Big Device Companies See Opportunity for AI in Healthcare

Brian Buntz

January 11, 2016

6 Min Read
MDDI logo in a gray background | MDDI

Major medical device companies are getting serious about the potential of artificial intelligence technology to improve patient care. Here's Medtronic's vision for using Big Data and IBM's Watson to help diabetics.

Brian Buntz

There is a growing trend of big device companies announcing partnerships with traditional tech companies. Take for instance Medtronic's recent presentation at CES, where it showed off its plans to develop diabetes apps powered by IBM's Watson. Last year, IBM announced its own plans to further enter into the healthcare through its $700-million acquisition of Merge Healthcare, a company that keeps track of medical images. In a similar vein, Johnson & Johnson announced last year that it is working on developing surgical robots that make use of Google's computing technology.

To learn more about this trend, Qmed reached out to Annette Brüls, president of Medtronic's global Diabetes Service & Solutions department. In the following Q&A, Brüls comments on the promise of Big Data to foster consumer-centric healthcare and sheds light on how the company plans on using IBM's computing prowess to help synthesize information from insulin pumps, continuous glucose monitoring devices, and potentially, artificial pancreas systems.
 

Qmed: Medtronic was prominently featured at CES for the first time ever in 2016. How does Medtronic plan to further take inspiration from the consumer technology sector this year?

Annette Brüls: Two of the key trends that we've seen over the last few years both across healthcare generally and within the diabetes community is the consumerization of healthcare and the rise of Big Data - essentially, consumers becoming the driver of their own health and wellness experience while companies are leveraging the power of data to provide real-time insights so people can make better healthcare decisions.

People with diabetes and their loved ones want convenient access to data from their diabetes devices anywhere and at any time to efficiently manage their condition. With the explosion of health-oriented mobile wearable devices and apps, there is a significant opportunity for connected technologies.

As we continue working with IBM Watson in the development of our app, we look to integrate additional information sources, such as wearable activity trackers, digital scales, geo-location data, calendar details and even the weather, to develop more valuable insights. We also intend to partner with other companies and incorporate their expertise in areas, such as nutrition. By combining new data sources and analyzing it in new ways, we hope to develop tools that will improve patients' ability to make better day-to-day and longer-term health decisions in the management of their diabetes.
 

Qmed: How did Medtronic's presence at CES this year come about? Is there a history in Medtronic co-presenting with IBM, or was CES the first time that the two shared a stage?

Brüls: Medtronic has participated at CES and even had speakers at CES in the past. This year, however Medtronic had an increased involvement.

Hooman Hakami, EVP Group President, Diabetes, Medtronic Diabetes, was invited to speak on the "Hyperdata and the Mobile Future" panel as a part of the Broadband conference track on January 5.

As a part of our partnership with IBM, CEO Omar Ishrak and I were both invited to participate in various IBM activities at CES. Omar joined Ginni Rometty, chairman, president and CEO of IBM, on stage as a part of her keynote address on January 6. I had the pleasure of joining William Evans, chief marketing officer of IBM Watson Health on stage in the Digital Health Summit Session: "The Dawn of the Cognitive Era" on January 7.
 

Qmed: Could you shed some light on Medtronic's relationship with IBM with respect to artificial intelligence used for helping diabetics?

Brüls: Medtronic insulin pumps and sensors generate a continuous stream of valuable health data. In the app being developed, we plan to have Watson synthesize information from Medtronic insulin pumps and continuous glucose monitoring devices--detailed information like the rate of insulin delivered, the constantly fluctuating glucose level and carbohydrate intake information.

As currently envisioned, Watson will identify both short-term and long-term characteristics of people groupings that have similar demographic, therapeutic and habitual profiles from the streaming data and apply machine-learned classification and prediction models to identify possible risk factors. Relevant insights would then be sent back to the customer's smartphone as a notification.

As our collaborative work continues, we expect to make more pattern discoveries that can be built into future additional predictive insights and especially combine our device information with other contextual information relevant to diabetes management, such as activity, meal intake, sleep patterns, etc. This will be incredibly useful for diabetes patients who have to manage so much information to make the right decisions. Our vision is to enable patients with actionable insights and personalized advice based on their own context, at the time a decision needs to be made.

IBM and Medtronic are also planning to develop a variety of other diabetes solutions that leverage Watson technology, such as solutions for healthcare systems and advancements in artificial pancreas technology.
 

Qmed: Can you summarize how this app could transform diabetics' lives?

Brüls: On a daily basis individuals with diabetes have to manage the balance between not having enough sugar in their blood and having too much. If their glucose level drops too low, they face the threat of hypoglycemia, which can cause confusion or disorientation and in its most severe forms loss of consciousness, coma, or even death. If it's too high over a long period of time, they risk cardiac disease, blindness, renal failure and amputation of fingers and limbs.

As an example, one of the important factors for managing blood sugar is to correctly estimate the impact of carbohydrate intake on sugar levels. But a person must also estimate the impact of physical activity.

The Medtronic app could, in the future, analyze these factors by integrating data from a wearable activity tracker, a picture of the meal eaten, and of course by continuously analyzing the data from a glucose sensor and insulin pump. Based on these factors, the Medtronic app could act as a personal assistant to help the users understand how to manage their glucose and carb consumption.

Although additional study is required, early research suggests that the cognitive technology may be able to predict low blood sugar patterns up to three hours before onset of a potential hypoglycemia - a low blood sugar emergency. This shows the potential for Watson to support diabetes management.
 

Qmed: How does Medtronic plan on using Watson technology in artificial pancreas technology?

Brüls: We plan to use Watson's machine learning capabilities to take our artificial pancreas algorithm to new heights. IBM's capabilities can help us and our partners develop closed-loop algorithms that can interpret more data, faster to provide a more personalized, seamless approach for each patient.

Like what you're reading? Subscribe to our daily e-newsletter.

Sign up for the QMED & MD+DI Daily newsletter.

You May Also Like