Technology Roundup: CGM Systems, Joslin Home Featured at 2020 ADA Scientific Sessions

Research presented at the American Diabetes Association's 80th Scientific Sessions offers new insights into the future of glycemic control, diabetes self-management, and remote care.

Data Highlight Successes of Abbott’s FreeStyle Libre System

Real-world data on Abbott’s FreeStyle Libre System, a continuous glucose monitoring (CGM) system, suggest its use in individuals with type 2 diabetes (T2D) is associated with significant reduction in glycated hemoglobin (A1C). Results from an observational, retrospective study, presented at the American Diabetes Association's Virtual Scientific Sessions, found the association exists in both individuals receiving long-acting insulin and those on non-insulin therapy, indicating use of FreeStyle Libre may prevent T2D from progressing to the point where insulin is necessary. A1C levels dropped by 0.8% after 6 months and by 0.6% after 1 year among study participants. “These new data underscore how our wearable technology provides actionable information to deliver positive outcomes in anyone with diabetes, emphasizing the power of the FreeStyle Libre system to change countless lives among the millions of people with diabetes,” said Mahmood Kazemi, MD, divisional vice president of global medical and scientific affairs and chief medical officer of Diabetes Care at Abbott.

Additional data released at the virtual conference show use of Abbott’s CGM technology reduces all-cause diabetic ketoacidosis-related hospitalization rates and acute diabetes events. As diabetic ketoacidosis (DKA) often leads to emergency room visits, the results point to potential cost savings for individuals with diabetes. A retrospective cohort analysis of 75,000 individuals in France found that use of the FreeStyle Libre system resulted in a 52% drop in annual DKA rates in those with type 1 diabetes (T1D) and a 47% reduction in those with T2D.

On June 15, the FDA cleared Abbott’s FreeStyle Libre 2 Integrated CGM (iCGM) in the United States for both adults and children above the age of 4 with diabetes. Using Bluetooth technology, the system continuously transmits glucose data every minute. Patients can also receive optional real-time alerts indicating glucose is too high or low, without scanning. The device is currently the only iCGM that can sustain performance for up to 14 days. The system is designed for use alongside a mobile app that Abbott is working to bring to market in the United States.

Is CGM Valuable for Managing T2D?

The system will be available at participating pharmacies and durable medical equipment suppliers in the coming weeks and will be priced the same as the FreeStyle Libre 14-day system, approved in July of 2018. Recently released data found that on a per person, per year basis, annual costs of using the FreeStyle Libre 14-day system among commercially insured individuals with T1D and T2D, were 61% and 63% lower compared with traditional blood glucose monitoring (fingerstick testing), respectively. Use of the system is estimated to save around 50% in average costs resulting from severe hypoglycemia, which can lead to hospitalizations and emergency room visits.At the conference, experts held a debate as to whether CGM is even a valuable management tool for individuals with T2D. Arguing in favor of the technology, Athena Philis-Tsimikas, MD, director of community engagement at Scripps Research Translational Institute and corporate vice president of Scripps Diabetes Institute, laid bare the behavioral and clinical improvements of CGM exemplified by numerous studies. In 2017, the United States spent $850 billion on treatment of diabetes, while most costs resulted from poor control of the disease and complications. Impactful and earlier interventions can help reduce projected costs in the future, she argued. Furthermore, CGM offers a remote, digital solution to diabetes care management while encouraging individualized feedback with behavioral modification incentives. Philis-Tsimakas cited one study which found participants using CGM were able to achieve greater glycemic control without intensification of pharmacotherapy.

In contrast, Elbert Huang, MD, a professor of medicine at the University of Chicago, countered that broad application of CGM may not be as cost efficient as other interventional methods. Pointing out the United States has the highest per capita healthcare spending in the world, Huang argued for better understanding of CGM via longitudinal studies before wholly embracing the technology for use in every patient. Controlling healthcare spending should be at the forefront when it comes to medical innovation and improving care. “We’re not done all the research needed both clinically and from a health economics perspective to really understand the full impact of CGM,” Huang said. He followed up by positing whether CGM actually implements behavioral change or if it merely facilitates individuals who have already decided to better control their diabetes.

Medtronic Releases Data on Hybrid Closed Loop Systems

Both experts noted intermittent use of CGM may be a cost-effective solution to the issue while future research is still needed to better understand the long-term economic and health outcomes of the technology.Medtronic released positive results from their investigational trial of MiniMed 780G Advanced Hybrid Closed Loop (AHCL) system, an automatic insulin delivery (AID) system. The system, which has a default target of 100mg/dL, programmable insulin action time within 2 to 8 hours, and issues automatic corrections every 5 minutes, met all study endpoints. Results from the 90-day at home United States pivotal trial revealed an average A1C of 7% among participants (aged 14-75), no severe hypoglycemia and DKA, and mean sensor glucose of 148 mg/dL overall, and 144 mg/dL at the default 100 mg/dL target. In addition, 96% of users indicated the system was easy to use, while integration of AHCL reduced system requests for fingerstick blood sugars by 46%, compared with the MiniMed 670G system. Without increasing hypoglycemia, lower target glucose and active insulin time settings were found to substantially improve Time in Range (TIR).

Joslin Diabetes Center Introduces Digital-Health Model

Are Machine-Learning Algorithms Reliable for Predicting T2D?

Data from an additional study revealed use of Medtronic’s AHCL in adolescents and young adults with T1D can lead to improved daytime blood sugar. The FLAIR study compared the efficacy of the new AHCL with a currently approved AID system in the United States (Medtronic’s 670G Hybrid Closed-Loop [HCL]). The study found the percentage of TIR over 24 hours was superior in those who used AHCL vs. HCL. At baseline, TIR was recorded as 57% and improved to 67% with AHCL use and 63% with HCL use. User satisfaction surveys also showed patients preferred AHCL over HCL. “This age group has traditionally been the most difficult group in which to optimize glucose management,” said Richard Bergenstal, MD, who presented the findings. “The FLAIR study shows that individuals using any type of therapy, even insulin injections without a pump or CGM system, can benefit from the next generation AHCL AID therapy.”Joslin Diabetes Center unveiled its new integrated digital-health model of diabetes care, Joslin Home. Citing limitations and pitfalls of the current clinic-based diabetes care model, Osama Hamdy, MD, PhD, introduced the internet-based model as an alternative. The model incorporates 5 pillars intended to support patient self-management: shorter visits conducted via telehealth, more frequent visits, direct 2-way electronic scheduling, easy one-step billing, and brief, focused electronic health record documentation. In a 6-month pilot study of 17 patients, researchers found the Home Model increased efficiency of visits and resulted in a significant drop in A1C (1.2%). Patients also reported high satisfaction with the program, which allows them to select a multidisciplinary care team from a wide range of experts. Researchers conclude the Joslin Home Model is “potentially scalable for remote and underserved areas where multidisciplinary approach is lacking,” and can aid high-risk patients who require frequent encounters with a diabetes care team.1A meta-analysis of the predictive ability of incident type 2 diabetes (T2D) using machine learning algorithms found existing models “are not yet satisfactory.” Researchers conducted electronic literature searches to find longitudinal studies including models constructed by machine learning algorithms. The studies’ reference standards were a blood test or physician diagnosis of T2D. “Pooled sensitivity and specificity (95% confidence interval [CI]) of 8 eligible studies were 0.70 (0.66-0.74) and 0.78 (0.69-0.85), respectively.” In addition, positive likelihood ratio (PLR) and negative likelihood ratio (NLR) were 3.25 (95% CI, 2.25-4.58) and 0.38 (95% CI, 0.34-0.43), respectively. Because PLR>5 moderately increases the probability of a disease given a positive test and NLR<0.2 moderately decreases the probability of a disease given a negative test, researchers concluded current machine learning algorithms are not sufficiently accurate to predict T2D incidence.2


  1. Hamdy O, Mitri J, Barrett A, et al. Joslin Home: an innovative integrated digital-health model of diabetes care in the technology era. Presented at: American Diabetes Association 80th Scientific Sessions; June 12-16, 2020; Abstract 18-OR.
  2. Sato T, Yamamoto M, Ishiguro H, et al. Predictive ability of incident type 2 diabetes mellitus (T2DM) using machine learning algorithms: a meta-analysis. Presented at: American Diabetes Association 80th Scientific Sessions; June 12-16, 2020; Abstract 841-P.