How OmniPod 5 ranks among the hybrid closed loop systems

What’s In A Number?

Here’s where OmniPod 5 ranks among the hybrid closed loop systems.

by Gary Scheiner MS, CDCES and Jennifer Smith RD, LD, CDCES

Integrated Diabetes Services LLC

Usually, when a product with a number after it hits the market, it was proceeded by similar products with lesser numbers (see Rocky 2 through 5, Medtronic 506 through 780).  The assumption is that the higher the number, the better/more advanced the product.  In the case of the OmniPod 5 (OP5), the new semi-automated insulin delivery (AID) system from Insulet, one has to wonder what happened to 1 through 4.

Perhaps those were versions that Insulet scrapped before bringing this version to market.  Without question, OP5 represents a major step forward for Insulet and a great option for many people with diabetes.  But it is good to know that there is no limit to the number of numbers that can follow the name (see Friday the 13th part XIII), because while OP5 is the first Insulet AID system to market, there remains a LOT of room for improvement.  Read on.

An in-depth look at the OmniPod 5 Insulin Delivery System by Insulet

How the OmniPod 5 Works

OP5 belongs to the category of AID systems.  This category already includes the Medtronic 670 and 770, Tandem t:slim with Control IQ, DiabeLoop, and the Loop/Open APS build-it-yourself systems.  These systems are comprised of three parts:  an insulin pump, continuous glucose monitor, and a computer algorithm that, to a certain extent, raises or lowers the pump’s insulin delivery based on the glucose level/trend and recent insulin delivery. When used properly, all systems can minimize high and low glucose levels, improve the time spent within one’s target glucose range, and reduce the mental burden involved in micro-managing one’s glucose levels.  OP5 is approved for users over age 6 who use a minimum of 6 units of insulin daily.

OP5’s automated-adjustment algorithm shares a great deal in common with Medtronic’s 670/770 system.  It takes the past two+ weeks of data to come up with an average daily insulin dose, of which it applies 50% to the daily basal requirement.  This basal requirement yields a flat 24-hour basal rate, called an adapted basal rate, which the system can then adjust based on the current and projected glucose value.  In OP5’s case, the system predicts where the glucose will be in the next 60 minutes and adjusts based on that value.  The prediction is based on the current glucose, trend, and amount of insulin on-board.  If the predicted value is above the set target, the system administers microboluses (which is a fancy way of saying that the basal rate is increased).  If below target, it is decreased.

OP5 allows the user to set a target of 110 to 150 mg/dl, in 10-point increments.  It also has a temporary modification (not sure I’d call it an “override” feature for physical activity which sets the target at 150 and reduces the adapted basal rate by 50%.  Like most AID systems, OP5 errs on the side of caution, focusing more on the prevention of hypoglycemia than achieving tight glucose control.  While basal delivery can be stopped completely, there are limits to how much it can be increased.  In situations where there is not CGM data available, such as during warmup periods, the system defaults to “limited” automated mode.  In this temporary state, OP5 delivers a basal rate that is equal to the adapted basal rate or the user’s preprogrammed basal rate, whichever is lower.

One unique aspect of OP5 is that it also adjusts bolus calculations based on the direction the glucose is headed.  If the glucose is rising, the recommended bolus may be increased by as much as 30%.  If it is falling, the bolus may be decreased by as much as 100%.  Users can look at the bolus “calculation” screen to see exactly how much was been added or subtracted from their usual dose before delivering the bolus.

In terms of hardware, the OP5 system uses the Dexcom G6 continuous glucose sensor and tubeless OmniPod disposable pump.  The computer algorithm is built into the pods themselves, so the CGM communicates directly with the pod.  A smartphone is required for performing pod change-outs, bolusing and making setting changes.  Insulet provides the smartphone “controller” with the app pre-installed (it can only be used to communicate with the pods).  The app can also be installed on a select few android devices, but Insulet plans to make the app compatible with a wide range of android and iOS operating systems in the near future.

Things we like about the OmniPod 5

First, let’s start with the Pod itself.  No other AID system features a pump that is as simple, discreet, or economical as OmniPod.  The Pod inserts its own canula and can easily be placed on many different areas of the body, which is great for those with limited dexterity or wanting to use hard-to-reach body parts.  There is no need for disconnecting, which eliminates issues caused by disconnection during sports/exercise and water activities.  The lack of tubing not only saves on insulin, it also does away with the daily inconveniences of wearing a pump and catching the tubing on drawers, doorhandles etc.…  And because the whole OP5 system costs far less up-front and can often be paid for through pharmacy benefits, it is accessible by most people with public or private insurance.

As far as the AID algorithm, it is nice that the CGM signal, receiver and computer algorithm are embedded in the pods themselves.  There is no need to have the controller at your side 24 hours a day; the pod will self-adjust basal delivery on its own.  The pod will beep if there is a problem that requires user involvement, such as when the glucose drops 55 mg/dl – a nice feature for those who choose not to have the controller with them during exercise.

OP5 is the first commercially-available system to adjust bolus dose calculations based on the direction the glucose is headed.  We have been teaching our patients how to do this manually for many years, so having a system do it automatically is a definite plus.  Our patients see better glucose levels 2-4 hour post-meal when the trend is taken into account when the bolus is calculated.

Data management is another plus with this system.  OP5 data is cloud-based (no need to plug anything in) and uses the Glooko program for generating reports.  Glooko is free for all OmniPod users, easily shared/accessible by clinicians, and produces a variety of insightful reports.

OP5 has the unique capacity to set up different target glucose levels at different times of day.  This can be useful for those concerned about nighttime highs, nighttime lows, or issues related to daytime physical activity.  Parents may also be more comfortable setting the target slightly higher when their child is not under their direct care.

Finally, OP5 is easy to set up. Just start up a pod, link the controller with the dexcom transmitter, and turn on “automatic” mode. Granted, the algorithm takes several days to learn the user’s typical insulin usage (it needs this to establish an “adapted” basal rate and estimate the maximum allowable basal delivery), but that’s really all the user needs to do.

Areas for improvement…

OP5 users cannot use the Dexcom receiver simultaneously while automated mode is running; the Dexcom G6 mobile app must be used.  This means that those who don’t own a smartphone will either need to get one or risk losing most of the functionality of their dexcom system.

Several features that current pump users find helpful for achieving optimal glucose control were taken away – including the ability to extend boluses and make temporary adjustments to basal delivery.  Even in manual mode, users can no longer bolus while an extended bolus is delivering (ever hear of something called dessert?)  Does every pump user make use of these features?  No.  But many people do, and they benefit from them.  There is no harm in allowing these features as an option for those who want them.  In our opinion, taking these features away does nothing to add to the quality of the product.

We really did not like having to keep the controller nearby almost all the time.  Even though the automated algorithm functions without the controller present, the user still needs the controller to see their CGM data (unless they also have the dexcom app on their usual phone), know the meaning of any alerts (beeps emitted from the pod), and program any boluses.  The “unlock” aspects of the controller mean that lots of button pushes are needed to clear any alerts, see any data or program boluses.  There are many instances where Insulet could have used a simple default to save steps (such as using the current CGM value populate for the bolus calculation) but did not.  They could also have expanded options for viewing CGM data within the app – only the past three hours can be seen on-screen and the user needs to look in another area to view the status of their current basal delivery.

Speaking of hidden information, we are at a loss as to why there is so much information about the algorithm that is unavailable to the user and their healthcare providers.  This includes:

• The exact percentage of adjustment being made to boluses based on the glucose trend
• The adapted basal rate
• How the maximum basal rate is determined
• Average daily insulin use (and proportion of basal vs bolus) – unless one goes to the Glooko reports

We are not fans of the flat “adapted basal rate” for several reasons.  First, for those with known variability in basal insulin requirements (for dawn phenomenon or evening growth hormone, for example), using a flat basal rate as a foundation leads to unnecessary glucose variability.  In essence, the algorithm is chasing/fixing rising and falling glucose levels rather than maintaining near the target.  We saw this personally during the night.  Our basal rates normally “peak” during the early part of the night and are much lower during the middle of the day.  This keeps us very steady through the night.  But with a flat adapted basal setting, our glucose consistently rises in the early part of the night before the algorithm kicks in with extra basal.

Another reason we are not fans of the adapted basal rate is the time it takes the system to adjust to sudden changes in insulin requirements.  Many factors can cause one’s basal needs to go up or down within a few hours, including illness, stress, menstrual hormones, steroid medications, and major changes in physical activity.  Because the algorithm adapts the adapted basal rate only after pod changes and bases the changes on two+ weeks of data, it is too slow to adjust to these types of situations and is also too slow to adjust back when things return to “normal”.  Our female clients see this with hormone shifts during the month.

And you’d better pray that you don’t lose or break your controller/smartphone (or have it malfunction).  If you have to start using a new one, the algorithm has to start fresh and “learn” your insulin requirements all over again in order to establish the flat adapted basal rate — despite the fact that each person’s insulin usage data is stored in a cloud-based server.

And here’s another funky quality of the flat adapted basal rate:  When the algorithm has to raise the basal during times when it is too low to maintain stable glucose, the extra basal is added to IOB – which will effectively reduce both meal and correction bolus calculations.  As a result, there are times of day when users may need to manipulate their bolus doses to avoid post-bolus hyperglycemia. This “extra IOB” holds true even when switching to manual mode to deliver the bolus.

As healthcare providers, we are always interested in how a hybrid closed loop system is adjusting basal insulin levels.  It lets us see if adjustments may be needed to other aspects of a person’s insulin program.  Unfortunately, downloaded reports of OP5 systems fail to show basal adjustments in any reports. However, experience has taught us that several settings may require adjustment when using “automatic” mode.  The system tends to under-deliver basal insulin any time glucose levels are near-normal – particularly going into meals. This results in frequent post-meal hyperglycemia. In order to compensate, many of our patients have had to make their insulin-to-carb ratios and correction factors more aggressive and reduce the duration of their insulin action curve. Some people also change their glucose target settings throughout the day depending on the situation. This type of setting manipulation can overcome some of the algorithm’s weaknesses, but it can also pose certain risks.  And the fact is, it just shouldn’t be necessary.

While OP5 generally does a good job at helping users avoid hypoglycemia, it can work against those looking to achieve very tight glucose control.  Users cannot set a target glucose below 110 mg/dl, even in manual mode.  And the way the algorithm functions, a target of 110 tends to produce glucose values at or above 110, but rarely below.  The “maximum” basal rate that it uses in automated mode is not based on the user’s set maximum basal rate.  The algorithm-set max basal is usually insufficient to reverse the course of a sharp glucose rise or fix very high readings.

Even the system’s ability to adjust bolus doses based on the glucose trend has a potential downside.  Rather than making a fixed dose adjustment (as is customary in advanced practice), OP5 makes percentage adjustments.  Fixed adjustments work because up/down arrows usually translate into a certain amount of rise or fall in the glucose level over the next couple of hours. By applying one’s correction factor, an appropriate amount of insulin can be added or subtracted from the calculated dose to offset the projected rise or fall.  For example, if one up arrow means that you’ll rise by about 30 mg/dl, and your correction factor is 40, 0.75 units should be added to offset the expected rise.  However, by making percentage adjustments, the amount of the adjustment could be way off.  For instance, adding 20% to a bolus dose of 2 units means that 0.4u would be added.  Adding 20% to a dose of 12 units means that 2.4 units would be added. This can result in high or low glucose afterwards the bolus is given.

Logistically, rotating pump infusion sites and sensors can be challenging for anyone who uses both devices simultaneously.  With OP5, the sensor must be placed within close proximity and “direct line of site” of the pod in order to make sure the pod picks up the CGM signal consistently.  For example, placing the pod on the arm and the sensor on the abdomen can be a problem since the CGM signal has to travel through the body to reach to pod.  We have found it best to rotate the pods around the sensor if space allows.

We also found the controller to be sub-par.  It is difficult to see the screen outdoors and almost impossible to see in direct sunlight.  The battery drains rapidly and requires up to four hours for a full charge.  Despite the large screen, the text and key icons are surprisingly small.  Information that can be very important to users, such as current IOB, is difficult to find.  Only the past three hours of CGM data is displayed.  And the system fails to let the user know if the algorithm is actively raising or lowering the basal insulin delivery and by how much (it only shows if the basal is “maxed out” or shut off completely).

How OmniPod 5 compares to other AID systems

So where does the OP5 system rank relative to the other AID systems?  I guess it depends on your perspective.  Fortunately, our team has had the opportunity to use and train/manage patients on every system, including the “DIY” versions.

• Omnipod 5 compared to Medtronic: Compared to Medtronic, OP5 has a clear hardware advantage.  The pump is smaller, simpler and easier to learn.  The lack of tubing is a plus for most people.  The need to carry the OP5 controller is a bit of a drag, but that should become a thing of the past when Insulet configures the OP5 app to work on a variety of smartphones.  And choosing between the Dexcom G6 and Medtronic’s Guardian 3 sensor is a no-brainer. Too many reasons to list here, but anyone who knows anything about CGM understands the differences in accuracy, reliability, convenience, flexibility and ease of use.  The OP5 algorithm is on-par with Medtronic’s, and perhaps a bit better given that it also adjusts bolus doses based on the glucose trend.
  8-) Advantage:  OP5
• Omnipod 5 compared to Tandem: OP5 has only a modest hardware advantage. Despite having tubing, the Tandem pump is compact and quite easy to program. And both use the Dexcom G6 as the CGM data source.  Tandem’s algorithm is relatively simple, and it works as intended from the moment the feature is activated.  OP5 has a bit of a learning curve; it often takes 4-6 pod changes before a reasonable adapted basal rate is created and glucose control starts to settle in. Because Tandem applies the user’s preferred basal settings as a starting point, those with varied basal needs tend to see more stable glucose levels overnight and between meals.  Tandem also allows users to extend their boluses and switch to secondary basal/bolus “profiles”, which effectively serves as a “temp basal” feature.  And in addition to an exercise “override”, Tandem features a “sleep” mode that allows for tighter glucose control at times when hypoglycemia is minimal.
  ?Advantage:  EVEN
• OmniPod 5 Compared to DIY (Loop): OP5’s only major advantage is that the app does not have to be built and maintained by the user.  Unlike DIY, Insulet offers excellent round-the-clock customer support.  DIY users must rely on fellow DIY users (via online platforms) for problem-solving.  But beyond that, DIY offers a number of advantages, including:Choice of pump types (OmniPod or a variety of older tubed pumps) and CGM systems (including Dexcom G6)
  > Rapid adaptability to new hardware and insulin types
  > Frequent software/feature improvements
  > Full programmability through most types of smartphones
  > Complete customization of system aggressiveness
  > Use of the user’s traditional/preferred basal pattern as a baseline
  > Unlimited customizable overrides for handling a variety of life’s challengesLoop in particular, has been shown to produce glucose levels that are superior to any system on the market.  However, achieving this requires a bit more user skill and engagement than with the commercially-available systems.
  ? Advantage:  DIY

In conclusion…

OP5 clearly holds a place in the diabetes landscape.  It will improve quality of life and glucose control for countless users.  But not all.  Those who are used to (or seeking) very tight glucose control may be disappointed.  And those who prefer to make adjustments and micro-manage on a day-to-day basis may find themselves wanting more.

Given our love for the OmniPod pump itself, we are eager and excited to see what OP6, 7, 8 etc. have in store.