We are expanding to further patient populations and health systems in the field of oncology. The robustness and extent of evidence supporting physical activity for improving clinical outcomes in oncology has led to the birth of a new clinical and research domain called Exercise Oncology, which we are excited to partake in.