With medical expenditures for diabetes reaching $176 billion, it is one of the most costly conditions in the U.S. Due to increasing obesity rates as well as issues surrounding lifestyle choices, the condition has become more and more prevalent as the number of American diabetics saw a fourfold increase from 1998 to 2014. About 22 million people in the country had diabetes as of 2014, compared with just 15.2 million in 2004 (an increase of about 30%), and 75,578 Americans died from the disease in 2013, the Centers for Disease Control and Prevention reported.
The costs associated with diabetes are alarming not just because they are high, but also because providing care for diabetics has been getting more expensive in a relatively short amount of time. In 2014, spending for diabetics who are covered by employer-sponsored insurance increased at a faster pace than spending for non-diabetics, totaling $16,021 per capita, according to a June study from the Health Care Cost Institute (HCCI).
From routine blood glucose testing and monitoring to improved eating habits, diabetes management requires many elements of care that make it too complicated for some to effectively address. However, health IT has opened new opportunities to substantially reduce diabetes-related spending and improve health outcomes.
Patients’ health data, especially for people with a condition like diabetes, can be found in a wide variety of places like labs and medical records. The use of big data in cognitive computing has the potential of connecting different data points to ultimately provide insights, including predictive analytics, that providers, researchers, and patients need for diabetes management.
IBM Watson Health partnered with the American Diabetes Association earlier this year to apply cognitive computing to the association’s data on diabetes. Its cognitive system can translate data into tailored and personalized insights that reflect individual risk factors, treatment regimens, and behaviors. Nudging patients according to their unique means and demands has tremendous value, Dr. Kyu Rhee, IBM Watson Health’s chief health officer tells Healthcare Dive.
Research findings have shown that diabetics can cost a healthcare system twice as much as non-diabetics or more. In 2012, the U.S. spent an estimated total of $245 billion on diabetics, which accounted for more than $1 in $5 dollars in the country’s healthcare system.
“If you can prevent diabetes, then you’re doing the right thing but you should be able to also reduce the costs associated with diabetes,” Rhee says. The idea behind using predictive analytics through cognitive technology is connecting what Rhee calls “the four v’s of data”: Volume, variety, velocity, and veracity.
While collaborating with Medtronic, Watson Health had access to de-identified data for 10,000 patients with type 1 diabetes and was able to predict with up to 81% accuracy a hypoglycemic episode four hours before it occurred.
Not too far from IBM's and the ADA's combined efforts, Healthy Interactions partnered with MedCurrent in June to introduce a platform that merges clinical decision support with patient engagement technology. The companies aim to enact behavior change among diabetics through technology and sharing accountability. Improving patients’ self-management of their condition can help drive harder outcomes like better exercise behavior.
Healthy Interactions has more than 30 million users in 122 countries demonstrating that conversations facilitated by a coach or facilitator through its Conversation Maps can help the patient experience, according to Chief Medical Officer Dr. David Moen. The platform also allows patients to set their own goals and share them with whomever they choose.
Physician ordering behavior
The other key lever of the MedCurrent-Healthy Interactions partnership is addressing physicians’ ordering behavior. Moen says there’s a lot of delay from the time when patients start using insulin to when it was appropriate to actually start using the medicine. That is partly because physicians sometimes don’t initiate the conversation early enough. “So the clinical decision support is a systemic way to look at a population and then begin deploying a tool within the electronic health records, which helps us change physician ordering behavior,” Moen explains.
The MedCurrent-Healthy Interactions system knows if a doctor is considering adding insulin to a patient’s treatment, presents choices and makes recommendations with regards to helping that individual patient.
Some have been calling diabetes an invisible disease for years, including Rhee. But organizations have figured out new ways to optimize its treatment thanks to health IT and other solutions may still be in store. “Through the power of cognitive insights, by translating big data into these insights, we are going to bring visibility to this disease,” Rhee says.