While the world tries to decide whether artificial intelligence is here to help us or hurt us, AI is quietly infiltrating our daily lives — from streaming recommendations to image recognition. And in health technology, AI is making a real difference to people across the world, saving lives in a multitude of ways.
Today Sweetch — a mobile health app that helps prevent diabetes and improve outcomes for people with diabetes by encouraging long-term behavioral change — has revealed the outcomes of its clinical trial conducted at Johns Hopkins University.
Directed by the university’s division of endocrinology, diabetes, and metabolism, the study shows that using Sweetch significantly lowered A1C levels — a diabetes biomarker for blood sugar. The app has been shown to increase physical activity and reduced weight for patients with early stage diabetes.
“Sweetch is a fully automated, artificial intelligence-based personal digital intervention program that helps individuals lose weight and become more active, with the goal of reducing their long-term risk of diabetes and other metabolic syndrome-related diseases,” Dana Chanan, CEO and cofounder at Sweetch, told me. “Sweetch’s core philosophy is that each individual has his or her own life habits, motivations, and behavioral change progress. Therefore, generic recommendations to walk 10,000 steps or be active for 30 minutes a day and eat fewer carbohydrates may not produce sustainable and meaningful behavioral change for all individuals, especially in the long run.”
That’s where AI comes in. Sweetch is able to personalize its recommendations to the individual, which means that the user is more likely to get the support and encouragement they need to see a material difference.
“Sweetch artificial intelligence automatically translates raw data streams originating from the patient’s mobile phone and Bluetooth digital scale into insights about the individual’s life habits, schedule, activity patterns, driving and walking routes, surroundings, and more,” Chanan said. “Completely free of human involvement, Sweetch presents personalized, contextual, just-in-time, just-in-place recommendations that guide him or her toward achieving recommended activity, weight reduction, and diet goals in a way that fits the user’s real-world life habits. Sweetch’s technology learns what types of message result in better compliance for the specific user at a specific context (e.g. day of the week, time, location, effect of consecutive messages of different types, etc.).”
The clinical trial found that being active for 150 minutes a week and reducing weight by 5 to 7 percent could reduce the chances of developing diabetes by 58 percent in its subjects. The three-month trial contained 55 pre-diabetic adults at different levels of obesity. Over the course of the study, Sweetch significantly changed participants’ behaviors, with retention rates as high as 86 percent. On average, participants achieved significant increases in physical activity, lost an average of 1.6kg, reduced waist circumference by 1.4cm, and had a clinically meaningful reduction in A1C of 0.1 percent. Based on previous studies, each 1kg weight loss translates into a 16 percent reduction in diabetes risk.
All the expenses for resources and time for the clinical trial are paid for by Sweetch. This includes the time and resources from John Hopkins university all the expenses related to the clinical trial participants which include the costs of blood work. The Sweetch app was given to the users at no cost.
So what is next for Sweetch, following the clinical trial?
“The company concentrates its efforts on business development and marketing, looking for self-insures, health providers, and payers that are interested in helping their employees, patients, and members reduce their long-term risk of diabetes and other metabolic syndrome-related diseases,” Chanan said. “In addition, Sweetch is looking for strategic partners that can help the company in its market penetration.”
The audio problem: Learn how new cloud-based API solutions are solving imperfect, frustrating audio in video conferences. Access here