While the Cancer Moonshot and Precision Medicine Initiative aim for, well, the moon in the world of Big Data, perhaps no chronic disease today is more reliant on data — big or — small – than diabetes.
“Diabetes is hugely data-driven,” said Dr. Jennifer Shine Dyer, a pediatric endocrinologist and app developer in Columbus, Ohio. Control of diabetes is largely dependent on managing blood-glucose levels. “It lends itself well to AI and apps,” Dyer said.
The world of data-driven diabetes care has been blown open in the last couple of years with continued advances in continuous glucose monitoring as well as new partnerships between unlikely entities. Then there is also the fact that future availability of an automated insulin delivery system which requires little patient input — the artificial pancreas — is a very real possibility.
In fact, Medtronic took the first step in that long road becoming the first to win Food and Drug Administration approval for a hybrid, closed-loop automated insulin delivery system that applies an algorithm to deliver varying amounts of insulin as needed to patients with Type 1 diabetes. Here’s an infographic created by JDRF (formerly known as the Juvenile Diabetes Research Foundation) that traces the development of the artificial pancreas.
No longer content to be a diabetes widget maker, the Irish medtech firm is also partnering with IBM Watson Health to create a cognitive app leveraging data on Medtronic’s insulin pumps and Big Data to predict blood glucose trends for patients. The goal? Leverage patterns in data to provide better glycemic control.
For its part, Dexcom has made headway in providing that glycemic control for patients through its continuous glucose monitoring system. FDA first approved a real-time CGM in 2005, according to JDRF.
A Deep-dive Into Specialty Pharma
A specialty drug is a class of prescription medications used to treat complex, chronic or rare medical conditions. Although this classification was originally intended to define the treatment of rare, also termed “orphan” diseases, affecting fewer than 200,000 people in the US, more recently, specialty drugs have emerged as the cornerstone of treatment for chronic and complex diseases such as cancer, autoimmune conditions, diabetes, hepatitis C, and HIV/AIDS.
“CGM creates a continuous stream of glucose data that didn’t exist before,” said Annika Jimenez, senior vice president for data at Dexcom.
The San Diego company made waves in early 2015 when it gained FDA clearance for its mobile apps tied to the G4 Platinum CGM. One app connects the Dexcom device to the smartphones of as many as five designated caregivers or family members, while another goes on the smartphone of the patient.
This “follower” feature helps loved ones keep tabs on patients to avoid complications such as hypo- or hyperglycemia. “This is especially important to Type 1 diabetes,” Jimenez said.
For some families, Dexcom’s CGM and related apps have been a godsend.
At last week’s MedCity ENGAGE conference in San Diego, Daniel McCaffrey, director of digital health, data and analytics at Dexcom, showed a video highlighting how data flows from a CGM device to a smartphone.
“This is a huge advancement in terms of care, management, and data,” McCaffrey explained. “Continuous glucose monitoring can provide a more accurate picture of a patient’s health and state of diabetes control, plus it’s much easier and convenient than the older method of self-monitoring blood glucose (SMBG) data collection.”
A typical Type 1 diabetes patient checks blood sugar multiple times a day through fingerstick testing. A continuous glucose monitor takes readings every five minutes.
“You only need to check your blood sugar 2-3 times a day to calibrate the monitor,” Dyer noted.
She added that the time savings and data advantage using CGMs is similar for those with Type 2, as are the outcomes.
CGM systems become more powerful when connected to other devices and tied to management algorithms. The Dexcom G4 series of monitors had been around for several years, but the FDA approval of the apps represented a sea change because the G4 Platinum facilitated connections to other devices.
“Things really took off when we brought Bluetooth into our continuous glucose monitor,” Jimenez said.
Dexcom has created an ecosystem for others to design products supporting the G4. The CGM now pairs with insulin pumps, clinician dashboards and patient nutrition and activity trackers from third parties. The ecosystem should grow exponentially because Dexcom is about to release an application programming interface, McCaffrey said at MedCity ENGAGE.
While Medtronic and Dexcom have been trying to make life easy for Type 1 diabetes patients, Omada Health, a San Francisco-based digital health company that aims to help people with pre-diabetes from developing full-blown Type 2 diabetes.
That company also believes in connectivity, though from a bit of a different angle.
Omada ships weight scales to participants in its remote diabetes management programs. The devices have 3G wireless connectivity built in, so there is no need to configure the scales to work on Wi-Fi networks or pair with phones or tablets by Bluetooth.
The scales are important because that company treats weight as a proxy for diabetes control.
“Weight gain is a strong sign of diabetes risk, said Eric Williams, director of data science at Omada.
Weight is easy to capture, which is important because people with pre-diabetes often do not collect hemoglobin A1C values more than once a month, according to Adam Brickman, Omada Health’s director of strategic communications and public policy.
However, weight is not the only measure, of course, and those in diabetes prevention programs should capture additional data points, Williams said. It is Omada’s role to help make sense of readings from scales, fitness trackers, patient-reported food intake and even communications on group message boards. (Brickman said at ENGAGE that said physicians often feel like they’re in a “data flood but a data desert.” The data has to inform lifestyle and clinical decisions, not complicate treatment, he said.)
“We really use data to empower health coaches,” Williams said. For example, an algorithm takes signals from the patient’s weight baseline, the constant data stream from a CGM, fitness trackers, message boards and food intake to develop and execute a treatment plan.
“It knows what is likely to happen to a participant in terms of weight gain even before” a change in diabetes risk or appearance of symptoms, Williams said of the Omada algorithm. Constant tracking of data and refining of the algorithm leads to better interventions and, ultimately, better outcomes, he said.
There still is a good amount of trial and error involved, however.
Omada originally gave every participant a pedometer and a goal of 10,000 steps a day. “We got a lot of feedback from participants, mostly about [the goal] being too hard or too easy,” Williams reported.
“We realized pretty quickly that it varied by demographics,” he continued. So Omada segmented patients by body-mass index, age, gender and similar metrics, and then developed more personalized goals.
Brickman reported that Omada has enrolled 75,000 people in diabetes prevention programs. Together, this cohort has conducted 10.5 million weigh-ins and generated more than 1 billion data points. “We believe we now have the largest collection of data on behavior change in human history,” he said.
Three to four weeks into the program for a given patient, Omada can now predict with 80 percent accuracy how an individual will do after 16 weeks, he said. “It allows us to empower our human coaches with data,” Brickman said. “We can start personalizing the program as soon as you come in.”
Personalization and prediction are where diabetes care certainly is heading.
Another diabetes-focused digital health company, Glooko, has been working with Joslin Diabetes Center in Boston on technology that would provide advance warning of hypoglycemic attacks so that patients with Type 1 diabetes can take a glucagon injection. Glooko aggregates patient data from wearables and apps for to collect information on meals, exercise, and sleep.
IBM Watson Health is teaming with Medtronic on similar technology.
“It’s a battle of math,” Dyer, the pediatric endocrinologist, said of diabetes control. She focuses on Type 1, so she is excited about the advent of the artificial pancreas, which, she said, “promises to flatten variations” in blood-glucose and insulin levels.
The artificial pancreas is not a panacea, though, since the newly approved Medtronic MiniMed 670G still requires people to input carbohydrate intake. “There’s a concern that the artificial pancreas is seen as a cure,” Dyer said. “You still have to do stuff. It’s not there yet.”
Indeed, there is much work to do in preventing and caring for a disease that affects an estimated 29 million Americans — plus another 86 million with pre-diabetes.
“We are just beginning to see the potential of data in diabetes care,” said Frank Westermann, CEO and cofounder of startup mySugr.
Photo: Flickr user Steven Depolo
