Tandem 4Q14 – Sales of $18 million rise 75%; artificial pancreas R&D project begins; 480-unit t:flex to ship in 2Q15 – February 24, 2015

This afternoon, Tandem CEO Kim Blickenstaff led the company’s 4Q14 financial update, its fifth since the company’s $125 million IPO in November 2013 and subsequent 4Q13 call. We found it to be yet again very strong results. Below, we enclose our top highlights from the call:

Financial and Business Highlights

1.  Sales of Tandem’s t:slim insulin pump and related supplies totaled a record-high $17.9 million in 4Q14, 33% sequential increase from 3Q14 and a 75% year-over-year (YOY) gain from 4Q13 sales of $10.2 million. Full year 2015 sales are expected in the range of $70-$75 million, representing growth of 41%-51% vs. 2014.

2. Tandem shipped 3,929 t:slim pumps in 4Q14, a 63% YOY increase from 4Q13 and a 34% sequential increase from 3Q14. Since launch in August 2012, 18,342 pumps have been shipped, which Tandem estimates is ~4% of the US insulin pump market.

3. Management believes 60 sales reps/territories is the “right size” for now, and the focus in 2015 will shift to increasing reps’ productivity. The average sales rep in 2014 sold ~15 pumps per month per territory.

4. Sales to distributors increased to ~79% of total sales in 4Q14, a sequential rise from 76% in 3Q14. Management attributed the increase to the new United Healthcare contract. More patient access is the upside, though margins and control are lower.

5. Tandem reported cash resource of $69.3 million, sufficient for operating needs for the next 12 months. This afternoon, a Tandem press release announced an underwritten common stock offering aiming to raise ~$60 million in net proceeds. This was not discussed on the call.

R&D Highlights

6. Notably, Tandem has initiated an artificial pancreas project to commercialize a predictive low glucose suspend (PLGS) or basal closed-loop system. The company plans to file an IDE in 2H15 with the FDA for a clinical study. A late 2017 launch is “not unreasonable.”

7. The 480-unit t:flex insulin pump will begin shipping in late 2Q15, and very modest sales of ~$1-3 million are expected in 2015. Currently, ~10% of Tandem’s customers have type 2.

8. Management expects FDA approval and commercial launch of the t:slim G4 in the second half of 2015. FDA interaction has been very positive.

9. New to Tandem’s pipeline is “Project Odyssey,” which will allow remote online updating of the t:slim’s pump software. An FDA 510(k) filing is expected in 4Q15.

10. Last week, Tandem released the t:simulator app on the app store, a free marketing tool that allows users to experience using the t:slim with their own Apple device.

Financial and Business Highlights

1. Sales of Tandem’s t:slim insulin pump and related supplies totaled a record-high $17.9 million in 4Q14, a 33% sequential increase from 3Q14 and a 75% year-over-year (YOY) gain from 4Q13. Management characterized it as an “outstanding” quarter, and indeed, this marks three straight quarters of 75%+ YOY growth and five straight quarters of 47%+ YOY growth. It’s still early days at this point, as Tandem has been public for just one year. As expected, the increased sales were driven by the 1Q14 sales force expansion (36 to 60 territories/reps), as well as a typically strong fourth quarter for pump sales – nearly identical to 2013, 36% of Tandem’s full year sales came in the fourth quarter. This year (2015) will provide quite an instructive look at the sustainability of Tandem’s growth, as the company does not have a sales force expansion planned, though potential launches of two products should certainly be expected to drive growth

• Full year 2014 sales of $49.8 million grew 71% from 2013. The results were at the top end of the revised guidance issued in 3Q14 ($48-$50 million), and still within the 2Q14 guidance of $48-$54 million. As a reminder, the guidance tightening in 3Q14 mostly stemmed from an underestimated impact of the sales force expansion (and associated disruption of existing territories).

Table 1: Tandem Sales – 4Q13-4Q14

Figure 1: Tandem Quarterly Sales – 2012-2014

• Full year 2015 sales are expected in the range of $70-$75 million, representing growth of 41%-51% vs. 2014. The more modest growth is perhaps to be expected given no planned sales force expansion, a larger base of sales, and the launch of the competing Animas Vibe in 1Q15 (management said it took the Vibe into account when building the guidance). We certainly believe the current strong sales force could become even more efficient as well as more productive. The new t:flex pump is only expected to contribute $1-3 million in sales in 2015, given the limited reimbursement of pumps in type 2 diabetes. Though approval of the Tandem/Dexcom t:slim G4 is expected in 2H15, it has not been baked into the guidance. The trajectory of 2015 sales is expected to follow a similar quarterly pattern to 2014: a lighter 1Q (~16% of total sales in 2014) and a stronger performance in the second half of the year.

2. Tandem shipped 3,929 t:slim pumps in 4Q14, a 63% YOY increase from 4Q13 and a 34% sequential increase from 3Q14. Since launch in August 2012, 18,342 pumps have been shipped, which Tandem estimates is ~4% of the US insulin pump market (up from an estimate of ~3% in 3Q14). Extrapolating, that would put the US pump market at ~460,000 type 1 patients. Management said that “roughly” half of Tandem’s customers report that they converted to the t:slim from MDI, consistent with prior calls – this was not stated in prepared remarks (only in response to a direct question in Q&A). The company’s first pump renewals (four years since launch) will come at the end of 2016. Management said that pump pricing has been fairly consistent at “around $4,000.” Currently, ~10% of Tandem’s customers have type 2 diabetes.

Table 2: Pump Shipments in 2014

Figure 2: Quarterly Pump Shipments (2012-2014)

3. Management believes 60 sales reps/territories is the “right size” for now, and the focus in 2015 will shift to increasing reps’ productivity. The average sales rep in 2014 sold ~15 pumps per month per territory. The expectation is to move towards 20 as the average for 2015 (translating to ~14,400 pumps for the year, or ~33% growth in pump shipments from 2014). Longer term, management hopes to move to the low-to-mid 20 pumps per rep/territory, especially once reps have multiple products to carry. Given no sales force expansion this year, a lot of the growth will rest on higher rep productivity and increased awareness of the company. 

4. Sales to distributors increased to ~79% of total sales in 4Q14, a sequential rise from 76% in 3Q14 and up from 66% in 4Q13. This represents five straight quarters of rising distributor sales, which translates into greater patient access but lower margins and less control for Tandem. There were a number of concerned questions on this front during Q&A. Management attributed the increase to the new United Healthcare contract, which is serviced through a distributor – 4Q14 was the first full quarter of this relationship. Given the size of United, management said the percentage of sales through distributors may actually continue to slightly increase, depending on the how mix of payers works out. Longer-term, Tandem believes 50% direct/50% distributor is an achievable and conservative goal. No time horizon was provided, though we imagine this would take at least a couple of years.

5. At end of 4Q14, Tandem reported cash of $69.3 million, reflecting a burn of $12.6 million during the quarter. Current cash is sufficient for operating needs for the next 12 months. Importantly, the company has managed to reduce its cash burn for past three quarters.

• This afternoon, a Tandem press release announced an underwritten common stock offering aiming to raise ~$60 million in net proceeds. Due to SEC rules, the company was not able to comment on the offering during the call. Tandem’s market cap sits at $311 million, down from a high of $687 million in January 2014 and ~$428 million when the company went public in November 2013.

R&D Highlights

6. Notably, Tandem has initiated an artificial pancreas project to commercialize a predictive low glucose suspend (PLGS) or basal closed-loop system. The company plans to file an IDE in 2H15 with the FDA for a clinical study involving its first artificial pancreas product – management said both PLGS and basal closed-loop system are options, and discussions with the FDA will help define the regulatory pathway and product configuration. In Q&A, a “late 2017” launch was deemed “not an unreasonable timeline,” given the major steps that need to happen (clinical work in 2016 + a ~12 month regulatory review). Tandem will use the Dexcom sensor (we assume Gen 5, given the references to the t:slim’s built-in Bluetooth) and take on the PMA submission. Management called the development a “very exciting step for company” and a “significant opportunity for near and long-term growth.”

• The decision to go for a predictive suspend or basal closed loop is a tough call for Tandem. We believe “basal closed loop” is effectively the same as “hybrid closed loop” (patients still take care of meals, but the system deals with basal needs in the background), and would consider the terms interchangeable.
  
  • PLGS is likely an easier and cheaper study to conduct, a simpler algorithm, and would probably move faster at the FDA. However, if a late 2017 launch is indeed reasonable, that would put Tandem quite far behind Medtronic’s MiniMed 640G in the US – as of JPM 2015, Medtronic had plans to launch it by April 2016 stateside, though the F3Q15 call this month expected a PMA submission later this calendar year. As noted in our ATTD 2015 coverage, the MiniMed 640G is already available in Europe and Australia, and Medtronic has a slew of studies ongoing.
  • Basal closed loop (aka hybrid closed loop) is a more complicated algorithm, probably a harder clinical study path, and probably a more expensive endeavor; the diabetes management upside is probably much greater than PLGS, though the company would risk more in funds to develop such a system. A late 2017 Tandem launch of a basal closed loop would put the company more on pace with Medtronic’s planned launch of the MiniMed 670G (April 2017 in the US).
• We wonder if Tandem would consider licensing an academic algorithm for basal closed loop, such as those developed at UVA, DREAM, UCSB, Cambridge, MGH/BU, Padova, or others. The DREAM consortium has already obtained a CE Mark for its algorithm (see our ATTD 2015 Day #3 report), and the UVA team shared plans with us to move its algorithm through regulatory approval (see our ATTD 2015 Day #4 report). 
• Aside from Medtronic, there are many industry “maybes” in the artificial pancreas game, and it’s good to see Tandem jumping in more fully. We look forward to seeing how quickly the company can move – certainly, the public announcement of this project is a good sign. Aside from Medtronic’s efforts to automate insulin delivery:
  
  • Insulet’s new management team seems most enthused about alternative drug delivery and the type 2 OmniPod – we’ve not heard anything on the artificial pancreas (see JPM 2015) since the team changed hands. The previous management team seemed to have very low interest in automating insulin delivery. Still, given partnerships with Dexcom and an unnamed CGM partner, an artificial pancreas application of the OmniPod is possible, especially with artificial pancreas global authority (formerly of Sansum) Dr. Howard Zisser there.
  • J&J originally signed a partnership with JDRF in 2010 to develop a first-gen automated insulin delivery system; a small in-clinic feasibility study was presented at ADA 2014, less ambitious than the one shown at ADA 2013.
  • Asante has expressed public interest in an artificial pancreas (see AADE 2014), though no concrete plans or timing have been announced. The company’s planned IPO earlier this month has been delayed.
  • Roche is working on its own CGM, but it’s still at least a year away from commercialization, and we assume an artificial pancreas application is beyond that. The company has alluded to some internal discussion (see JPM 2015), so there may indeed be something to report on in the coming year or two.
• As a reminder, Tandem’s t:slim does have a Bluetooth LE radio inside it, and a research version of the pump has been used in several artificial pancreas studies around the world (e.g., Bionic Pancreas, UVA’s DiAs system.). Interestingly, the Bluetooth radio is included but not enabled in the commercial t:slim. We wonder if Tandem could turn it on via a remote software upgrade, as it is thinking about with Project Odyssey (see below).
• Tandem has not mentioned the dual-chambered pump since it went public, and we heard nothing on the call on this front. We’re not sure if this is still in the pipeline or would be part of this project.

7. The 480-unit t:flex insulin pump will begin shipping in late 2Q15, and modest sales of ~$1.0-$3.0 million are expected in 2015. That projects to the average territory initially selling one additional pump per month and eventually scaling to three pumps per month – at the midpoint of two pumps per month in the 60 territories for the latter six months of 2015, that projects to ~720 t:flex pumps this year. Currently, ~10% of Tandem’s customers have type 2 diabetes.

• Management put forward its view that this will not be a major sales catalyst this year, given the undeveloped market and the challenges of reimbursement in type 2 diabetes. This pump will be billed using the same reimbursement codes and pricing as the t:slim pump. That said, Tandem do better than it is noting –  pumps in type 2 will continue to be more widely discussed, we believe, especially as Medtronic gears up its type 2 diabetes sales following the results from the OpT2mise trial.
• As a reminder, FDA clearance of t:flex came in January following a lightning quick two month review (blowing away management’s expectations from the 3Q14 call for a six-month review). The t:flex is brilliantly designed, using the exact same t:slim pump body but a larger 480-unit cartridge that slides into the pump (for context, the original t:slim cartridge is 300 units). The new offering should give Tandem an upper hand vs. options from competitors, since many patients require over 100 units of insulin per day, particularly type 2s. According to Tandem’s market research, two-thirds of endocrinologists cited limited capacity as the #1 barrier to pump adoption in patients with type 2 diabetes. At 480 units, Tandem’s t:flex will be the largest reservoir on the market by far (second is the Roche Accu-Chek Combo at 315 units).

8. Management expects FDA approval and commercial launch of the t:slim G4 in the second half of 2015. This is right on par with the expectation for a 12-18 month review following the PMA submission in July 2014. CEO Kim Blickenstaff said conversations with the FDA continue to be “positive and interactive.” The Agency completed its site inspection in December, and management was “very encouraged” by the outcome. The company is actively working to respond to the Agency’s recent questions. Management spoke very positively and confidently on this timing and dialogue and some preparations are already underway. Like the Animas Vibe, the t:slim G4 will not be able to upgrade to the new G4AP algorithm – that could be a deal breaker for some patients, as one analyst highlighted in Q&A.

9. New to Tandem’s pipeline is “Project Odyssey,” which will allow remote online updating of the t:slim’s pump software. An FDA 510(k) filing is expected in 4Q15. This seems similar in theory to what Dexcom did with the G4AP algorithm (remote online upgrade of the already commercialized G4 Platinum receiver). The goal of Project Odyssey is to provide a user experience similar to smartphones, where a new version of the device’s operating system can be downloaded very easily. Initially, the goal is enhance the customer’s experience through better software – we imagine things like user interface, menus, etc. could be optimized. Longer-term, Tandem has a broader vision to offer a next-gen universal pump platform – through Odyssey, patients and providers will have the flexibility to select the “most fitting software for their needs.” This seems brilliant in our view, as it could allow a pump user interface more tailored to the individual needs of the patient (e.g., a simple setup for a type 2, a more complicated setup for an intensively managed type 1).

10. Last week, Tandem released the t:simulator app on the Apple app store, a marketing tool that allows users to experience using the t:slim. The app is available for free and an Android version is in development. The t:simulator is used for demo purposes only and seems especially targeted at those who want to explore what using the pump is like for some of the most commonly used tasks. We appreciate efforts like these to de-risk the scariness of going choosing a pump for those on MDI, since it’s hard to try-bef0re-you-buy (with the exception of Asante’s free trial program).

Pipeline Summary

Questions and Answers

Q: Could you reflect on representative productivity? As we think about representatives broadly, is it a bell curve distribution of productivity or are there more representatives that are further or less along than expected? Moving forward, in terms of peak productivity, are you still comfortable with 25 pumps per month as an accurate metric?

A: In terms of productivity, we are not going to comment on the scale, but there is a range of productive representatives, with some very highly productive and some at the lower end of our expectations. The average for the year was in the mid-teens, which we are quite happy with, and the  expectation is that they will move toward 20 as an average for 2015. Longer term, we expect overall productivity to move to the low to mid 20s.

Q: Does the guidance for 2015 reflect any potential? I am going to call it a sort of G4 sensor “warehousing” as patients that may have been t:slim candidates decide to wait until the t:sensor comes out.

A: There is no sort of new sales impacted by t:slim G4. At this point, we are assuming we don’t have it. That may impact us from competitive standpoint, especially with Vibe coming to the market.

Q: Is there any update on a propriety infusion set?

A: It’s something we have been interested in and have evaluated, but from an R&D perspective as well as from an overall business perspective, there is no definitive project to speak to today other than to acknowledge that it is something we have looked at and are considering. We are heavily market research focused and will continue to be.

Q: Could you provide more detail on the G4 products. Perhaps you can comment on site inspections? Timelines? When will you be able to say more definitively that you will be able to launch in the second half of 2015 versus something closer to the end of the year?

A: Site inspection is a key step – to have that completed before Christmas was a big deal for us. That was one of the big things on the check mark. Responding to deficiency letter is the second major milestone. We think it will be within the range of the guidance we’ve given and that we can get that wrapped up and into the marketplace in the second half of the year. Overall, our interaction with the agency was pleasing, and from what I can tell, they want the product on the market. We are preparing to commercialize.

Q: Is it correct that there is no assumption of any sales for G4 in this year’s guidance of $70-75 million?

A: That is correct.

Q: On warehousing, it sounds like you have taken into consideration that Vibe is out there and that there may be some patients trying to decide which product to use. Is this correct?

A: Yes, it has been a consideration for full year and quarterly guidance.

Q: How much of Q4 sales are coming from gaining share from competitors on exchanges versus expanding MDI?

A: It is roughly consistent with what we have said in the past, about a 50/50 split.

Q: On the predictive low suspend, there  was quite a bit of talk in Paris about the 640G from Medtronic. You mentioned a submission for an IDE in the second half of this year and assuming that a PMA takes a year and working with the FDA might take another year, is late 2017 a good estimate for something like this?

A: That is not an unreasonable timeline. Those are the main activities that would have to happen in order to get into the marketplace.

Q: Who’s PMA is that?

A: That would be our PMA.

Q: But you would be using the Dexcom sensor?

A: Yes, the Dexcom sensor.

Q: Ok. The reason I ask is because  you have a 510(k) pump and they have a PMA sensor, but when you put them together, it is your PMA?

A: It’s a PMA, correct. We do not do the clinical trial part. All that will be determined as we move into the predictive low glucose suspend or the closed-loop basal algorithm. That may be the clinical trials we’re signing up for.

Q: Does guidance assume representative disruption is fully resolved?

A: It is fully resolved. Now the focus is working the productivity, up from low expectations.

Q: Can you explain your guidance in terms of the percent of sales from distributors versus from direct sales? The quarter came in a bit higher than expected, at 79% – is that correct? How are you thinking about the percent of sales coming from direct channels next year? Could you also comment on the trajectory of the conversation over the next few years? Moving forward, how should we be thinking about all of this?

A: It was roughly 80/20. Now that we have a full quarter of United Healthcare being serviced by distributor, we have seen that sort of impact. That is the big plan. It may continue to increase slightly, depending on how the mix of payers works out. Long-term, I still believe we will move toward something like 50/50 as an achievable, conservative goal.

Q: It is early in the launch of Vibe, but, on the competitive front, what are you seeing in the field, and what is the sales approach to that product? Could you also comment on the Asante products and what you’re seeing in field from them?

A: On Vibe, I think the fact that it is sensor-enabled increases the number of companies with sensor enabled products, so that is essentially raising a sales barrier. It is important that we make progress on our PMA, so that we can nominally be in that game and improve our products over the second half of the year. They will be fairly strong coming out of the gate, because they are #2 with that sensor-integrated offering. As far as Asante goes, we are doing a sort of free sampling program – try it before you buy it. It was a regional offering; it wasn’t nationwide. We haven’t seen much change from that strategy. That’s part of what we saw on the road show activity as the free sampling program was one of their major selling points. But we will not know how recent financing activities will impact their nationwide rollout plans which we have somewhat factored into our guidance for the year.

Q: Can you quantify a level of volume necessary to break even on cartridge sales, and do you think you might reach this in 2015?

A: We are not providing detail at that level for cartridge manufacturing or gross margins by product. We are making meaningful progress, but it is still something where we expect to see higher margins with our pumps than we would see in our supplies in the immediate year ahead of us.

Q:  Can you talk a bit more about your newly announced products? For the Odyssey software, do you expect to file at the end of the year? What type of features will be upgradeable for customers in the first generation of the product? Could you elaborate on your long-term vision? It sounds like the plan is to eventually make a decision support tool collaborating with payers.

A: The concept of Odyssey is that it’s a conduit to upload new versions of software to the various products we are selling. We’ll have the base t:slim, tflex, and t:products or t:slim G4. They allow us to remotely upgrade software on each of those three pumps to make each of them the latest and greatest version. And there are a lot of smaller esoteric features. The load sequence can be improved. This is a mechanism to get that software improvement out into the field, just like you upgrade your phone. Once it is 510(k) approved, we can update pumps in the field. The flow through will be the 510(k) if it is needed or a letter to file the type of application for regulatory upgrades for the specific device itself.

Q: The second round of competitive bidding has ended recently. What has the impact of this been?

A: It’s dropped off the map. Competitive bidding really isn’t the threat it was a year and a half ago. In fact, it’s not part of the competitive bidding packages that are out there now.

Q: What is the distributor mix like this year? Is it correct that we may see something higher than 80/20 on a full year basis in 2015?

A: It is expected to come down later in year. 80/20 is what we saw this year.

Q: What makes that happen?

A: A lot of it is signing the new contracts and taking advantage of the contracts we have in place that have not been utilized.

Q: In terms of getting a first generation, low threshold suspend and even a second algorithm platform to market, what do you think the artificial pancreas is going to cost?

A: We don’t have any guidance we can give at this point in time, but we have a clearing agenda on the part of our R&D group. In last two years, our R&D group has been spending so much time on catchup 510(k)’s, the PMA approval, and the development of flex and G4. I think capacity is going to free up to allow us to do more work, but there is going to be a clinical component that we haven’t had to deal with before. I don’t think it’s tens of millions of dollars, but it’s going to be some incremental spend, and we have not yet given any projections or given any guidance about that.   

Q: Closing in on three years plus in the market in the US. You will have some of your own pumps starting to roll up later this year or early in 2016. How do you plan to hold onto those patients? What are you seeing on pump pricing, and what is built into the guidance for 2015?  

A: Renewals will start to hit at the end of 2016. We launched in Q3 of 2012, so it’s roughly a four year renewal cycle after which we’ll start to see some renewal activity. Hopefully at that point, we’ll have the Dexcom G4 on the market, which will hopefully be an upgradeable option for these folks who come off our t:slim. That’s one opportunity for us. Pricing has been fairly consistent and steady, roughly around $4,000 – that hasn’t really moved that much in the last few quarters.

Q: Are you seeing any different levels or change in competitor responses – for example, promotions – aside from Asante’s programs in terms of more substantial competitors?

A: There is the Asante program. Our largest competitor, Medtronic, has used some economic incentives. That hasn’t changed in last few quarters. It’s still a significant part of our sales barriers as we come up against them.

Q: Is the t:slim G4 a completely different pump than the regular t;slim? Is the actual device different?

A: There are different radios - different innards on the inside.

Q: Down the road, would you have both systems – the t:slim G4 and the regular t:slim – or would you just go with one? Then, could the patient and or physician decide to enable the G4?

A: They are going to be separate SKUs. They will be separate product lines that will be maintained, and their softwares will be slightly different because one has integrated CGM display information while the other does not.

Q: So, it is not just a software difference?

A: There are also hardware differences, and our future generations might be different pumps as well.

Q: When is the right time to start thinking about providing promotions or upgrades to patients? When Medtronic launched the 530G, I think they were a year early.

A: We hope we don’t have to get into this upgrade game. We hope we’ll get approval quickly enough and that patients will be able to delay a purchasing decision because they wanted to integrate with a CGM, and they will have access to that sometime later this year. They do have discretion to keep using their pumps if they keep working. We’re hopeful that we don’t have to get into this upgrade game that others have tried to play in order to hold their installed base.

Q: If you come in at the earlier end of your expectations – that it takes 12 months – how soon would you be able to start manufacturing and selling?

A: We’re starting to prepare a little bit now. Assuming we get an approval, there will be some lag time at which point the sales force would have to get trained, develop their marketing and sales strategies. Once they start to sell, they will have to go through the normal cycle of generating sales orders as well as going through the insurance verification process, which typically takes around 30 days. Usually 2-3 months until we start selling. It is similar to the t:flex.

Q: One thing we have picked up on in our survey work with some of the existing Animas customers that also have the Dexcom CGM is that the G4AP algorithm with the CGM device has them so hooked that they would be reluctant to upgrade to the Animas Vibe since it does not have the G4AP algorithm. Would the Odyssey Program be something that could enable that upgrade? Is this upgradeable without Odyssey?

A: Odyssey would not be able to upgrade that.

Q: Would that require an amendment to the pre-market approval?

A: It would be a new PMA. Any future generation would be a separate project. We’ll address that strategy on an interim basis at a later date as we talk about evolving the Dexcom relationship.

Q: Can you speak to the questions that you have received from the FDA – are these easy to answer or would they require a lot of additional work?

A: It is a relatively minor amount of work, just some retesting. Nothing is fatally flawed. It is just more time rather than risk. We feel as though we’re in a good position to respond in a timely manner and be within the timeframes we’ve given.

-- by Adam Brown and Kelly Close