Novel mobile app provides personalized weight loss solution
For patients across the globe, obesity heightens the risk factors of potentially fatal conditions like stroke, diabetes, heart disease, and certain types of cancer. As the incidence of obesity increases, new digital tools for weight management have become available. But many nutrition and weight tracking apps don’t include foods from a variety of cultures and dietary restrictions. Moreover, access to high-speed internet is not equitable across developed and developing countries. Mobile phones, however, are ubiquitous.
According to a report published by the World Advertising Research Center, it is estimated that three-quarters of users will access the internet solely via smartphone by 2025, representing nearly 3.7 billion people. And in emerging economies where internet infrastructure is not yet robust, mobile phone ownership is growing – broadening access to the internet and to Internet of Things devices.
Associate Professor Valencia Koomson of the Tufts University Department of Electrical and Computer Engineering and colleagues from the University of Ghana’s Department of Computer Engineering and Department of Nutrition and Food Science had that knowledge in mind when they put their heads together on a new project. The team developed an artificial intelligence (AI) based mobile application driven by a genetic algorithm. Users could wield the app as a new tool to track their energy balance – the calories taken in through eating, compared against the calories burned during physical activity – and to plan meals, providing a personalized weight loss solution for people with diabetes and hypertension.
In research published in the International Journal of Telemedicine and Applications, the team detailed its methods and findings. Thirty volunteers at the University of Ghana tested the model, inputting information on their meals, levels of physical activity, diabetes status, cholesterol levels, gender, age, weight, and height. The algorithm took that data into account along with information on foods drawn from a database containing locally-known Ghanaian foods. The app then provided tailored recommendations for meals. “The app tracks macronutrients (carbohydrate, protein, and fat) in order to help users track the quality of foods, rather than just counting calories,” says Koomson. “A 200-calorie snack consisting of apple slices and almond butter is quite different from a 200-calorie bag of potato chips, for example.”
The researchers found that the system was able to track a number of factors, including daily nutritional needs and calorie intake, recommended meals to eat to maintain health, and a user’s weight loss progress over time. The model reviewed the user’s condition and recommended foods based on that condition and on macro- and micronutrient requirements – ensuring that the user received the necessary nutrients and met their daily calorie needs, while avoiding compromising potential health concerns like diabetes or their cholesterol.
Between the foods considered and its personalized predictions, the team’s app stands out. Many apps that track diet and macros offer large databases of foods but focus solely on Western diets. Koomson and the University of Ghana team ensured that their database moved away from a Western-centric model and featured staples of Ghanaian cuisine like okro stew, fried fish, and waakye. “Popular apps do not have a personalized prediction feature that offers the user suggestions for meal planning, based on local foods, in order to stay within macronutrient limits,” says Koomson.
Going forward, the AI-based application could be a useful tool for patients and health management personnel managing obesity-related medical treatment and weight loss progress.
Department:Electrical and Computer Engineering