image_printPrint Page

Gary Scheiner, MS, CDCES
Owner and Clinical Director of Integrated Diabetes Services evaluates
the Beta Bionics iLet AID system

article by gary scheiner

Usually, when evaluating some new form of technology, I go into it looking for faults and shortcomings.  But this time was different.  Based on what I already knew about the inner workings of the iLet system from Beta Bionics, my expectations were quite low.  So…

I Let preconceived notions about the system fade from my frontal cortex.

I Let Beta Bionics’ trainers go over everything with me in detail, just as they would with any user.


I Let the algorithm do its thing without “creative interference” on my part.

Just so everyone is on the same page, iLet was originally called “The Bionic Pancreas”.  The brainchild of Dr. Ed Damiano in Boston, the system was conceived as a dual-hormone pump that would infuse insulin when glucose levels rise, and glucagon when it drops.  However, early challenges associated with procuring stable liquid glucagon led to development of an insulin-only automated insulin delivery (AID) system named iLet.

iLet pump CGM system reviewA tubed pump with a monochrome touch screen, iLet currently syncs with the Dexcom G6 or G7. 

It was designed with simplicity in mind for both users and healthcare providers.  Unlike other AID systems which require initial basal rates and bolus settings (including insulin-to-carb ratios and correction factors), iLet only requires the user to enter their weight and a target glucose level of 110, 120 or 130 mg/dl.  The algorithm makes a conservative estimate of the user’s initial basal rate, meal boluses and insulin sensitivity based on their weight, and proceeds to fine-tune these estimates based on the past several days’ glucose results.  In addition, the algorithm functions like other AID systems by raising or lowering insulin delivery based on minute-to-minute changes in glucose levels.

A very unique aspect of the iLet is its method for determining meal boluses. 

Rather than asking the user to enter grams of carbohydrate and glucose levels for bolus calculations, the user only needs to inform the system when they are having a breakfast, lunch or dinner, and whether that meal is “usual” size, “less,” or “more” than usual.  The algorithm then delivers a predetermined bolus dose and adjusts future meal doses based on what happens over the four hours that follow that meal.  For example, let’s say you entered a “usual” lunch but wound up needing extra insulin from the algorithm due to a significant rise after the meal.  Next time you enter a “usual” lunch, the dose will be a little bit greater.

For people who really dislike carb counting or have no interest (or opportunity) to be trained on carb counting, the small/medium/large estimation entries can be very appealing.  However, users still need to be able to recognize and quantify the general amount of carbs in their meals.  Likewise, patients (and clinicians) who lack the time, skill or motivation to estimate and fine-tune their own basal/bolus settings will enjoy the “plug and play” nature of the iLet.

iLet offers several other nice features: 

  • It allows the user to choose an aggressive, conservative, or “usual” target glucose, an option that is not available with all AID systems.
  • It is already compatible with both the Dexcom G6 and G7.
  • The glass insulin cartridges can be ordered prefilled, and are easy to fill on your own with fewer bubbles than plastic cartridges.
  • Software updates can also be made easily through the iLet phone app.

Where iLet is Lacking

Many of the drawbacks of the iLet are related to its quest for absolute simplicity, minimizing the burden experienced by the user and their healthcare provider. 

However, in striving for simplicity, iLet makes it more difficult to manage glucose levels in a number of common situations.  For example, with no means for bolusing for snacks, the user must abdicate snacking, bolus for snacks by entering a “small” meal (which may deliver too much insulin), or not bolus at all (which produces a sharp glucose rise).  The user has no means for correcting high glucose levels other than waiting for the system to fix it on its own.  Entering “fake meals” to fix a high tends to mess up the algorithm that is used to estimate meal bolus requirements.

Likewise, there is no effective strategy for managing glucose levels while exercising.

There is no mechanism for manually reducing basal or bolus insulin doses other than “lying” to the system by entering a smaller meal than is actually consumed.  And again, the effects of exercise on the post-meal glucose can falsely alter future meal boluses.  iLet also has no options for temporary overrides or extending meal boluses, so hypoglycemia can occur with slowly-digesting foods.

Unlike systems that function fully the moment they are activated (such as Tandem Control IQ and open-source Loop and iAPS), iLet has an extended learning curve. 

During the first several days, basal and bolus doses are estimated purely based on a person’s weight.  Not body composition, but weight.  For people who are insulin resistant, the system will greatly underestimate their insulin requirements.  For those who are insulin-sensitive, it will overestimate.  This can contribute to significant highs and lows early on.  And for optimal fine-tuning of meal doses, it is best for the user to refrain from snacks, exercise, stress, and other unusual conditions for the first 3-7 days – not an easy task for most people.   Post-meal hypoglycemia also tends to fool the system.  After treating lows, it is common for glucose levels to rebound for a few hours.  This causes the algorithm to think that MORE insulin is required the next time the same-sized meal is consumed.

bent canula

Algorithms and glucose management features aside, there are several aspects of the iLet design that could be better. 

> While the tube priming process is fairly rapid, there are many parts and pieces involved in cartridge change-outs, and sealing the cartridge (which holds only 160-170 units) into the pump requires considerable hand strength.

> The plain black & white screen can be hard to read in sunlight, and the touch screen icons (along with the on/off button) can be difficult to activate.  The programming menus have confusing names and it is difficult to extract insulin delivery history.

> There is no way to adjust the site change reminder or low insulin alert setting. The pump requires charging every 2-4 days via a docking station, which essentially requires disconnection from the pump since the charger must be plugged into a wall outlet.  Acknowledging alarms/alerts involves a multi-step process.

> And one feature that I find particularly troubling is the lack of an angled infusion set option; only 90-degree soft cannula and steel needle options are available.  We see 10X more problems with 90-degree soft cannula sets than with angled sets.

iLet Go

Overall, I think iLet meets the needs of people who prefer to not be involved much with their daily diabetes management and are just looking to get their glucose levels out of dangerously high places. 

It will do that, especially if you have a relatively bland/consistent schedule – three square meals a day with few snacks, little exercise, and minimal challenges from a hormonal standpoint.  iLet will also be a hit with healthcare providers who just don’t have the time to work with their patients on fine-tuning their pump settings, since iLet really doesn’t have any settings (other than the target glucose).

gary scheiner tries the iLet AID systemFor those who don’t mind applying the skills they’ve acquired over the years, or are happy to obtain self-management skills in order to achieve reasonably tight and stable glucose levels, iLet is likely to disappoint.

The designers took away too many tools and features that allow users to manage effectively through the rigors of daily life.  Like it or not, life is complicated for the vast majority of people and no automated system can handle everything on its own.

Personally, I’m more than happy to put some effort into my daily management in order to achieve better results.  If staying in-range means looking at data once in a while to dial-in my basal/bolus settings, estimating carbs whenever I choose to eat (we don’t have to be all that exact with carb counts when using an AID system), count me in.

image_printPrint Page