Medtronic 2016 Analyst Meeting – 7 new CGMs to launch in next 5 years; 670G, Guardian Connect in US by April 2017, Watson app this summer; 20 million patients by 2020 via new care and payer models – June 6, 2016

Medtronic held its biannual 2016 Analyst Meeting earlier today, following previous iterations in 2014 and 2012. Diabetes Group President Hooman Hakami shared an illuminating 27-slide presentation, highlighting major plans to launch seven new CGM products over the next five years. “We will launch more new products in the next five years than we have in the last 15.” We include product pictures below in the top pipeline highlights, followed by business and financial highlights. [Note that Medtronic’s fiscal year runs from May to April, meaning “FY17” on a slide denotes May 2016-April 2017. We’ve used calendar timelines for simplicity.]

The bigger focus on CGM is a smart strategic move for Medtronic: (i) expanding into professional CGM and the gateway to a broader patient population than with insulin delivery alone; (ii) entering a less penetrated, higher growth market than pumps; (iii) equipping its Service & Solutions business with better measurement tools to drive and document outcomes; (iv) potentially better margins than pumps as sensors become fully disposable; and (v) countering strong competitive products and growth from Dexcom and Abbott, who threaten its installed base.

CGM Pipeline Highlights

1. Personal CGM sensors will see four products over the next five years if the schedule holds as planned by management: (i) Enlite 3 by April 2017 (in the 670G and Guardian Connect); (ii) Harmony 1 by April 2019 (10% MARD, 10-day wear, one calibration per day); (iii) Harmony 2 by April 2020 (<10% MARD, 30% reduced size, and “additional biometrics”); and (iv) combo seven-day CGM-insulin infusion set by April 2021. Eliminating fingerstick calibrations was not mentioned.

2. Professional CGM will see three new products over the next five years: (i) iPro 3 by April 2018 (single-use, blinded, clamshell transmitter, MARD: ~11%); (ii) iPro 4 by April 2019 (adding real-time data to iPro 3); and (iii) iPro 5 by April 2019-April 2020 (dramatic hardware upgrade, appearing as a slim, round bandage, seven-day wear, <10% MARD).

Insulin Delivery Pipeline Highlights

3. Medtronic still plans an FDA submission of the MiniMed 670G/Enlite 3 before the end of June, and slides still pegged it to launch by April 2017 (FY17). This implies an FDA review of ~9-10 months at the most (assuming a June submission). Following last week’s very positive FDA artificial pancreas webinar, that ambitious timing seems more realistic. Medtronic believes it is ~2-3 years ahead of the competition with the 670G. We think that may be ambitious given how fast the smaller companies can move, particularly on factors like user interface, but there are still many benefits Medtronic has related to infrastructure and clinical trial experience.

4. The next-gen closed-loop product after the 670G appears to be a smaller touchscreen pump with smartphone control (!), an algorithm that performs automatic bolus corrections, and “biometric,” “multi-parameter” sensing. Launch is expected after May 2020, meaning this is still early. We speculate below on why the delay will be so long between the 670G and this follow-on. We think conservative timing is smart.

5. Seven planned infusion set products will launch over the next three years: the MiniMed Pro-set (with BD), a new Quick-serter device, an extended wear set, two different CGM-insulin delivery combo sets, and two unknown innovations that looked like infusion sets (“SC2,” “Solo”). No other insulin delivery plans were shared, suggesting either CGM is a bigger R&D priority now or that fewer iterations on the pump platform may be needed. We believe CGM should be a bigger priority given the goal to help a much larger percentage of patients.  

Data and Analytics Pipeline Highlights

6. The first-gen app with IBM Watson will launch “this summer,” and based on the slide, include retrospective pattern analysis at minimum. In ~summer 2017, Medtronic will launch “glucose prediction,” followed by a ~summer 2018 launch of “therapy dialogue” (i.e., Q&A chat with Watson). As of last month, it sounded like regulatory discussions were still underway and will drive what the app includes in the first-generation. There is big interest and excitement on this front from patients and clinicians.

7. Over the next three years, Medtronic will launch four new analytics products to aid provider analysis of CGM: a Logbook app by April 2017 (adding behavioral context to CGM), Outcomes Snapshot by April 2018 (population health, quality metrics, benchmarking), and Patient Snapshot by April 2020 (personalized care plans via proCGM and Biometrics).

Business and Strategy Highlights

8. Medtronic Diabetes expects to double its revenue to ~$4 billion in five years, growing “mid-teens” operationally YOY. The aim is to serve 20 million patients (a ~20-fold increase) by the end of that time, and turn into a “diabetes management” and “outcomes” company with “globally balanced” revenue. Calendar year 2015 revenue totaled just over $1.8 billion – the company hasn’t yet achieved $500 million in revenue in a single quarter, though its last reported quarter was very close. We saw the tightest mission statements yet for the three Diabetes divisions: “Artificial pancreas as standard-of-care” (Type 1); “Make glucose a vital sign” (Type 2); “The leader in VBHC [value-based healthcare] for diabetes” (Service & Solutions). The last one in particular seems very smart given the payer landscape globally.

9. Medtronic estimates the global pump market at ~$2.4 billion annually, expected to grow 10%-11% in the upcoming year (FY17). US sales were flat in 1Q16, so this seems like a high growth rate. Global sales of personal CGM were estimated ~$800 million, with 25%-30% expected growth in FY17. Professional CGM is still a tiny market, recording <$100 million in sales in FY16, according to Medtronic. Assuming ~80% of Medtronic’s FY16 sales were in pumps (~$1.5 billion) and ~20% were in CGM (~$375 million, most OUS) gives it a global market share by sales of ~62% in pumps overall (~71% in durable pumps) and ~46% in CGM. Since the last analyst day, Medtronic has held share in its durable pump business, but ceded significant share in CGM.

Integrated Care and Access Highlights

10. Mr. Hakami detailed impressive pilot work with a US health system to prevent hospital readmissions in uninsured people with diabetes. Fifty-two patients have been enrolled in the program thus far, with blood glucose down from 242 mg/dl to 164 mg/dl after eight weeks, and a remarkable 3.5% A1c reduction (baseline: 10.8%) in the five completers thus far at 14 weeks. Zero patients have been readmitted.

11. Medtronic Diabetes is experimenting with new business models to expand access globally: a “risk-sharing model” in Colombia, a subscription model for CGM in Spain (where CGM reimbursement doesn’t exist), and a direct-to-consumer approach in Thailand.

12. Mr. Hakami acknowledged that the UnitedHealthcare-Medtronic partnership has led to “some noise” in the community about restricted pump choice. His response: “This is not about limiting choice. This is about us being in favor of evidence and outcomes.” He suggested that other payers may gravitate towards this type of relationship, and echoed comments from the earnings calls last week.

HM: The Diabeter clinic will add a fifth location in the Netherlands, and a slide highlighted expansion potential in the UK, Middle East/Africa, and now the US. Diabeter is in the lowest quartile of hospital admissions and the lowest decile of cost across all pediatric type 1 clinics in the Netherlands. Medtronic has two ways to scale the model globally: turnkey and franchise.

CGM Pipeline Highlights

1. Medtronic plans to launch four new personal CGM sensors over the next five years. All will be Bluetooth-enabled, with a greater push to make them smaller, more accurate, easier to use, and longer lasting. Dexcom and Abbott have clearly pushed Medtronic to close the gap on the CGM accuracy, reliability, and wearability fronts – all places where Enlite and Enlite Enhanced have lagged behind for the past few years. We’re especially glad to see the large clamshell transmitter will be retired in favor of smaller, disc-like form factors (i.e., like FreeStyle Libre).

• Enlite 3 by April 2017: 11% MARD, seven-day wear, two calibrations per day, one-year transmitter life. This is expected to launch in the US by April 2017 with the Guardian Connect standalone CGM system (under FDA review) and the MiniMed 670G (FDA submission before the end of June 2016). We’re glad to see the timing is consistent with JPM expectations shared in January. The expected accuracy – MARD 11% - is exactly what we saw in an ATTD 2016 poster, and a big improvement over the current Enlite and Enlite Enhanced sensors (MARD ~13%-14% with 3-4 calibrations per day).
• Harmony 1 by April 2019: 10% MARD, 10-day wear, one calibration per day. A pivotal trial will now occur in the first half of 2017, behind last month’s expectation that this sensor was “about to start an FDA pivotal study.” Harmony 1 will include intelligent, redundant sensing, and be indicated for abdomen or arm wear. It will retain Medtronic’s same historic clamshell transmitter design, a wearability disadvantage to what Abbott and Dexcom will likely have out by that time. We last saw data on this sensor at ATTD 2016 (MARD: 10.9%, n=55 sensors vs. the Bayer Contour Next Link meter).
  
  • Harmony 1 will match Dexcom’s plans for G6 – 10-day wear, one calibration per day – though it is running about a year behind G6’s timing (pivotal study in late 2Q16 or early 3Q16, per Dexcom’s April call). We wonder if the July 21 Dexcom-FDA meeting on a replacement claim for CGM is a gating factor in this pivotal’s timing. Alternatively, Medtronic may not go for a replacement claim with Harmony 1.
• Harmony 2 by April 2020: <10% MARD, 30% reduced size, and “additional biometrics” (not specified). This is the first we’ve ever heard of Harmony 2, which will see a solid improvement in accuracy according to management and what looks like a significant overhaul to the on-body hardware to make it more like the disc-like FreeStyle Libre patch. It looks fully disposable from the picture, but we’re not sure. The slide said this CGM will have 30% reduced size, though we’re not sure if that refers to size on the body or size of the sensor implant. The on-body size looks slightly taller, though the wearability looks much better relative to the current clamshell.  Calibration was not mentioned, which presumably means one fingerstick per day will still be required.
• Combo seven-day CGM-insulin infusion set by April 2021: the single insertion will house the Harmony 1 CGM sensor and a cannula for insulin infusion. Nice! We first heard about this at last month’s Advocate Forum, where management called it a “formidable challenge,” but said it could be transformational. As a reminder, Medtronic’s MiniMed Duo combination insulin infusion-CGM sensor set launched in Europe in June 2014, but the three-day wear set was not ideal for the CGM sensor. We assume the company learned a great deal from this product that it will integrate into a more commercially viable, longer lasting set.

2. Medtronic expects to launch three new versions of the iPro2 Professional CGM over the next five years. All will include Bluetooth, and starting with iPro4, share real-time data. The focus is on making professional CGM single-use (disposable) and smaller on the body, and therefore, easier to prescribe and wear. The Qualcomm partnership announced two weeks ago was not mentioned, but we assume it will help develop all three “single-use” systems, particularly iPro4 (adding real-time data) and iPro 5 (bandage-like). This is exciting given the massive need to improve diabetes management in people with type 2 diabetes, particularly those on basal insulin who need to intensify and those on MDI who are not taking the right amount of insulin but do not know how to improve their insulin settings.

• iPro 3 by April 2018: This product was labeled as “single-use” on the slide, which presumably means both the sensor and transmitter will be disposable – an upgrade from the reusable iPro 2 transmitter. iPro 3 will operate only in masked mode (similar to the current iPro), and offer an MARD of ~11%.
• iPro 4 by April 2019: The big upgrade here is the addition of real-time data, which was emphasized in the Qualcomm partnership announcement two weeks ago. Otherwise, iPro4 resembles iPro 3 in terms of a single use design, clamshell transmitter, and ~11% MARD.
• iPro 5 by April 2019-April 2020: iPro 5 looks like a dramatic upgrade in wearability and form factor, and appeared to be a slim, round bandage. The seven-day sensor is expected to have a <10% MARD. The slide did not mention “real-time,” which could mean it will be masked to achieve the dramatically improved form factor. This would really keep Medtronic competitive with Dexcom/Verily’s bandage-like CGM and potentially a future version of Abbott’s FreeStyle Libre Pro.

Insulin Delivery Pipeline Highlights

3. Medtronic still plans an FDA submission of the MiniMed 670G/Enlite 3 before the end of June, and slides still pegged it to launch by April 2017 (FY17) – this was excellent to see, and implies an FDA review of ~9-10 months at the most (assuming a June submission). Following last week’s very positive FDA artificial pancreas webinar, that ambitious timing seems more realistic. Said Dr. Courtney Lias, “Medtronic has just finished its pivotal study – that’s in the public domain – and one would hope if they put in a high quality submission, it doesn’t take long to review.” Of course, this is the first hybrid closed loop, a new sensor, and a new pump platform, so many things have to go optimally for a speedy review. Still, it was terrific to hear the comments from Dr. Lias!

• A slide listed “potential benefits” of the MiniMed 670G: reduce A1c, increase time-in-range, and reduce patient burden. Mr. Hakami’s commentary added, “preventing dangerous outcomes.” That would be an impressive slew of benefits out of the gate! We were thrilled to hear outcomes beyond A1c mentioned and emphasized, given patient priorities. Mr. Hakami mentioned that pivotal results will be shared at ADA later this week (99-LB). The three-month, single-arm trial wrapped up in March, and though there wasn’t a control group, this study will give an early sense of the 670G’s efficacy. What can we expect on time-in-range and hypoglycemia reduction? Will it reduce A1c? Will there be information on reduced stress and anxiety? On user interface and training?
• “Beyond the clinical results, the other impressive thing is the patient reaction to this product.” Mr. Hakami reiterated the impressive fact 80% of patients in the 670G pivotal have continued on the system through the FDA’s continued access program. We have heard user interface needs to improve with the 670G, but even without that, patients are enthusiastic about continuing. It is among the most engaged patients that currently take part in trials so we hope continued attention is put on improving user interface across the board.
• Medtronic believes it is ~2-3 years ahead of the competition with the 670G. Our automated insulin delivery competitive landscape shows at least seven systems with potential to be commercialized over the next few years, though Medtronic could be right if companies cannot meet their timelines and if Medtronic achieves its goals. Animas plans to launch its hypoglycemia-hyperglycemia minimizer by November 2017 (only seven months behind the 670G), though its pivotal will only start later this year, implying a very tight turnaround time. Tandem plans to launch a predictive suspend system in 2017 and a hyperglycemia-hypoglycemia minimizer in 2018, though both are still in feasibility studies. Meanwhile, Bigfoot Biomedical and Beta Bionics plans to run pivotal studies in 2017, which likely puts them two years behind Medtronic with ~2018 launches. 

4. For the first time ever, Medtronic shared a picture and detailed product plans for the next-gen closed-loop product after the 670G: it appears to be a smaller touchscreen pump, smartphone control (!), an algorithm that performs automatic bolus corrections, and “biometric,” “multi-parameter” sensing. Launch is expected after May 2020, meaning this is still in early development. Aside from what looks like a dramatic hardware improvement, this will integrate the DreaMed algorithm to close the loop further – instead of only increasing basal insulin to gradually mitigate highs (670G) and bring blood glucose back to target, the next-gen algorithm will add automatic bolusing to correct highs. As a reminder, the 670G algorithm has a cap on insulin delivery per hour, and its use of basal modulation to keep blood glucose in range is what makes it a hybrid closed loop – it still requires meal boluses. The implication is a missed meal bolus on the 670G hybrid closed loop could still mean several hours above range, as the higher basal rate will take a while to bring blood glucose back in zone; it sounds like the more automated next-gen version will add this feature in.  

• Why is the next-gen version of the 670G four years away? We have several hypotheses:
  
  • Medtronic needs a more accurate and robust CGM sensor before adding automatic correction boluses to fully close the loop. This is entirely possible, and based on the above CGM pipeline, it will take a few years to bring this into the fold.
  • Time to learn from the MiniMed 670G in real-world use. How will Medtronic need to adjust the next-gen algorithm based on what it learns from the 670G’s regulatory review and real-world use?
  • Smartphone control of a pump will take time to implement, particularly via regulatory discussions.
  • Comfort that the 670G is well ahead of the competition, leaving room to perfect a second-gen system. On the other hand, we would be cautious if we were Medtronic – the landscape is packed, and other companies may be able to move faster through FDA and reimbursement once Medtronic paves the way. Ease of setup and use is also a question with the 670G, areas where other companies like Bigfoot and Tandem will compete aggressively.
  • Conservative timing: There are many unknowns with this planned system, so the timing may simply be conservative and easy to beat by the time the next analyst day rolls around in 2018 (!). We wonder about the addition of biometric sensing – what will be measured (e.g., activity, sweat, stress, heart rate), and what will be incrementally useful vs. a potential false signal?

5. An overall pipeline slide (20 products) indicated seven planned infusion set innovations over the next three years: the MiniMed Pro-set (with BD), a new Quick-serter device, an extended wear set, two different CGM-insulin delivery combo sets, and two unknown innovations that looked like infusion sets (“SC2” and “Solo”). The most near-term product is the MiniMed Pro-set with BD, which will see an initial launch by September 2016, with full commercialization ~October-December 2016 (per BD’s May 2016 call). Though these products could be very impactful for patients, it was not surprising they were skipped over in prepared remarks – they do not have the glamor of a new pump or CGM sensor. Still, innovation in consumables is key, since it’s a huge chunk of Medtronic’s revenue (~$800 million per year, per the 2014 Analyst Meeting), is highly profitable, and is recurring revenue.

• We’re elated to see the longer wear set is an R&D priority. This could be truly impactful from a wearability, hassle, and cost perspective, though there is a lot of biology to sort out with a seven-day wear set.
• Today’s slides and remarks did not share other insulin delivery plans. Is CGM a bigger R&D priority now? The MiniMed Flex product (a “hybrid” durable pump with a smaller footprint that can be worn on OR off the body) was last mentioned at the 2014 Analyst Meeting, and we wonder if it has been dropped from the pipeline. Patch pumps and simple insulin delivery devices also were not mentioned. It seems that Medtronic is moving the R&D focus to sensor innovation, which makes sense given the potential for professional CGM to expand into most type 2s globally.

Data and Analytics Pipeline Highlights

6. The first-gen app with IBM Watson will launch “this summer,” and based on the slide, include retrospective pattern analysis at minimum. In ~summer 2017, Medtronic will launch “glucose prediction,” followed by a ~summer 2018 launch of “therapy dialogue.” This is consistent with previous comments to launch a cadence of Watson apps, though the most specific timing and insight into what each version will include. In zooming up to read the screenshots, the first Watson app will include past-looking insights (“In the last 30 days, high glucose pattern found usually after glazed donut for breakfast”), while glucose prediction adds the forward-looking pattern insight we first saw at CES in January (“In the next two hours, there is a high likelihood that you will go low”). The third update, therapy dialogue, is the ultimate compelling vision for this “Personal Diabetes Assistant” – Watson: “Hello William, how can I help?” William: “I would like to do a 30-minute run today. How should I prepare?” Bring it on!

• As of last month, it sounded like regulatory discussions were still underway and will drive what the app includes in the first-generation. In addition to retrospective data analysis, we assume hypoglycemia prediction (fairly low risk) could be added to the first-gen version. Today’s screenshot, however, only showed retrospective data analysis in the first-gen app.
• A slide indicated the Watson app could be a direct revenue generator, implying for the first time that Medtronic might charge for use upfront or on an ongoing basis. We assume this could work like the Pro version of the free mySugr app, which costs $2.99 per month and adds premium features like a bolus calculator, search capabilities, etc. Perhaps Medtronic would similarly have a free version of the Watson app with more basic capabilities, and then charge on an ongoing basis for bonus features. Alternatively, the model could be to charge health systems or payers to give the app to patients, assuming Medtronic can demonstrate meaningful outcomes. It will be fascinating to see patient reception to a subscription if that is the future direction – pure speculation at this stage.
• We assume this app will launch in the US first, though will be expanded globally once MiniMed Connect and/or Guardian Connect are available outside the US. The key to the whole system, of course, is real-time glucose and/or insulin data streaming into the app.
• Medtronic is still positioning the Watson app as a product for those not on automated insulin delivery (e.g., Guardian Connect standalone CGM users). Presumably it will read data from other devices and apps on the phone (food, activity, sleep), or ask patients to enter food and insulin data from within the app. We’re not sure how many features will be incorporated in the first-gen version, how high the entry burden will be, and how much the analysis will improve outcomes.
• Consistent with previous remarks, the Medtronic/IBM Watson “Personal Diabetes Assistant” will ultimately address three questions: How have I been doing? How will I be doing? What should I be doing? We first saw this at last month’s Advocate Forum. We’d also love to see the Medtronic/IBM Watson “PDA” address “what should I be doing less of?”

• Retrospective analysis: “In the last 30 days, high glucose pattern found usually after glazed donut for breakfast.” The example showed a CGM trace, making the food-glucose cause-and-effect relationship very clear. This has been missing for too long in diabetes, though of course it requires accurate and specific food entry – not easy.
• Future prediction: We first saw this über hypoglycemia prediction capability at DTM last fall and heard more about it at CES. The system reportedly has 80-90% accuracy predicting hypoglycemia within three hours of bolus insulin delivery. Alarm annoyance factor and false alarms are the big unknowns here, but this kind of hypoglycemia prediction could be very useful.
• Real-time advice: Chat box: “Hello William, how can I help?” William: “I would like to do a 30-minute run today. How should I prepare?” We LOVE this, and today’s commentary put this about two years away.
• Will the Medtronic/IBM Watson app live up to the hype? Could it make Medtronic’s CGM a more compelling choice vs. competitors? How gamechanging could a Personal Diabetes Assistant be? How much useful real-time and retrospective insight will the app provide in the first-generation? Will the hypoglycemia prediction capabilities generate too many alarms? Can the app actually bring meaningful therapy dialogue to patients – “Watson, how much insulin should I take right now?” We also wonder to what degree, if any the Medtronic/IBM Watson PDA will be able to address questions related to stress and anxiety. “I want to be done with diabetes, Watson!”

7. Over the next three years, Medtronic will launch four new analytics products to aid provider analysis of CGM: a Logbook app by April 2017 (adding behavioral context to CGM), Outcomes Snapshot by April 2018 (population health, quality metrics, benchmarking), and Patient Snapshot by April 2020 (personalized care plans via proCGM and Biometrics). These were not detailed further, though all are consistent with the greater push into professional CGM, and the move with IBM Watson to make data more actionable. These reports build on the launch of Pattern Snapshot last fall, a one-page iPro CGM report that prominently shows a patient’s top three glucose patterns and lists up to six possible causes for each one.

Financial and Strategy Highlights

8. Medtronic Diabetes expects to double its revenue to ~$4 billion in five years, growing “mid-teens” operationally going forward. Mr. Hakami still aims to serve 20 million patients (a ~20-fold increase) by the end of that time, and turn into a “diabetes management” and “outcomes” company with “globally balanced” revenue. This vision really started at the Analyst Meeting two years ago, where newly hired Mr. Hakami first shared the plan to start a type 2 diabetes business unit. The 20 million by 2020 vision first came out in our interview with Mr. Hakami and Anette Brüls last April following the Diabeter clinic acquisition. We love the notion of Medtronic helping so many more patients globally outside of intensively managed type 1 patients. This is definitely the vision of Minimed Founder Al Mann.

• Medtronic is definitely building a solid foundation to realize this vision, though it will require a serious step-function change, and much hinges on CGM innovation, new business models, and global expansion. These are the only way it will penetrate into a broad type 2 diabetes population, which will be key to serving so many more patients globally than today.
• Mr. Hakami shared the tightest mission statements yet for the three Diabetes divisions: “Artificial pancreas as standard-of-care” (Type 1); “Make glucose a vital sign” (Type 2); “The leader in VBHC [value-based healthcare] for diabetes” (Service & Solutions). We have not ever heard the Service and Solutions mission statement put so succinctly. The focus on glucose monitoring for type 2 first emerged at the company’s Advocate Forum last month, and was reiterated in the Qualcomm partnership and last week’s earnings call. The artificial pancreas has always been a focus in the core type 1 business, but we are encouraged to see how far Medtronic hopes to take it.

• Medtronic expects Diabetes revenue from Emerging Markets to triple in five years, rising from 8% of annual Diabetes revenue today ($140 million of $1.8 billion) to 11% of revenue by FY21 (~$440 million of ~$4 billion). No further details were shared, though new pricing and business models will obviously be key in driving this significant expansion. Reported emerging market sales have been roughly flat since Medtronic Diabetes started sharing them five quarters ago ($35 million, $31 million, $35 million, $37 million, $37 million), though they saw strong 23% operational growth in 1Q16 (they don’t state reported sales).

9. Medtronic estimates the global pump market at ~$2.4 billion annually, expected to grow 10%-11% in the upcoming year (FY17). The slide pegged the durable pump market at ~$2.1 billion (8%-9% expected growth in FY17) and the patch pump market at $0.3 billion (20%-25% expected growth in FY17). For context, our 2015 Industry Roundup pegged the worldwide pump market at ~$2.2 billion in 2015, fairly close to Medtronic’s model. [As a reminder, we confirmed with Medtronic earlier this year that roughly 80% of its sales now come from pumps, down from ~85% based on a chart presented at the 2014 Analyst Day. That said, that means that $375 million is CGM revenue, which sounds high to us., but is mostly international, since Dexcom’s sales are so US-weighted.] Insulet’s worldwide OmniPod sales were $250 million from 1Q15-1Q16 (i.e., $0.3 billion in the equivalent of Medtronic’s FY16).

• To see 10%-11% pump industry growth over the next year, the field will need to see real acceleration from 2015. Given guidance from Medtronic, Tandem, and Insulet, this seems possible if US growth re-emerges. As a reminder, we estimated pump industry sales growth at 1% as reported YOY in 2015, meaning operational growth was probably in the mid- to high-single digit range. Assuming the upcoming 10%-11% growth is an operational growth forecast (it was not specified), it is possible given: Medtronic’s expectation to grow in the mid-teens overall in FY17, Insulet’s guidance to grow mid-teens in the US and 50% internationally in 2016, and Tandem’s plan to grow 48%-58% in 2016.   
• Medtronic estimates worldwide personal CGM sales were ~$800 million in FY16, with 25%-30% expected growth in FY17. Professional CGM is still a tiny market, recording <$100 million in sales in FY16, according to Medtronic. However, growth in the latter is expected to be 65%-75% in FY17, a testament to the low base and Medtronic’s new focus on this segment.
  
  • For context, our 2015 Industry Roundup pegged the global CGM market at ~$744 million in 2015, up a strong 41% YOY. Last year’s sales growth makes Medtronic’s forecast for 25-30% growth seem pretty reasonable, particularly because Dexcom expects strong 35%-40% growth in 2016 ($540 million-$565 million in sales).
• Assuming ~80% of Medtronic’s FY16 sales were in pumps (~$1.5 billion) and ~20% were in CGM (~$375 million) gives it a global market share by sales of ~62% in pumps overall (~71% in durable pumps) and ~46% in CGM. These were not listed on the slide like in the last analyst day, where Medtronic reported a 60% share of global CGM and a ~70% share of durable pumps. In other words, the company has held onto share in its durable pump business, but ceded significant share to Dexcom in CGM. As of 4Q15, we estimated that Dexcom had ~66% market share in the US by sales, while Medtronic held ~68% of the international market.

Integrated Care and Access Highlights

10. Though it’s with a small group, Mr. Hakami detailed very impressive work with a US health system to prevent hospital readmissions in uninsured people with diabetes. Fifty-two patients have been enrolled in the program thus far, with a 90% retention rate, blood glucose down from 242 mg/dl to 164 mg/dl after eight weeks (n=14 so far), and a remarkable 3.5% A1c reduction (baseline: 10.8%) in the five completers thus far at 14 weeks. Zero patients have been readmitted. The hospital pilot was not detailed in full, but includes engaging with patients at the point of discharge; opting them into the program; assigning a Medtronic health coach to partner with patients; and use of the iPro2 blinded professional CGM and an unspecified mobile app. Mr. Hakami said it is “very early days” and “just a pilot,” but the company is encouraged by the initial results. The case study of one patient at 14 weeks showed a 4.7% reduction in A1c (!) (baseline: 11.5%), a 23 lb decrease in weight (baseline: 308 lbs), and robust reductions in cholesterol (206 to 130 mg/dl) and blood pressure (160/70 to 130/70 mmHg). Put this in the water! (Just kidding – but maybe the water of the most out of control type 2 patients who seem engaged about CGM?)

• We’re not sure how scalable the health coaches are, but this move makes perfect sense for Medtronic to take on, given its hospital management work in other therapeutic areas. Better post-admission follow-up in people with diabetes could save the system significant unnecessary costs, and we look forward to seeing what kind of difference Medtronic can make in some of the toughest patient populations. For the health system, hiring Medtronic for this purpose is presumably cheaper than footing the bill for readmissions in people without insurance.
• In the future, IBM Watson will help identify the patients that are at greatest risk of a hospital readmission upfront, enabling Medtronic to engage them and promote adherence. On the back-end, Medtronic and IBM will use data to optimize protocols and pathways for future patients (“a learning system”). This is where we see digital health and connected devices making a huge difference – collecting data where it is historically absent, tightening the feedback loop in high-risk patients, scaling human expertise to care for a larger patient base, and using outcomes to learn and drive continuously better results for future patients.

11. Notably, Medtronic Diabetes is experimenting with new business models to expand access globally: a “risk-sharing model” in Colombia, a subscription model for CGM in Spain (where CGM reimbursement doesn’t exist), and a direct-to-consumer approach in Thailand. Management said the latter has driven much more awareness and “patient pull” for therapy in an out-of-pocket payment environment. These updates follow plans shared last month to pilot a pay-as-you-go subscription model for sensor-augmented pump therapy in Mexico. Medtronic also has an at-risk contact with the Chengdu, China government to expand access to insulin pumps and manage a population of patients at lower cost.

• These models could meaningfully expand access in many countries, as few can currently afford the full upfront out-of-pocket cost of diabetes technology in the traditional DME business model. We love that Medtronic is thinking of new business models, which are undoubtedly needed in diabetes. Several startups are using or planning to use subscription business models (e.g., Bigfoot, LabStyle, Livongo, OneDrop), which makes us wonder if it will be common industry-wide in a decade. The Insulet business model is actual higher-revenue overall than the traditional pump model, even though there is less revenue upfront.

12. Mr. Hakami acknowledged that the UnitedHealthcare-Medtronic partnership has led to “some noise” in the community about restricted pump choice. His response: “This is not about limiting choice. This is about us being in favor of evidence and outcomes.” He added: “This was based on data and evidence. United saw what others have seen. We have a system that takes action ... we have compelling data, we can drive differentiated outcomes, and those outcomes lead to lower costs.” This paralleled comments on last week’s earnings call, which suggested that clinical and economic evidence was a driving factor in UHC’s decision to select Medtronic as its preferred provider of insulin pumps (effective July 1). We assume Medtronic brought a body of convincing evidence to UHC: STAR-3, ASPIRE data (MiniMed 530G), outcomes from the CareLink database, potentially other cost data, and perhaps 640G and 670G clinical trials.

• “This is the first time in diabetes that there is a preferred partnership around DME ... There are going to be others [payers] that will gravitate towards this, as they understand this approach. We are truly on the same side as the patient: outcomes based,” foreshadowed Mr. Hakami. This is the worry of smaller companies like Tandem, who lack capital to run large trials that show their products are clinically and economically superior. Will deals like this increasingly lock them out of the market? It is clear that in this healthcare environment, pump and CGM companies will have to prove their products’ effectiveness in convincing outcomes trials – and then go to bat with payers. We are fans of evidence and salute Medtronic for blazing a trial to collect it. On the other hand, this trend will make it much harder for smaller companies to compete, which in turn, could negatively impact innovation.  
  
  • We assume there is an adherence difference between different pumps and CGM, but payers likely don’t appreciate it – there is no comparative data to prove it! Can companies bring evidence to show their product drives better outcomes vs. competitors? The most innovatively designed product (alone) will no longer be automatically paid for in this environment, though presumably that will have something to do with outcomes. We hope more real-world pilots and payer experiments are possible – the publication lag for diabetes technology is too long, while the cost to running RCTs is too high.

Honorable Mention: The Diabeter clinic will add a fifth location in the Netherlands, and a slide highlighted expansion potential in the UK, Middle East/Africa, and now the US. While Europe expansion has always been in the strategy, and we heard about the Middle East and Africa last month, this was the first mention of the US. The slide cited “double digit growth,” which presumably means sales but was not specified. Medtronic has created a standardization approach so it can scale Diabeter in two ways. One is turnkey model, where Medtronic acquires an entire clinic and copy-pastes the Diabeter model in. In this scenario, it becomes a Medtronic-owned facility. The other model is a franchise model, where providers own the clinic, and Medtronic brings in Diabeter’s expertise to help with care pathways, IT, and benchmarking. Management said there is high interest in both models.

• Diabeter is in the lowest quartile of hospital admissions and the lowest decile of cost across all pediatric type 1 clinics in the Netherlands. What is the secret? How does an ~40-person staff at four clinics take care of ~1,700 pediatric and adolescent patients and provide world-class care? Mr. Hakami cited a few key factors: (i) an IT system with algorithms that upload data from patients on a daily basis, crunch the data, and provide a triaged list of patients every morning that the team can engage with; (ii) a multi-disciplinary team that includes physicians, nutritionists, NPs, psychologists, etc.; (iii) a unique engagement model where patients sign a contract and adhere to responsibilities as Diabeter patients.

-- by Adam Brown and Kelly Close