Lilly developing smart pen with insulin titration and AID with its own pump and Dexcom CGM; launch expected in ~2-3 years, trials to start by end of year – November 22, 2017

Executive Highlights

  • Lilly plans to commercialize two tech-enabled insulin delivery systems in ~two to three years: (i) automated insulin delivery with its own pump, an acquired closed loop algorithm (Montreal), and CGM; and (ii) smart pens with a paired dose titration app. Clinical trials will begin by the end of this year. Lilly itself will submit both systems to the FDA and commercialize both, pulling together all the components.
  • As we wrote yesterday, Dexcom announced a “development agreement” with Lilly, bringing its sensor into the first clinical trials with both systems. It is non-exclusive and not a commercial agreement. This is a strong competitive move vs. Bigfoot/Abbott’s ambitions and it’s great to see so many players establishing smart insulin delivery or lifting their game.
  • Lilly’s AID system includes a new, custom pump designed by Deka (Dean Kamen). The tubed pump holds three days of insulin and is “a white disk about the size of a shoe-polish tin.” The pump has an infusion set, and can be worn direct on the body or in the pocket. It will be controlled wirelessly from a handheld and come with a companion app. It will require a PMA and will not launch prior to the AID system.
  • The “Integrated Insulin Management” system includes a connected smart pen based on the disposable Kwik Pen platform – the team couldn’t tell us whether it is a cap or another attachment that enables dose capture. However, we did confirm it will work on disposable pens. The paired app will be capable of titrating, basal, bolus, and basal+bolus, which was very good to hear – it will be a 510(k) or PMA, depending on use of BGM or CGM.
  • Lilly did not send out its own press release (this came out in WSJ), but the digital/tech commitment here is “a massive effort,” according to our call today with Ms. Marie Schiller and Dr. Howard Wolpert. Lilly is taking a wide perspective on connected care and aims to pull many pieces together: insulin, connected devices (glucose monitoring, insulin delivery), evidence-based algorithms, and surrounding IT. Lilly’s Cambridge Innovation Lab has now hired ~40 people and it is our sense there have been team members on “connected care” in Indy for quite some time.

The Wall Street Journal (WSJ) published an exclusive, very positive and well-written article sharing Lilly’s ambitious goals to launch two major tech-enabled insulin delivery solutions in two to three years: (i) its own automated insulin delivery (AID) system with a custom disk-sized tubed pump (designed by Dean Kamen’s Deka), CGM, a Class AP algorithm (Montreal), a wireless controller, and a companion app; and (ii) an Integrated Insulin Management system with a connected pen (based on the disposable KwikPen), a paired dose titration app, and CGM or BGM. Clinical trials for both systems are expected to begin by the end of 2017 and will be posted on

Dexcom sent out its own press release announcing a development agreement with Lilly, meaning that its CGM will be included in Lilly’s “Connected Diabetes Ecosystem” – the two insulin delivery systems noted above as well as other possible analytical tools. The Lilly/Dexcom agreement is non-exclusive and not commercial at this time. The Dexcom sensor model is not specified, though we assume the trial might use G5 or G6, and the ultimate commercial system would likely be no calibration. This certainly brings an answer to what Bigfoot and Abbott are doing to automate insulin delivery for pumps and injections.

Both systems have been in development for two years, ever since Lilly’s Cambridge Innovation Center came online. Lilly will submit both products to the FDA and commercialize them, pulling all the components together and developing pretty much everything internally. Pricing and reimbursement models are unclear, though bundling seems like an obvious play – insulin, devices, and apps for one price (our speculation).

Read on for much more on the two systems’ planned features, which we gleaned in an enthusiastic call with Lilly’s Dr. Howard Wolpert and Ms. Marie Schiller. We also include more on the Cambridge Innovation Center and Lilly’s connected ecosystem vision; implications for Lilly’s current pump/pen partners; and what Novo Nordisk and Sanofi are doing in digital.

Lilly’s AID system: Disk-Sized Pump Designed by Deka (“Hybrid” Wear), Algorithm from Montreal Group, Handheld Controller, Companion App

  • Lilly’s pump-based AID system includes a new, custom pump designed by Deka Research and Development Corp. (led by Segway inventor Mr. Dean Kamen), an algorithm acquired from Montreal-based Class AP (tested in Dr. Ahmad Haidar’s McGill group), and Dexcom CGM.
  • The tubed pump holds a three-day supply of insulin and is described as “a white disk about the size of a shoe-polish tin.” The pump has an infusion set, but can be worn in two ways (“hybrid”): direct on the body like a patch pump or in the pocket like a traditional tubed pump.
    • We’re not clear on the durable vs. disposable components of the pump. We have not seen pictures of it, and the WSJ article was not clear on this. A durable version seems more likely – given the disposable infusion set concept – though we’re not sure. 
    • Ms. Schiller emphasized that it’s been built from the ground up for automated insulin delivery, so presumably some thoughtful design choices have been made. We assume the physical pump will not have a screen, but are not sure.
  • The AID system will be controlled wirelessly from a handheld, and will also come with a companion smartphone app. Direct pump control from the app depends on regulatory progress, but it is a stated goal. Insulet and Bigfoot are also aiming for this, so we’ll be fascinated to see who gets there first and how quickly the whole field moves to it.
  • Lilly will submit the PMA and commercialize the AID system, the first insulin company to commit so significantly to closing the loop – especially in developing its own pump.
  • Lilly acquired an evidence-based dosing algorithm for its connected diabetes ecosystem from Class AP located in Montreal. The algorithm was invented by Dr. Ahmad Haidar, a student and now professor at McGill University, in collaboration with Dr. Remi Rabasa-Lhoret and Virginie Messier from The Clinical Research Institute of Montreal (IRCM) and Dr. Laurent Legault and Dr. Benoit Boulet from McGill University. The algorithm has been tested in multiple clinical trials conducted in more than 200 people with diabetes. See the group’s page here.
    • We expect it will be hybrid closed loop, meaning meal boluses will still be required. Like Insulet and Tandem, presumably some modifications will be made to enable commercialization. We don’t know the details of this algorithm intimately. It was clear from our conversation that Lilly is thinking diligently about integration into clinical practice and the requisite clinical training/setup burden.
  • Notably, Lilly has acquired both the insulin-only and insulin+glucagon versions of the Class AP algorithm. This gives obvious potential to expand into bihormonal or glucagon-only delivery, though we assume this is for a future generation of the product.
  • The pump will not be prefilled in the first-gen product, but it has been designed to enable this in future generations. This would be a spectacular patient convenience win and something Lilly could drive in a very meaningful way.

Lilly’s Integrated Insulin Management System: Smart Pen and Insulin Dose Titration App

  • The “Integrated Insulin Management” system includes a connected pen based on the disposable Kwik Pen platform – the team could not tell us whether it is a cap (like Common Sensing or Bigfoot/Timesulin) or another type of attachment (like BD) that enables dose capture. However, we did confirm that it will work on disposable pens.
  • The paired app will be capable of titrating, basal, bolus, and basal+bolus insulin – it will be a 510(k) or PMA, depending on use of BGM or CGM. (It can use either as an input.) Lilly expects to have one app that offers a complete set of titration solutions: “It’s a platform for an entire diabetes management system for people on insulin.”
  • As a reminder, Lilly received FDA clearance for the Go Dose system for titrating Humalog (bolus) inulin last December. It has not launched to our knowledge. It’s possible this submission was to gain early experience and inform future submissions and product designs. The app, available by prescription only for patients with type 2 diabetes, provides prandial Humalog dosing recommendations using manually-entered blood glucose values. Go Dose is the patient-facing app, while Go Dose Pro allows the provider to determine a starting dose and send the recommendation to the patient’s app. As the patient manually enters glucose values, the software generates more individualized dosing recommendations until the provider determines that titration can be stabilized or stopped.

Lilly’s Connected Care Vision and the Cambridge Innovation Center

  • Lilly did not send out its own press release (it went for a broader, independent story in WSJ), but the digital/tech commitment here is “a massive effort,” according to Ms. Schiller. She added that it’s “as big” as a product development plan on the drug side.
  • Lilly is taking a systems perspective to diabetes/digital health and aims to pull many pieces together: insulin, connected devices (glucose monitoring, delivery devices), and evidence-based algorithms. Surrounding these are software and an IT environment, an area where Lilly has “invested a lot” in underlying infrastructure.
    • We are glad to see Lilly committed in this way, as it easily could have handed the component(s) off to someone else to commercialize (e.g., “Here’s a smart pen to use in your system”). Though building the internal hardware and software expertise is harder upfront – especially for a pharma company – it is arguably more likely to deliver a well-integrated product.
    • Enrique Conterno, head of Lilly’s diabetes business, fears Lilly’s insulin business would become “obsolete” if the company remains largely an insulin provider. As quoted in WSJ, he asked: “Do we want to be just an insulin provider that just goes into a system, or do we want to be the integrator of the system? To me, it’s clear where the business is going.”
  • Ms. Schiller and Dr. Wolpert highlighted a few key design principles:
    • “We have the opportunity to put Dr. Howard Wolpert in a box.” This was definitely our favorite quote of our call!
    • “Re-envisioning the whole delivery of diabetes care.” This is something Dr. Wolpert is particularly passionate about, given the trends in endocrinology and primary care, the rising number of patients, and the lack of data-driven, personalized advice.
    • Creating “best-in-class solutions” within a “connected care ecosystem”
  • Why announce this now after two years? “We were in early research,” said Ms. Schiller, “and we wanted to make sure things were at the stage worth talking about. False promises are not what this is about.”
  • We sense Lilly’s efforts are at least as significant as its competitors – perhaps more so. Novo Nordisk is working with Glooko, IBM Watson, and smart pens, while Sanofi is working with Verily (Onduo), Voluntis, Common Sensing, and One Drop. Of course, since these efforts are still mostly stealth, it’s hard to know who is the furthest along and what products will deliver the most value.
  • Lilly’s two-year-old Cambridge Innovation Lab has now hired ~40 people. Included in that bunch are mechanical engineers, material scientists, and industrial designers. The Lab occupies two floors in a new building on the fringe of the MIT campus.
    • Dr. Wolpert and Ms. Schiller emphasized the big advantage to Lilly’s approach: the Cambridge center can function as an agile startup within the context of a well-resourced pharmaceutical company. “It’s the combination that is going to be an important ingredient.”

What about Lilly’s Existing Investments/Partners: Companion Medical, Beta Bionics, and Insulet?

Lilly is taking a broad approach to smart pens, pumps, and AID, now augmenting its three partnered efforts with internal R&D. We’ll have to see how this plays out, since none of these products – internal and partnered – are publicly available yet. In the past, Lilly has had a mixed record in terms of partnerships – in 2002, it announced a major partnership with Amylin that later dissolved when it built its own internal GLP-1 (Trulicity); the same thing happened with Adocia, when Lilly partnered but later dropped out, deciding to do its own work on ultra-rapid acting insulin.

  • Companion Medical/Lilly: In May 2015, Lilly led a $3 million Series B raise for Companion to develop a durable Bluetooth-enabled insulin pen and app, which is now within days or weeks of launching in the US (see our coverage from last week). Companion Medical management has been unwavering in its commitment to making its system insulin agnostic. Ms. Schiller emphasized that Companion has a reusable smart pen solution, which augments Lilly’s internal plan for a smart pen that works with disposable delivery. We’ll see soon see how Companion Medical does in the US – CEO Sean Saint said last week that payers are covering InPen “more than 50% of the time” and copays are typically ranging from $0 to $75. InPen’s actual price ($665 wholesale, $549 cash discount) could be more expensive than a disposable offering – depending on how Lilly prices. It will be very important to see the value that all the systems bring.
  • Beta Bionics/Lilly: Similar to Companion, Lilly invested $5 million in Beta Bionics, and Lilly’s Global Brand Development Leader Ms. Deirdre Ibsen sits on the Board of Directors. Leaders did not comment on implications for Lilly’s internal automated insulin delivery project on Beta Bionics, who is also building its own pump platform and using Dexcom CGM. The insulin-only Beta Bionics system is slated to possibly launch in 1H20, on a roughly similar timeline to Lilly’s system, and the potentially differentiating bihormonal system has a pivotal set to start in early 2019, though FDA submission and launch timing depend on a stable glucagon, among other things.
  • Insulet/Lilly: As a reminder, the Lilly U500 and U200 Insulet Omnipods are on track for a U500 launch in 2019, followed by a 2020 launch of the U200 Omnipod (per the 3Q17 call). There are no publicized plans to automate delivery of U200/U500 insulin with Insulet, and we’re not sure if Lilly’s own pump/handheld will be compatible with U200/U500. On the plus side, Lilly has a tubed pump and it sounds larger than the Omnipod, both clear advantages for Insulet. This partnership has had significant investment from both companies, and we saw both co-present at DTM (sharing the first-look at the very well-designed U500 user interface). Of the three existing Lilly partnerships, Insulet seems to be in the safest position (from what we can tell); we have not heard mention of Lilly’s pump efforts before reading Peter Loftus’ WSJ piece and are not sure to what degree other partners were aware of the internal pump efforts at Lilly.

CGM and Pump Competitive Implications

  • The Lilly news is a big victory for Dexcom, who has clear path to commercialization with an insulin leader. This also gives Dexcom a strong answer to Bigfoot/Abbott’s plan for AID and injection auto-titration. Dexcom has been teasing undisclosed pen partnerships for quite some time now, so it’s great to see this fially publicly disclosed. For automated insulin delivery, Dexcom now has a growing list of partners: Lilly, Tandem, Insulet, and Beta Bionics in the US, plus Diabeloop and Cellnovo in Europe. 
    • The Lilly/Dexcom deal is a “development agreement” – similar to what Bigfoot had – so we’ll have to wait and see if a future “commercial” deal is signed. Pricing will obviously be a critical component of this, especially with pressure from Abbott’s FreeStyle Libre. Ms. Schiller and Dr. Wolpert said Lilly will take an “open architecture” approach, meaning FreeStyle Libre (real time or professional) could be integrated too. “We will continue to look at the options.”
  • For Bigfoot/Abbott, this is definite competition – and with such smart teams on both sides, competition should just result in innovation. Lilly’s vision is similar to Bigfoot’s at a high level with a two-pronged product roadmap - pump and smart pen. Lilly has some clear advantages: ability to bundle with insulin, large sales rep footprint, payer relationships, etc. Of course, Bigfoot also has deep digital and algorithm expertise, a market-ready pump (Asante), and a factory calibrated CGM (Abbott’s FreeStyle Libre). We imagine both partnerships will launch quite competitive, well-designed products that make insulin delivery far safer and more effective and better than what was previously available.
  • For all pump companies (i.e., Medtronic, Tandem, Insulet, Roche, Ypsomed, Cellnovo), Lilly’s brings strong potential competition on the product and business model fronts. We assume Lilly is quite confident in the pump and manufacturing it at scale, meaning big cost/business model changes might be possible as Lilly comes to market. If Lilly is successful in bringing a strong CGM + insulin dose titration to market, it could either reduce or increase the number of patients that go on pumps (or change payer dynamics), depending on a variety of factors (perceived ease of use of both, perceive and actual value shown), etc. Lilly can also offer a portfolio with both options, allowing patients to switch back and forth (similar to Bigfoot’s plan) – this could encourage more customers to stay within Lilly’s ecosystem, though of course formularies now dictate what insulins patients use and pumps, to a lesser degree. On the flip side, Lilly will need to successfully execute on manufacturing and scaling up as a pump company, not a given for a pharmaceutical company.
    • We wonder if Lilly would be open to selling other companies’ pumps within its Ecosystem, so that patients can choose the hardware that best fits them – but still get the other benefits of what Lilly can offer.

Insulin Companies Moving Quickly Into Digital: Novo Nordisk, Sanofi, Lilly’s Efforts

This announcement is another striking reminder that traditional pharmaceutical companies no longer envision a secure future just manufacturing drugs. To stay relevant in the evolving healthcare climate, they are investing heavily in devices, connectivity, and support. Even just in the past 30 days alone, buzz has elevated significantly from all three major insulin manufacturers. We are excited for this burgeoning ecosystem approach, and look forward to seeing how pharma navigates new business, service, and payment models, all the while developing and iterating hardware/software tools on a much faster time scale than they are accustomed.

  • Novo Nordisk: At the Palo Alto HITLAB Symposium, Novo Nordisk VP of US Device Research Thomas Miller shared a valuable 90-second video (55:40-57:16) of the company’s digital health “vision” – complete with a connected insulin pen (it looked like an auto-injector); a patient-facing app with dosing guidance, automatic carb counting, and real-time CGM; and a healthcare provider dashboard with clinical decision support. Novo Nordisk also has connected pens piloting in Sweden and being used in clinical trials stateside. Meanwhile, it has launched a first-gen app with Glooko (Cornerstones4Care), and a currently recruiting trial is evaluating Glooko’s mobile insulin dosing system (MIDS) with Tresiba. MIDS feasibility data at DTM looked solid: an 18 mg/dl drop in mean glucose, improved BGs in range over three weeks. Novo Nordisk’s recent Triple Bottom Line quarterly magazine focused exclusively on digital health. Of course, it also has the exciting IBM Watson partnership, on which there has not been an update in some time – but the potential is big for insulin titration, drug discovery, clinical decision support, and beyond.
  • Sanofi: Onduo (Verily+Sanofi) and the Blue Cross Blue Shield Association announced plans last week to pilot a type 2 diabetes management program in three states in 1Q18, including connected CGM, apps, and continuous lifestyle and medication support through Onduo’s healthcare experts. Sanofi is also in the midst of a pilot with Innovation Health, Common Sensing, and One Drop. Sanofi also has its own FDA-cleared My Dose Coach app for titrating basal insulin (currently available, though doesn’t seem widely launched yet), an investment in Common Sensing (dose capture), and has pilots expected with Voluntis’ Insulia this year.
  • Lilly: See above!

Close Concerns’ Questions

Q: What will the clinical trials starting next month look like? Sites? N? Outcomes? Devices? Inpatient or outpatient?

Q: Which of the two systems will be harder for Lilly to build? If Lilly had to pick one, which would it focus on? Which has the most internal risk? The most external risk? Which has the most upside? Which will have the most competition?

Q: How will Lilly capture data from disposable smart pens? How will it compare to Common Sensing, BD, and Bigfoot/Timesulin? What is the level of accuracy?

Q: Why hasn’t Go Dose launched? Will that be available before the ultimate product vision in 2-3 years? What’s been holding up a launch?

Q: How will the market segment between reusable vs. disposable smart pen products?

Q: What does the Lilly pump look like? Will the “hybrid” wear concept be popular? Will the pump have a screen? How far away is prefilled? How will the controller compare to others?

Q: How solid is the Montreal algorithm? How will it compare to others? Will it require meal announcement? What info does it need to start up?

Q: How will Lilly’s partners take the news – Beta Bionics, Companion Medical, Insulet? Which partnership is most at risk? Which partnership has the most upside, even with Lilly’s own internal efforts?

Q: Will Lilly seek to drive AID into type 2 diabetes?

Q: What kind of pricing models will Lilly use for both products? Will it bundle insulin with the devices + apps? Or will these be separate hardware/service systems? 

Q: How interoperable will Lilly’s system be? 

Q: What is the cost profile of Lilly’s pump? Was this a key reason to develop a pump internally instead of working with existing pumps? Will it leverage highly automated manufacturing?

Q: How will closed loop with a pump compare to a smart pen + insulin dosing advice? What will outcomes look like, and what will the cost differential be? Will payers discriminate between them?

Q: Will Novo Nordisk and Sanofi build their own pumps? Will they sign a deal with Dexcom or Abbott?

Q: Will this Dexcom development agreement turn into a commercial agreement? What has Dexcom learned from the Bigfoot deal negotiations?

Q: What other work is going on at the Lilly Cambridge Innovation Center?

Q: How does GLP-1 play into this? SGLT-2? Will algorithms help titrate these non-insulin drugs? 


-- by Brian Levine, Adam Brown, and Kelly Close