All types of diabetes are characterized by the body’s inability to control its blood sugar levels naturally. For type 1 diabetics, this inability stems from an immunological attack on the pancreatic cells responsible for insulin production. Left without an intrinsic source of insulin, type 1 diabetics are permanently dependent on external sources of insulin for blood sugar regulation – though, thankfully, this stuff can work just as well as the body’s own homebrew. But there’s a catch: natural insulin levels are highly dynamic and replicating them effectively requires attentive observation of blood sugar levels and responsive dosing throughout the day.

Unfortunately, blood sugar management often feels like a high-stakes balancing act. The frequent blood tests, stressful predictions around mealtimes, and logistics of having the necessary supplies on hand at all times can create quite a burden on patients, especially given the knowledge that less-than-optimal management can have serious health ramifications.

Inspired by the pancreas itself, researchers have spent decades attempting to emulate the organ’s normal function with technology in pursuit of a vision of radically simplified diabetic care. Recent research may represent the furthest development in that direction yet. Read on to learn more about this new technology, dubbed the “bionic pancreas”, and the many ways in which it improves upon its predecessors.

Current Options for Monitoring and Insulin Delivery

The finger prick method

At the simplest end of the sampling spectrum, patients monitor their glucose by pricking their fingers to draw a small blood sample and analyzing that sample with a portable electronic device. Patients are typically advised to test several times per day, including before meals and at bedtime.

Options for the finger-pricking method do exist that cover a range of ease-of-use, cost, sample-site flexibility, and connectivity with other devices. Still, many find this method tedious, painful, and inconvenient, and lament that the few samples per day leave large information gaps between samples.

CGM (Continuous Glucose Monitoring)

Continuous glucose monitoring (CGM) represents a sizeable improvement over the finger prick method. CGM systems are carried with the patient throughout the day and connected to the patient’s body by a small sensor on the skin. Regularly at short intervals, the CGM device will take a blood glucose reading and store that information either locally or by transmitting copies to a connected device like a smartphone (depending on the model).

CGM systems remove the pain of needles, and the inconvenience of manually sampling one’s blood, and increase the quantity of daily samples many times over, providing a much more complete picture of a patient’s blood sugar levels throughout the day. And, as we will explore, their ability to provide to-the-moment information has proven useful in the development of bionic pancreas systems.

MDI (Multiple Daily Injections)

The manual multiple daily injection (MDI) method is like the finger prick method of insulin delivery. It represents the lowest-tech option for taking one’s insulin, and simply consists of administering several injections to oneself throughout the day, using good old-fashioned needles and syringes. This method obviously leaves plenty to be desired around convenience, accuracy, and comfort.

Insulin pumps

As on the monitoring side of the equation, insulin delivery has seen technological improvements. Insulin pump therapy, sometimes called continuous subcutaneous insulin infusion (CSII) therapy, replaces manual injections, instead delivering a steady supply of insulin through a semi-permanent subcutaneous delivery port.

These battery-powered pumps must be configured to deliver the appropriate rate of background insulin and can be instructed to deliver additional insulin around mealtimes. Still, simple versions of the device only do as instructed, meaning that user feedback based on glucose monitoring is still required. More sophisticated models (called sensor-augmented pumps, or SAPs) use CGM data to modulate their outputs to a degree, though they still fall well short of full automation.

The Bionic Pancreas

Conceptual history

The limitations of existing technology have long since crystalized the goal of a fully functional artificial pancreas for type 1 diabetics. Such a system, if successfully implemented, would represent the gold standard of type 1 diabetes care, essentially replicating a healthy pancreas and virtually removing the role of user oversight altogether.

Early on, researchers developed a conceptual model for a closed-loop glucose management system, meaning one that measures glucose and delivers insulin without any user input. The device would simply require a CGM, an insulin delivery device, and some form of algorithmic connection between the two. The hard part would prove to be developing the existing technology for all three of those major components to the level of sophistication that reliable full automation would require.

Throughout the 2000s and early 2010s, significant advances toward full automation were made. The 2000s saw SAP technology connect a CGM to an insulin pump, which allowed the pump to respond to real-time blood glucose data. And, in 2016, the first FDA approval was granted to a hybrid closed-loop system, so categorized for its ability to operate without user input throughout the nighttime only, but still requiring prompting at mealtimes. These systems could, for example, suspend the administration of insulin where hypoglycemia was detected, limiting the severity of hypoglycemic episodes. However, systems to date have fallen short of the mark for true closed-loop technology, requiring some level of continuous input to properly function.

Recent developments

A recent trial put a new, truly closed-loop bionic pancreas system to the test. The results are groundbreaking and suggest that this new technology may represent the next great advancement in type 1 diabetes care.

The massive multi-site trial was conducted with 15 collaborating clinical sites across the country and co-ordinated by Dr. John Buse, Ph.D., at the University of North Carolina – Chapel Hill. Researchers studied a pool of 326 participants with type 1 diabetes, all of whom had been using insulin for at least a year prior to the study date. Participants were randomly split into two even groups – one of which would continue to receive their normal insulin therapy, and one which would switch to bionic pancreas treatment – for a total of 13 weeks.

The trial results, published in the New England Journal of Medicine, were significant. While the control group saw no significant change in their glucose management as measured by A1C levels and time spent within a healthy range, the artificial pancreas group saw marked improvements in both metrics. Encouragingly, the bionic pancreas seemed to cause few adverse effects, with hypoglycemic episodes being the most commonly reported (though still infrequent) event, and the total number of hypoglycemic episodes being roughly the same between both groups.

In addition to the primary publication, several companion papers were published in parallel in various other journals, which expanded on the results from the primary trial. One of the companion studies switched the control group to bionic pancreas treatment after the initial study period had concluded and saw that group experience similar improvements to the initial test group. Another study replicated the primary study but used fast-acting instead of standard insulin in the bionic pancreas. They, too, reported levels of efficacy similar to the primary study.

Put more simply, these results seem to suggest that modern closed-loop bionic pancreas systems are better at maintain healthy blood sugar levels than other current standard type 1 diabetes treatment options.

How does it work?

The system at the heart of these recent studies follows the conceptual blueprint that has existed for decades. Simply put, it incorporates a CGM monitor and an automated insulin delivery pump through an AI interface that uses complex algorithms to fine-tune insulin dosages in real time. It uses its intelligent computing power to identify trends in blood sugar data that allow it to administer insulin proactively without users having to report their carb consumption.

Benefits of a bionic pancreas

It has been reported that less than 25% of type 1 diabetics consistently achieve their target blood sugar levels with conventional methods, with the figures being even lower for diabetic children. Researchers have speculated that this is likely due to the “complexity and difficulty of management”, rather than any inherent inefficacy of conventional self-administered methods.

Given the significance of consistent blood sugar maintenance in a diabetic patient’s prognosis, methods that facilitate easier, less-involved monitoring and insulin administration could therefore play a critical role in improving treatment efficacy. Bionic pancreas systems represent the ultimate advancement in this direction by minimizing the need for a patient’s reliance on manual methods that are easily missed or skipped for convenience. This advantage is especially relevant for young, old, or cognitively limited diabetic patients who may struggle the most to keep up with their treatment demands.

Precursors to the newest bionic pancreas technologies already limited patient input to recording basic information like the timing and amount of food consumption, with the software taking care of the rest. These designs have been improved even further by the algorithmically-advanced technologies studied recently, which only require users to input their body weight at the outset, and eliminates the need for periodic dosing adjustments otherwise made by patients or their doctors. This elimination of the potential for human error is likely responsible for much of the improvement in A1C levels and time-in-range observed in the recent studies and represents a huge benefit for patients who struggle with consistency.

Challenges and considerations

Because the artificial pancreas technology studied in these recent publications represents the cutting edge of advancement in diabetic care, relatively little real-world data exists about the challenges that patients using this technology might face. However, precursor technologies have existed on the market that achieve much of the same functionality, which can offer some insight into considerations for patients interested in artificial pancreas technology.

Like any sophisticated new technology, artificial pancreas systems present a learning curve for optimal function. While the entire point of the system is to limit user input and oversight, patients will still need to adjust to and become familiar with the requirements of the system in order for it to function reliably. Especially for older patients who may be less comfortable with technology usage, the digital connectivity and interfacing of these systems may also pose an intimidating obstacle.

New technologies in medicine are also usually associated with substantial cost premiums. Many diabetic patients without health insurance already struggle to afford the basic supplies to perform manual insulin injections, so the accessibility of artificial pancreas for populations with fewer means may be more limited. Of course, accessibility will vary on an individual level depending on a patient’s budget, health insurance coverage, and device selection.

In the world of technology, increased complexity is often accompanied by increased vulnerability to malfunction. Bionic pancreas systems rely on multiple technical components that all communicate properly and co-ordinate their activity on the basis of intelligent algorithms. An issue with any of these components can cause disruptions that may not be easy to catch for the average user, who likely relies on their tech to “get it right”.


The studies explored in this article may represent the crowning achievement of decades of medical research. From as early as the 1970’s, researchers have shared the conceptual vision for a system that would use technology to give type 1 diabetics what they needed most: a functioning pancreas. Now, these studies seem to confirm that such technology has substantially arrived and is capable of greatly improving outcomes for type 1 diabetics with no apparent safety issues.

Closed-loop and hybrid-closed-loop systems still may not be accessible or medically appropriate for everyone. But all patients can and should be excited for the future of care for type 1 diabetes, as research continues to pave the way toward a standard of care that maximizes freedom and minimizes stress for diabetics.