September 16, 2018 - 6:22pm EST by
2018 2019
Price: 97.00 EPS 0 0
Shares Out. (in M): 24 P/E 0 0
Market Cap (in $M): 2,300 P/FCF 0 0
Net Debt (in $M): -50 EBIT 0 0
TEV (in $M): 2,250 TEV/EBIT 0 0
Borrow Cost: Available 0-15% cost

Sign up for free guest access to view investment idea with a 45 days delay.



Quick overview

iRhythm sells a heart monitoring device, called the Zio patch, which is worn for 14 days and records heart activity for people with irregular heart rhythms. The legacy standard of care for these patients is a Holter monitor, which is a bulky monitoring device with up to 12 electrodes that is worn for 24 to 48 hours and may not diagnose some patients with infrequent symptoms given the shorter monitoring period (and thus repeat the process or use a secondary test).  IRTC has come up with a single electrode patch that is easier to wear, thus has higher compliance, and records for an extended period, resulting in a higher diagnostic yield than traditional Holter monitors. This $300 device must be removed and mailed back to iRhythm after the 14-day monitoring period. iRhythm then has a team that uploads the data to the cloud and uses their machine learning algorithms which generates a report that a technician than reviews for accuracy which is then sent to the physician for review.   Until recently the Zio was the only “extended Holter” patch on the market which allowed for outsized sales growth.

IRTC went public in October 2016 as a JOBS Act IPO valued at $360 million which has quadrupled to $1.6 billion.  During this time sales have grown at an impressive 63%; however, this has not translated to scale as losses have dramatically expanded and forecasts for timing of profitability have been continuously pushed out.  As Lakewood pointed out, in early 2017 analysts expected some operating leverage and forecasted EBITDA losses to go from a loss of $16 million in 2016 to a loss of $8 million in 2018. Nevertheless, the opposite has occurred, and only one year later revised 2018 forecasts now call for losses of $34 million. I believe this is a good short because IRTC is a one product company that no longer enjoys being the only product on the market (increasing competition), losses are increasing with no articulated path to profitability and trades at an excessive valuation.  The stock trades at 11x forward sales; however, using aggressive assumptions for penetration for extended Holter patches, sales are likely to decelerate ~2020, at which point the multiple is likely to be more in-line with the group at 4-5x implying $41 stock (on 2020 rev estimates with no discounting) or 40% downside from here.  Additionally:

  • Valued as a medical device company when it’s a services company (where the device enables the service -i.e. lower margins, lower op leverage)

  • Bull case for the TAM to double is likely nothing more than a hope

  • Insiders and VCs are out

  • Studies have shown Zio to underperform on an apples to apples basis (Zio’s only advantage is length which is easy for others to copy)

What’s in the stock:

The momentum and sentiment factor for this stock is a 9-10.  People are excited for wearable health technology and IRTC is the only pure play and grew 53% last year and 43% last quarter with the multiple expanding to 11x sales from 7x.  The market is focused on sales growth and penetration and ignoring cash burn. The following is a summary of why the market is building in several years of uninterrupted growth.

Out of an estimated 4.8m procedures last year, Zio participated in 299k or 6% penetration.  It’s hard to estimate how much share Zio (or extended Holter patches) will ultimately reach but expectations are patches could get 40%+ and the sell-side and IRTC management think Zio will get the majority of that and are projecting smooth growth to get there.

Screen Clipping

But that’s only half the story.  The market is building in another 3m patients that aren’t included in the TAM.  These patients are asymptomatic, but given they have Afib, have a higher risk of stroke.  IRTC thinks that these patients should be monitored in order to measure Afib burden to decide whether they need anticoagulation drugs for stroke prevention.  The company spends a lot of time on calls discussing this opportunity and their recent press releases pump up this opportunity. These incremental 3m patience would essentially double the company’s existing TAM (existing TAM is 2.6m Holters, 1.3m Event monitors, 400k mobile telemetry)

The market believes that IRTC has a large competitive moat due, to great degree, to first mover advantage and resulting patient data (over 1.5m) for continued improvement of its machine learning algorithms to diagnose patients.  IRTC also provides the cardiologist with a report from its algorithms to save them from having to go through the data themselves, in addition to other back end services, such as claims processing. Bulls think this adds to the moat as doctors are already overworked and the report and back end services are another selling point.

Why is the bull case of doubling the TAM is unlikely?

I’ll answer this question with the question that many physicians at the recent Heart Rhythm Society conference were reportedly asking: Who would pay for it? I’ll show below why I think it is uneconomical to spend an incremental $1b to monitor 3m patients with no symptoms of Afib.  

There are 795k strokes per year which translates to a 0.25% chance of having a stroke.  According to the National Stroke Foundation, people with Afib have 2.5x greater chance of having a stroke.  A recent study funded by IRTC puts that number closer to 6x, leading to the likelihood someone with Afib will have a stroke at 0.60%-1.5%.  

The population of salient/asymptomatic Afib is~3m.  At an average cost of Zio of $300, this would cost $900m per year to monitor.  The average cost to treat a stroke (not including lost time from work) is ~$9,000 (we only care about ischemic stroke with is much cheaper than hemorrhagic stroke).  This would require that monitoring this population prevents ~100k strokes per year to break even or 3.3% of the high risk population; which works out to be over 2x the actual incidence rate.  So, to make this economical, the cost of stroke needs to double or the actual incidence rate needs to be 5x what we previously thought it was.  

While I understand this is a simplistic approach to analyzing the problem, and indeed the company is currently conducting a 3 year study to better understand the degree of risk of this population and economic impacts, the magnitude of the delta as illustrated above and sheer confusion the physicians had at the HRS conference  (according to Northland analyst) about this actually coming to life makes me think this is more of a hope than opportunity.

Why the company does not have a competitive moat and better products are coming

  1. First let’s start with the machine learning algorithms.  While being the first to market, the company was able to collect a large amount of heart beat data, but they have been very quiet in terms of the accuracy of their algorithms.  We have no idea what the false positive rate is, however the fact they have a team of technicians to go over each report to correct mistakes indicates it’s higher than they’d like.   Furthermore, there are no studies or data from the company illustrating how Zio changes clinical decision making. So far, this part of the story appears to be more sex appeal than an actual economic advantage.

  1. No true IP exists.  While bulls will point to the IP around the algorithms and patch, my understanding from a short conversation with a former systems engineer with the company is that no true IP exists since everyone uses R-R wave analysis, and the MIT database for algorithm refinement at  

  2. The Zio Report is not unique. Another claim from the bulls that insulates IRTC from competition is that the Zio report is a big sell to physicians as IRTC’s large investment in AI saves Docs work on the back end.  This isn’t true. The Bardy CAM, the other patch on the market today, also provides physicians with a report. Additionally, the topic of liability was also a growing one at the heart rhythm conference. It is my understanding that physicians don’t just trust the reports and make clinical decisions based on them due to potential liability.  Like I stated before, these algorithms are kept secret and the accuracy/methods are not disclosed so doctors understandably don’t take them at face value and instead use them more for a starting place.

  3. The benefits of Zio’s monitoring time of 14 days are overstated.  The company talks about this chart (below) a lot showing the higher diagnostic yield of the Zio.  They point to all the IRTC funded studies which supports the high diagnostic yield of the Zio. However, diagnostic yield is simply the number of arrhythmias detected.  A 14 day device vs a 48 hr device better detect more arrhythmias. What they don’t disclose, and recent studies are starting to show (the first one came out this month, more below), is that the Zio has a lower diagnostic yield to other devices on the market on a like for like basis. Additionally, 14 days is a long time for a high risk patient to be monitored with no data going back to the doctor for a full 21-30 days. What if the patient has an emergency arrhythmic event in that time?  Importantly it takes another 7-10 days after the monitoring period for the physician to receive a report due to the process of having to ship back the patch to IRTC, for IRTC to upload the data, run analytics, QC check on the analytics and provide the physician with a report.

Screen Clipping

Furthermore, I have not had the chance to confirm this, but based on the data (chart below), I think the 7 day monitor time for the Bardy CAM is a good balance of diagnostic yield, potentially lifesaving time saved, and money saved.  As illustrated on the chart, the vast majority of the incremental arrhythmias detected beyond the 2 days of a Holter happens by day 8 (diminishing returns). Leading to the question, is 14 days necessary and if so in how many cases? If diagnostic yield is the biggest concern than why not use an MCOT with a higher yield (absent from chart above, 88% diagnostic yield) which can transmit data wirelessly at the time of arrhythmia?

Screen Clipping

More on the competitive environment

  • Bardy: We have discussed only Bardy so far given they are the most near-term threat, but there are others discussed below.  Bardy performs better on a like for like basis. A recent study published just two weeks ago (first head to head study for the same time period) in the American Heart Journal shows the Bardy CAM “identified significantly more arrhythmias and resulted in better, more informed clinical decision-making over the iRhythm Zio® XT patch.” Interestingly, the Zio also underperformed when compared to the legacy Holter on a like for like 48 hour time frame.  The company is small with only 10 sales reps (per linkedin) compared to the 66 IRTC had entering 2018, but it appears they are growing the sales force fast. They have doubled it in the last 6 months (per linkedin) and there are another 6 openings on the website.  I was able to speak with the Chief Commercial Officer at Bardy who told me they are going on a “hiring spree” after this study was published.

  • BioTelemetry (BEAT): Beat is the largest competitor and sells a whole host of products ranging from Holter monitors, to Event monitors, to MCOT’s.  They have developed the ePatch (no studies yet). Importantly Apple partnered with BEAT as part of the Apple Heart Study in Nov 2017.  It is my understanding that one would get a notification if the AAPL watch detects an irregular heartbeat/arrhythmia. The person would then call American Well (TDOC competitor), and a doctor would prescribe ePatch if appropriate.    IMO the leverage BEAT can get off the AAPL brand along with the potential size and scale of the trial and resulting ability to drive increased awareness around ePatch is underappreciated. Additionally, per the most recent Q1-18 conference call, the reason BEAT isn’t more aggressive in this market is due to the temporary CPT codes and associated reimbursement risk and indicated they will participate in a big way when there is more certainty.  Furthermore, BEAT already has the sales force, customer relationships, distribution and complimentary products that should allow them to scale fast.

  • National Cardiac. National Cardiac, a private company based in CA, has developed an interesting patch AECG called Liba3. The Liba3 is a multi-function 4-1 monitoring device that includes Mobile Cardiac Telemetry, Wireless Cardiac Event, Holter, and extended Holter monitoring in a wearable patch. This device is not yet cleared for sale.

  • Icentia: Incentia has developed the CardioSTAT, a 14-day patch monitor that supposedly achieves low impedance through an electrolytic gel, therefore purporting to obtain high signal quality for ambulatory  ECG monitoring. The company announced in 2016 that the CardioSTAT was chosen to be the patch monitor of choice in a NeuroVision study being jointly conducted by the McMaster Hospital in Canada.

  • VitalConnect has developed the Vital Patch single-lead ECG monitor that continuously monitors ECG activity for 4- days. The Vital Patch is 510k and CE mark cleared. The level of sales from Vital Patch is not clear.

  • Preventice Technologies, based in MN, is a private company that has developed the BodyGuardian Heath patch ECG monitoring system. The BodyGuardian Heart Monitor wirelessly records ECG data and connects via a smartphone to the Preventice Patient Platform. The patch monitor includes patient triggered and device-detected episodes. It has been jointly developed by the Mayo Clinic and ST Microelectronics. Even though the device is 510k cleared, I don’t have an idea about the current level of commercial sales or adoption rates.

Lack of opex leverage

As illustrated on the chart on the left, the company posted some operating leverage ahead of the IPO but began deleveraging again in 2017 and more so in Q1-18.  The company blamed a growing sales force; however, I think that explains only a small piece of it as sales productivity was flat in Q1 at 1.3m per head and increased in 2017 from 1.0m/head in 2016.  The company has limited disclosures so it’s difficult to understand where the lack of leverage is coming from(no unit disclosures, opex breakout etc), but I think it has more to do with them being a service company and not a device company.  For example, on the Q1-18 earnings call, management cited increased demands from customers on the back end which I view as an underappreciated expense. The company and analysts tout the Zio report that physicians get, and thus limited work the physician needs to do, as a competitive advantage; however, it seems the analytics and Zio patch is not enough for cardiologists to adopt the product/service as IRTC is now doing other back end work such as claims processing to further incentivize health systems to adopt it.

It’s difficult to understand where the leverage will come from given the variable cost nature of service companies.  The largest leverage I think the company can get is on sales rep productivity. Management thinks it can ultimately get to $2m per head and bring gross margins to 75-80%.  Another leverage point would be on R&D but IMO, not much as they are at 13% vs peers at scale of 10-17%. Giving management the benefit of the doubt and assuming 80% gross margins, $2m per rep, 10% R&D margins, 2017 EBIT% would have been less than 10% and using more base case assumptions would have been less than 5%.  


The company trades like a med tech/device company when in reality it is a services company and its margins at scale will look more like a services company.  Given that it is valued on a P/S basis, it is possible that it can grow into its valuation; however, using aggressive assumptions for penetration for extended Holter patches, sales are likely to decelerate ~2020, at which point the multiple is likely to be more in-line with the group at 4-5x implying $41 stock (on 2020 rev estimates with no discounting) or 40% downside from here.  If it still commands a Med tech multiple of 8x in 2020, the stock would be at $76-$77 or 10% upside.  

Ways can be wrong

  • Current level of service (analytics/report, back end claims processing) is what makes the model non- economical; thus private competitors that care about profits don’t follow suit and thus don’t take much share and IRTC growth doesn’t slow.

  • The market continues to agree with management that revenue growth and share gains matter and profits don’t.

  • Grows into valuation: management can continue to push the TAM opening narrative for high risk, asymptomatic patients for another 2-3 years as the 3 year mSToPS study is only one year complete.  I believe it will ultimately be a hard sell to payors which may not be obvious to the market for another 2 years

  • Machine learning algorithm automates back end resulting in operating leverage  


I do not hold a position with the issuer such as employment, directorship, or consultancy.
I and/or others I advise do not hold a material investment in the issuer's securities.


market stops being so frothy...

    show   sort by    
      Back to top