Description

Semler has been written up twice before on VIC by rhianik and those give a good background on the product and market. Basically SMLR has a much simpler and lower cost way to detect Peripheral Artery Disease and the test is seeing increasing adoption, including for asymptomatic patient screening which will grow the market tremendously.

The stock shot up to an enthusiastic valuation and is now down by about 80% from all time highs. I calculate the true value to be 2x-5x the current price and have chased down many of the potential concerns to judge whether the product and company have a strong future.

During any rapid drop from bubble territory a large number of bearish theses arise. I tried to prove each one right, and will mention why each bearish thesis appears incorrect.

Value proposition

How medical groups get paid

Many medical groups these days get paid a capitation rate that pays a certain amount of money per person whose medical care they handle. This rate rises for people with known medical conditions, such that the doctors get more money for caring for a patient with worse conditions at a rate statistically calculated to cover expected additional care while motivating the doctors to keep the patient healthy to minimize actual required care. If your doctor has ever seemed overly eager to add new conditions to your medical record, this is why. 

Thus the most profitable patient is one with a documented medical condition that is easily cared for and minimally disruptive (more pay, least care). Conversely, the least profitable patient is one with an undiagnosed condition with serious consequences (no more pay, but more care turns out to be needed).

Now let us consider Peripheral Artery Disease (PAD), a stiffening of the arteries associated with high cholesterol and/or diabetes which some of you dear readers probably have (statistically speaking). PAD is associated with a dramatic increase in mortality, but is often undiagnosed. The test for PAD, before SMLR and their QuantaFlo devices, requires enough specialized skill and equipment that it has historically not been applied until after symptoms are reported. However, population studies of undiagnosed patients over 50 have shown that when the entire population is tested around 30% of the tested populations have undiagnosed PAD. Further studies following those patients have shown that the 30% with previously undiagnosed PAD have just as much medical risk as those who were diagnosed due to reporting symptoms. This means that around 30% of a given doctor's older population patients may fall into that “least profitable” category - they have a risky medical condition but the doctor isn’t getting paid for the work they will likely need to do.

Semler scientific’s QuantaFlo product is a cheap, easy, and quick screening test for PAD. Small light emitting and receiving clips are placed on two fingers and two toes. A couple minutes later the test interface on an ipad or laptop gives a PAD evaluation and the doctor gets a readout on whether the patient has PAD, doesn’t, or is borderline.

Bear Thesis: “The test is too simple to be real. Lots of small medical device companies survive for a while by gaming reimbursement even if they don’t work, and that applies to SMLR”  - Honestly this was the Bear Thesis I really expected to win. The Semler devices appear so simple. I’m an engineer by original training with experience working with medical devices so I dug into the Semler technical details with skepticism. The key Semler patent describes things pretty well, the optical backscatter is analyzed for Spectral Flatness Measure (SFM) and 1-SFM is a circulation index. This seems like a plausible way to measure elasticity (or the lack thereof) for circulatory vessels since stiffer vessels will flex less and thus provide more spectral flatness in the return optical signal. I was surprised at the cleverness. Additionally, study after study comparing QuantaFlo to other evaluations of PAD have shown that QuantaFlo is at least as good a test as ABI (Ankle Brachial Index) testing, the previous existing “definitive” test for PAD.

Bear Thesis: “Can doctors even get compensated for that test? It could take forever for insurance to allow such a test.” - We dug deep on this one, working through billing systems from multiple angles. Anyone familiar with medical payments will not be surprised that the answer appears to be ‘it’s complicated’. Some payers will NOT pay for a PAD test if the test is done without the particular patient having some reason for the test: leg pain is the classic example (and a classic PAD symptom). Other reasons could include risk factors like Diabetes. Notably, some of the doctors we talked to didn’t care about getting reimbursed for the test - at a 30% positive rate within the target population the value of increased capitation payments is far more valuable than the cost of the tests (at about $40 per test - less in bulk). Other payers WILL pay for the PAD test. Semler has settled on two forms of payment they allow: fixed monthly fee or per-test fees. Once we started exploring how compensation for the test works this made sense. Some users get paid ~$80-$150 for doing each test so they want to pay Semler per-test as it gives them known operating economics (large customer Signify Health is in this category for their home “Evaluations”). Other medical practices want to pay a fixed price as a known total monthly overhead and then screen nearly everyone to increase average capitation rate - doctor-owned medical groups appear to be the core of this user type. One of the doctors we talked to said they not only don’t get paid for QuantaFlo testing but they have to send positive PAD QuantaFlo patients for ABI testing because that is the only form of PAD diagnosis some insurance accepts, and they still want Quantaflo because the minor cost is still hugely rewarded at the capitation payment level. 

We consider complications around insurance company acceptance and large practices having pre-existing workflows as the main growth rate limiters for Semler. Those are real barriers but they both fade with time so it bodes well for the long term while cautioning that the initial burst of growth may slow.

Finally, we think the main barrier to QuantaFlo growth is just medical group inertia. Many larger groups already have “an ABI person” with a dedicated room for the equipment and office staff who know how to code the tests for reimbursement. I talked to one doctor who just commented that if she wanted a PAD test she’d just send the patient to “our ABI guy”. This is standard adoption curve behavior and suggests the better option (QuantaFlo in this case) will make market gains over time.

Medical Trends

Previously, PAD testing has typically been applied when clinically indicated - for example due to patients reporting leg pain. There is still a bit of overhang in the medical community where doctors expect PAD should only be tested when there are symptoms because the testing requires a specially trained person and equipment. The newly developing consensus is trending towards viewing PAD testing as a beneficial screen for anybody in appropriate risk groups such as those over 50 years of age. This had not typically been considered practical due to the amount of effort required for ABI testing (the previous test standard).

Bear Thesis: “PAD testing isn’t really needed if not clinically indicated.” This is the view taken by a recent report that made a headline out of the fact Signify Health is using QuantaFlo on everyone over 50 in some areas, not just on those with symptoms. First, even if QuantaFlo is only used when clinical indicators of PAD are present they still have room to grow (at a lower rate than historical). More importantly, the headline mistakes a change in care standards for something sinister - PAD screening is becoming a standard. Even WITHOUT QuantaFlo (using the much higher effort ABI test) Johnson&Johnson has launched one example of a mobile cross country screening effort testing even those without symptoms. Meanwhile mass-screening studies have indicated somewhere between 40-75% of PAD cases are asymptomatic but detectable by QuantaFlo/ABI. Even when asymptomatic, PAD cases should be treated because asymptomatic PAD raises morbidity risks as much as symptomatic PAD as far as can be determined statistically. Thus the increasing use of screening (testing all patients in a risk group, even those without symptoms).

Valuation

Recent financial and stock price performance 

The recent “bubbly” maximum valuation was during Oct 2021. Let’s walk through the market narrative:

Figure: Stock price versus recent revenue performance. Green and Orange lines are revenue rate for fixed and variable fee testing, respectively, (left Y-axis). The Gray line is the number of home “Evaluations” reported by Signify health, believed to be SMLRs biggest variable fee customer (right Y-axis). The blue shaded overlay is COVID cases in the US (arbitrary vertical scale).

As with many companies, the initial drop to near zero revenue on COVID shutdown was forgiven by the market and when testing came back strong the shares rallied hugely. The third quarter of 2021, however, shows a slowdown in variable fee revenue which popped the price bubble at the beginning of Nov 2021 . The bad news continued to the end of the year (the following dot) and the stock kept falling. Having lost its momentum, SMLR stock has bumped along since then.

Looking at the chart above with a skeptical eye we can see several features. The first is that the downward bumps in the orange plot (after the shutdown) are magnified versions of downward blips in the gray plot - reduced Signify patient evaluations caused bt COVID case rate climbs lead to reduced variable fee testing - unsurprising since QuantaFlo is a test that requires contact. The proportional drop in 2021 variable revenue is larger in proportion than Signify’s drop in evaluations, but that seems expected considering they could still be doing preplanned Evaluations with a noncontact protocol that skips certain tests. Applying a bit of math, it looks like Signify was doing Semler tests at about 31% of their home Evaluations during 2021 until the COVID spike when it dropped to a low of 18% and has already recovered to 26%. The 31% number suggests only some Signify Clients are currently using QuantaFlo.

Bear Thesis: “Semler’s variable fee income is so choppy and uncertain!” - Yes, yes it is. I am comfortable with what I see here in the details in terms of reasons for that volatility. COVID restrictions will not last forever (or people will adapt). The things I’ve learned about the practicality and usefulness of the test and the medical need also reassure me that the long term should be a good story, but I fully expect the variable fee line to be choppy.

Another detail we can see here is that the fixed fee revenue has been steady. Medical groups did not cancel their subscriptions.

One can estimate the number of actual tests done based on an average price of $40/test for the variable fee business and assuming the fixed fee folks are doing more tests per dollar than that - otherwise they would save money switching to variable fee. Such a guesstimate puts us in the range of maybe 1mm tests in 2021. Medicare advantage alone has 28 million participants. Even testing ONLY medicare advantage patients biannually gives Semler a LOT of room to grow.

Since the March 2022 drop below the $70 dollar range there have been no insider sales. Even share awards to directors, which are commonly sold quickly for tax reasons, have been hoarded by insiders instead.

Looking at the price action since March shows long stable price plateaus. That’s exactly what I see when I’m spending a month or so building a position that exceeds volume by a lot, but this time it isn’t all me. Someone else is onto this too :)

Discounted Cash Flows - What should the share price be?

I typically only consider DCF calculations as a disqualifier or rough guide to valuation. There is simply too much unknown to rely on them heavily or to expect them to equal the valuation of the stock market in the near term. With that in mind here are some scenarios for consideration and thought. (For share count, the current shares are 6.855 million but if all options authorized are issued and executed the fully diluted total will be 8.139 million.)

Base Case: 13% growth for ~ 10 years (compare to 17-70% historical annual rev growth range since 2017). Margins stay the same (historically they have been improving). Final patient share works out to something like 20-40% of just the medicare advantage market in 2031 depending on how often the screening is done. Share price in this scenario is something like 4-5X too low depending on how you prefer to treat the cash on the balance sheet.

Lower Growth Case:  Growth drops to 5% (which would be really low for this company, product, and sector). Current value is more than 2X market cap

Ultra low growth and lower margins: In this case long term growth drops to 3%, roughly the long term GDP growth rate and the FCF conversion drops by 25%. And the stock would STILL be worth ~50% more than today’s price

Other Things

As always, I’ve made this too long already….so, quickly

I skipped some Bear Arguments I’ve seen that I found less substantial but feel free to ask anything in the comments. 

They sold a bunch of shares of Mellitus to the SMLR founder, then signed a distribution agreement and prepaid Mellitus a few months later. Like, really? 

Main product patent expires in 2027 but there are reasons that it will take a while for competition to matter.

C.R. Bard (SMLR distributor) reported an investigative subpoena in Nov 2015 covering SMLR product marketing as part of a list of topics, in July 2017 a separate civil investigative demand was served. Both appear to have been resolved without material incident in 2020.

Catalyst

Catalyst: As COVID impact fades over time and adoption grows, the value will become clear.