As technology to treat type 1 diabetes becomes ever more sophisticated, the amount of data available to make blood sugar management decisions grows at an exponential rate. This tsunami of information is only valuable if it can be easily accessed and interpreted by clinicians and patients.
When confronted with such a volume of data, people with type 1 diabetes often feel they can only make significant adjustments in their blood sugar management at the clinician’s office. At these all-too-brief appointments, health care providers often struggle to download and interpret all the pump and CGM data in a way to make actionable and helpful recommendations for patients.
The Israeli tech company DreaMed Diabetes aims to ensure that diabetes device data can help people with type 1 diabetes and their health care providers make decisions on blood sugar management. The project, spun out of the Petch-Tikva (just near Tel Aviv) Schneider Children’s Medical Center of Israel, began in 2006, when Dr. Moshe Phillip, director of the hospital’s Institute for Endocrinology and Diabetes, recruited engineer Eran Atlas to join himself and endocrinologist Dr. Revital Nimri to create the hospital’s new integrated medical technology center.
Streamlining and enhancing the doctor-patient relationship
Dr. Phillip and Dr. Nimri spoke with T1D Exchange at the 79th Scientific Sessions of the ADA in San Francisco. T1D Exchange and DreaMed announced last week they are partnering to provide the T1D Exchange Quality Improvement Collaborative access to DreaMed Advisor Pro, an artificial intelligence (AI)-based decision support technology. The partnership will leverage DreaMed’s technology to engage with high-risk patients and improve their therapies, helping them achieve better outcomes. It also marks a key milestone as T1D Exchange forges its first international partnership.
In the interview, they said the idea behind DreaMed is to streamline and simplify the process of adjusting basal rates, bolus decisions, and correction factors. DreaMed works with the Glooko, diabetes management data platform to gather and log data from a wide range of meters, pumps, and other technology into a repository of information, which can then be analyzed with big data techniques.
Recommendations can be made after the input of just 12 to 21 days of data. If, for example, the system sees a pattern of high blood sugars after breakfast paired with hypoglycemia in the afternoons over the course of 12 to 21 days, it might recommend to increase the carb ratio for a narrow band of time in the morning, then decrease it after lunch.
With DreaMed Advisor Pro, Dr. Phillip said, the health care provider still is in the driver seat when it comes to recommendations.
“We made sure the physicians will never feel as if they’re losing their autonomy,” said Dr. Phillip. “In the present version, the physician is getting the interpretation of the data – it’s specific advice like changing the basal by small amount – but still the physician can make changes and he or she has to push the button to send it to the patient.”
Building a first-rate medical research facility
It is perhaps no surprise that DreaMed was borne out of Schneider Children’s Medical Center. Since the center is the only fully independent pediatric hospital in Israel, it recruited talented and specialized staff while focusing on projects that could bring the greatest benefit to their patient population – including 1900 children and young adults living with type 1 diabetes. The result has been impressive, with innovation and research being conducted in a wide variety of fields including artificial pancreas technology, big data analytics, and more.
Originally, the engineers lived within the institute as they conducted research and worked on projects. As their numbers grew, they moved outside the hospital to an attached facility and expanded their operations. DreaMed was spun off into a separate firm when the organization sought out a CE mark for its technology from EU health regulators
The young company celebrated several key milestones early on, including the licensing of artificial pancreas technology to Medtronic in 2015 and being awarded a $3.4 million grant from the Helmsley Charitable Trust in 2016 to develop an algorithm that mimics clinical interpretation of type 1 diabetes device data.
“We have named (the algorithm) MD Logic – Medical Doctor Logic,” said Dr. Phillip. “We actually translate the way our physicians are thinking into an algorithm. We are trying to make decisions in a split-second, without spending a lot of time.”
Clearance from FDA was done in a de-novo process, creating new product code for using algorithms to retrospective analyze CGM data and provide recommendations on insulin therapy. This regulatory change likely will mean that DreaMed’s algorithms will be in the news in the coming years for future innovative decision support algorithms that provide solutions also for people with diabetes under multiple daily injections therapy.