Tandem’s t:slim, A Next Generation Pump

The design success of Apple, Inc. with products like the iPod, iPhone and, iPad, has raised the bar of consumer expectation. It’s not enough that a product works; it must be pretty and “intuitive,” like Apple’s products. Every commodity must be an experience– from the opening of the pretty, exactly-sized box with small, precise fonts, to the feel of the shiny and cold surfaces, to the fluidity and ease of the software interface.

And now that we have and know these Apple products, we raise our eyebrows at some of the other products we encounter. Why should I have to search around through five different menus to save a phone number on this Motorola phone? Why does this tea kettle have such an awkwardly placed handle? Why is the bevel on this monitor so large and unattractive? Why can’t this be more like my iPod, my iMac, and my iWhatever?

Which brings me to a question I’ve had for a while, and one I’ve heard coming from many other diabetics as well: why on earth can’t my insulin pump be more like my iPhone? I love and appreciate the Minimed pump I have, and I know I’m  better because I wear it every day– but why is it so big? So round and plasticky? Why does it have so few buttons and such a low-resolution screen, like an old Casio watch? Why do I have to iterate through eighteen zillion menu screens just to change the time? Why can’t they design a battery compartment cap that doesn’t wear away every time I open and close it? Why is it so hard to read in the dark? Why oh why would you design the web interface for data uploads to work only in Windows?

Sigh. There are answers to these questions, and usually the answers I hear go something like: the medical device world is not like the software world you are used to. Everything takes longer, and change is dangerous. You can’t just make it, just ship it, and fix problems or add features when users complain. People’s lives are at risk. The FDA has to approve everything, every step of the way. We’re dealing with a different kind of consumer here– you can’t just shake things up and hope everyone adapts. So, development is slower, and the cutting edge is trailing.

But. Still. Does it have to look like a pager? Can’t we aim for at least 2005?

I am not alone in my plaintive requests. And, to my surprise and delight, I learned recently that I am not unheard either.

Recently I had the chance to visit the offices of Tandem Diabetes Care, a biotech startup that promises a “new era” of diabetes products.  Tandem Diabetes Care was started in 2007 with the help of Dick Allen, an angel investor and long-time player in the health care and biotech industries. Allen’s granddaughter has type 1 diabetes, and so he was intrigued upon hearing about new insulin pumping technology being developed by a fellow Baxter International alumnus, Paul DiPerna. DiPerna was the president of Phluid Corporation, which had developed and patented a method of infusion that could allow for much tighter control and monitoring of the amount of fluid leaving the pump. Phluid’s system could potentially offer more precise dosing and a simpler, more compact mechanism. Though he was not an expert in pumping technology, Allen understood how useful such improvements would be for diabetics like his granddaughter, and he decided to invest.

Since that early stage, Tandem has recruited a number of big names to join the team, both on the management and the investment side. DiPerna continues to guide the development of the pumping technology as the Chief Technology Officer, and Kim Blickenstaff, former CEO of Biosite, Inc. and former Dexcom board member, joined the company in late 2007 as the President and CEO. The Board of Directors includes such notable names in the diabetes world as Ted Greene, founder of Amylin Pharmaceuticals, and John Livingston, former VP at Minimed, while Dr. Steven Edelman serves on the Clinical Advisory Board. Meanwhile, on the investment side, over $68 million has been raised in three funding rounds from venture capital firms like Delphi Ventures, TPG Capital, and HLM Venture Partners.

All this is to say: Tandem Diabetes has its name-dropping credentials in order, and its cash on hand. This is not just an exercise, or a couple of diabetics in a room with folding chairs, putting together their personal wish list. And that’s important; the insulin pump market is controlled by a small number of players right now, including Minimed, Johnson and Johnson’s Animas, Insulet Corporation, and Roche, and entering that market as a new company with a unique product may prove difficult.

Given the barriers to entry, what does Tandem intend to offer to make their mark in the market? Well, soon after the initial partnering of Allen and DiPerna, it was clear that the new pumping technology alone would not be a sufficient selling point for investors and users; Tandem needed to figure out what pump users and potential pump users wanted that they couldn’t get from the insulin pumps currently available.

And here’s the part that impressed me: when I visited Tandem, three years in to product development, it was obvious that they didn’t decide what a pump should be within a locked room full of medical device manufacturers, pouring over FDA analyses of current systems. They had done their market research, and they were listening to what diabetics were saying. They originally thought everyone would want a patch pump, like the OmniPod, and were prepared to design one. But, when they spoke to pump users, they found that the tubing wasn’t the main complaint; the main complaint was that the pumps looked like hospital devices, and worked like cell phones from the eighties. Over and over, the Tandem team heard that patients wanted smarter, slimmer, more modern pumps. Diabetes, they came to understand, is a self-managed disease, and diabetics clearly wanted some panache in the tools they used for management.

So, with the voice of their populace in their ears, Tandem is developing the t:slim— an insulin pump from the iPhone mould. It doesn’t have the same set of push-buttons, a single set of up-and-down arrows like on old alarm clocks; instead, the pump has a touch screen. The current prototype touch screen allows the user to manipulate settings, view multiple configurable profiles, access keypads, set basals, and administer boluses by directly interacting with the menus and objects, just like on an iPhone or iPod.

The software reflects the company’s focus on user experience and an awareness of human factors, and takes advantage of basic design principles common in consumer devices but absent from so many medical devices. For example, numbers are entered via a number pad, rather than with up-and-down arrows. (Yes, that’s right: the t:slim includes an actual ten-key number pad for entering boluses, time, and so on. It’s like insulin pumps have finally discovered the Arabic numeral system!)

It was clear to me from speaking to the engineers and from sitting in on a company-wide meeting that Tandem was designing carefully the look and feel of their device, and was concerned not just with basic usability, but with how much patients liked the pump. The interface was designed in the Apple-style white-text-on-black, which greatly increases the contrast and therefore readability of the information on the screen, but, more importantly, makes the pump look like a modern device. Put another way, unlike with my current Minimed pump, a stranger might guess that the t:slim is a cell phone or an mp3 player.  She probably won’t guess “pager.”

The novelty of the t:slim, however, does not stop at the user interface; the original infusion mechanism that began Tandem is still an integral part of the system, and one that may make the t:slim an extremely precise and consistent pumping system. All of the pumps currently on the market deliver insulin via what is essentially a large, motorized syringe– a cartridge holds the insulin, and has an opening that is directly connected by a tube and catheter to the body. A mechanized pumping system gradually moves forward a stopper at the back end of the cartridge, pushing insulin out of the cartridge and into the body. The problem with this system is that it is not perfectly accurate; the stopper at the back of the cartridge is pushed forward, but the actual rate and amount that it moves forward varies slightly due to the natural friction and resistances of the materials involved.

In other words, it’s a mechanical system, and there is room for minor error, especially early on in the life of a cartridge, as the stopper and plunger “get into the groove” of the pumping action. Typically, therefore, insulin pumps chart the accuracy of the amount of insulin dispensed as compared to the amount that was supposed to be dispensed in “trumpet charts,” like the one below, which comes from the Minimed Revel User Guide:

A syringe-style pump will report, as the Minimed does, a delivery accuracy of within 5% in either direction, but really what this means is that the amount that is actually delivered will converge to within 5% of the amount that is supposed to be delivered– in the case of the trumpet curve above, a basal rate of 0.1 units per hour– over an observation interval of several hours. As the chart shows, the actual error rates can be quite extreme, especially early on in the observation; however, over time, as the stopper adjusts and the negative and positive deviations from the desired amount even out, the amount actually delivered will be sufficiently near the amount intended to be delivered.

Now, as a pump user who is sensitive to even small amounts of insulin, I don’t like the idea that converging to within 5% over several hours is considered adequate, and I especially don’t like the idea that the rate and amount of insulin delivery is so variable at any given point. As a result, the hope of a much more accurate and verifiable system of pumping was very interesting indeed. The t:slim’s delivery mechanism is not based on the traditional motorized syringe system. Instead of having a cartridge with an open exit for insulin at the top, the Tandem design uses a “micro-shuttle valve.” The valve is like a tiny gate at the top of the cartridge that is pushed back and forth across the opening by a small motor. When the gate is opened, insulin can be delivered in remarkably small increments.

Due to its unique mechanisms of action, the t:slim is smaller than many standard pumps, but still holds up to 300 units of insulin in its cartridge. Furthermore, the gated design avoids some of the problems with the traditional the syringe design, which Kim Blickenstaff at Tandem describes as a “loaded gun” of insulin pointed at the patient. In other words, mechanical failure of a traditional pump could theoretically empty the entire cartridge of insulin into the patient, since there’s nothing blocking or monitoring the exit of insulin from the top of the cartridge. In the Tandem pump, however, even full mechanical failure couldn’t push out all remaining insulin into the patient.

An additional innovation– a pressure sensor within the cartridge– allows the system to report almost exactly how much insulin has been delivered. Traditional pumps, when measuring “insulin on board,” or the total amount of insulin output within a given period of time, do not really measure how much was let out; instead, they measure how far the plunger was pushed, and calculate accordingly. The valve and sensor system of the Tandem model, on the other hand, tracks how much fluid has actually exited the cartridge,

So the t:slim prototype has razzle and dazzle, and may have more accurate mechanisms to back them up– what’s lacking? The most notable shortcoming is the lack of Continuous Glucose Monitor (CGM) integration. According to the Tandem team, they recognize the value of the CGM, and intend to have an integrated system as soon as is possible, but they didn’t want that feature to slow the approval or release of the first generation of the t:slim.

And, I hate to say it– but it’s no iPhone. They’ve done a remarkable job with the insulin pump, and I liked the prototype more than my current pump in terms of feel and design, but Apple gets some of the details right in ways that prove really difficult in a medical device that also has to be affordable. It’s hard to pinpoint exactly; perhaps it’s that the t:slim’s screen lacks the luster and indestructibility promised by the iPhone’s Gorilla glass, and perhaps it’s the fonts or the colors or the glossiness of buttons. What I mean is, it’s a great interface– it just doesn’t have the Apple “Aha!” moment of intuition and tactile pleasure.

I suppose that’s more a comment on the success of Apple, though, than on any real shortcoming of Tandem; it’s certainly the sexiest insulin pump I’ve seen thus far. So the next important question is– when can you actually get one? As with many new diabetes technologies, that proves to be a question for the FDA. Tandem is required to apply for clearance for the t:slim under Section 510(k), meaning they must demonstrate that their device is “substantially equivalent” to existing insulin pumps.

In the meantime, Tandem is slowly exiting stealth mode, and had its first company showcase at the American Association of Diabetes Educators Conference in August. Not surprisingly, they met many interested diabetics and educators on the conference floor.  They must be careful with the marketing of the t:slim until the device is FDA approved, however, and not make any claims about the device’s viability or availability.

As a result, much of the marketing focus right now is toward building a company brand– Tandem Diabetes Care, bringing new technology backed by research. And to be frank, it must be working, because as nice as I think the t:slim is, I’m more impressed by the company and the people I met. The pace of innovation and development at many medical device companies is appreciable, but far from sufficient, and an awareness of user experience often feels altogether lacking. So when I see a company coming together with a push towards not just accuracy, but also product look and feel– I take note. And I feel a little bit more optimistic about diabetes, and the future of diabetic care. If I have to meticulously manage this disease until there’s a cure, it would be nice if the management were as easy and intuitive as possible, so I am extremely happy to see a new player in the game, pushing everyone closer to more usable and more likable insulin pumps. It will take companies like Tandem, with their Apple-worthy ads showing just a hand manipulating a touch-screen pump above the tagline “I control my therapy,” to shake up the insulin pump market like Apple shook up the consumer electronics market– and that sort of disruptive capitalism I can get behind.

Karmel Allison

Karmel was born in Southern California, diagnosed with Type 1 Diabetes at the age of nine, and educated at UC Berkeley. Karmel now lives in San Diego with her husband, where she is loving the sunshine, working in computational biology at the University of California, San Diego, and learning to use the active voice when talking about her diabetes.

Karmel Allison

Karmel was born in Southern California, diagnosed with Type 1 Diabetes at the age of nine, and educated at UC Berkeley. Karmel now lives in San Diego with her husband, where she is loving the sunshine, working in computational biology at the University of California, San Diego, and learning to use the active voice when talking about her diabetes.