Diabetes is one of the most common disorders in the world, with over 100 million cases diagnosed worldwide. And while recent technological advances have made treating it easier than before, it remains far from a walk in the park. People living with diabetes have to make about 180 decisions about food intake, insulin, sleep, and physical activity each day to keep their blood sugar levels in check. And in the severest of cases, a seemingly harmless slip-up, like skipping lunch or having an extra cup of coffee, can cause hypoglycemia, a dangerous dip in blood sugar that can lead to fainting — or death.
To ease the stress somewhat, IBM and medical device company Medtronic have teamed up to develop IQcast, a predictive tool built into Metronics’ Sugar.IQ app for diabetic patients who require multiple daily injections. By applying machine learning algorithms to readings from Medtronics’ Guardian Connect continuous glucose monitoring system, IQcast can predict the likelihood that a person will experience a low-glucose event within 1-4 hours and recommend proactive steps to reduce the chances of future dips.
“Simply put, IQcast acts like a weather forecast for people with diabetes so they can better prepare for their day,” Dr. Robert Vigersky, senior director of medical and clinical affairs at Medtronics’ Diabetes Group, said. “By predicting the likelihood of a low glucose event … [it] helps people with diabetes better prepare for their day so they can live their life with greater freedom and better health.”
When IQcast detects that a user is about to enter a hypoglycemic state, it communicates through Sugar.IQ the risk severity: low, medium, or high. The accuracy of its predictions improves as the blood sugar dip becomes more imminent, according to IBM.
“Hypoglycemia, or ‘going low’, is one of the most acute and frightening events that a person living with diabetes can experience,” Dr. Lisa Latts, deputy chief health officer at IBM Watson Health, said. “Fueled by the right data, AI and machine learning can play a powerful role in helping to alleviate the burden of diabetes and the worry of a hypoglycemic event, and we’ve built the new IQcast features with this goal in mind.”
It’s not the first time machine learning has been tapped to make predictions about diabetes, to be clear.
Google recently launched an AI program in Thailand to screen for diabetic eye disease. Beijing-based 4 Paradigm created a model that’s 88 percent accurate at detecting the likelihood a patient will develop diabetes within 15 years. Klick Health developed algorithms to predict blood glucose levels 30 minutes into the future. And One Drop, a biomedical company that produces blood glucose monitoring systems compatible with the Apple Watch, recently introduced a feature that forecasts glucose values over time and provides behavioral recommendations.
But beyond tools like IQcast, Latts says that IBM is investigating ways to improve diabetes care with artificial intelligence and predictive algorithms. In a study conducted by IBM’s Watson Health division, for example, researchers recently explored ways machine learning can be applied to medical data for the purpose of identifying patients at risk of complications from type 2 diabetes. And building on work it undertook a year ago, IBM recently partnered with JDRF, a global nonprofit organization funding type 1 diabetes studies, to investigate novel type 1 diabetes forecasting techniques.
More broadly, IBM is teaming up with academic institutions to advance AI health care applications. Through IBM’s AI Horizons Network, which aims to pair faculty and graduate students with IBM researchers to “accelerate” machine learning technologies, the company worked with the Rensselaer Polytechnic Institute (RPI) in Troy, New York to establish the Center for Health Empowerment by Analytics, Learning, and Semantics (HEALS).
“Technology alone cannot alleviate the burden of managing diabetes,” Latts said. “As we continue to work with our partners to find greater ways in which technology can help clinicians and those living with diabetes to manage their condition, we are confident that the ultimate best solution will be one in which patient engagement, professional care, technology, data and informatics all come together to optimize treatment.”