LVGO - Livongo

After reading this thread, Bert’s article yesterday, and the Seeking Alpha by Richard Chu, I was ready to pull the trigger on an initial position in LVGO, swapping out some of my MDB holdings.

However I spent some time looking further into one of the links provided by the OP, as well as a further link in a comment to Richard’s SA article, as I am a naturally skeptical person. I append these below for ease of reference:

https://validationinstitute.com/the-livongo-study-cora-pro-i…
https://dismgmt.wordpress.com/2019/08/26/are-livongos-outcom…

Following the links I found yet another article suggesting deception about patient outcomes and savings to payers by (digital) wellness companies, including competitors to LVGO such as Omada.

https://validationinstitute.com/the-diabetes-prevention-indu…

My education is in mathematics (though not statistics), and I found the arguments in all these links logical and compelling.

Much, if not all, of this originates from the Validation Institute, which is “an independent, objective, 3rd party resource for health vendors and purchasers.” It was acquired in mid-2018 by the Health Value Institute, allowing them to “expand business operations into validation, consulting, in-person and online training and certification of the HR, Benefit, and Wellness executives.”

Possibly many HR executives aren’t correctly discerning between peer-reviewed studies and savvy marketing campaigns, especially when one is disguised as the other. It could be that LVGO is the first mover on the digital wealth app gravy train because they see an easy sale and are moving ahead with a product that has not (yet) been proven effective.

I tried to find other peer-reviewed studies of LVGO’s effectiveness but to no avail. I’m still open to buying LVGO (the numbers look great!) but wondering if anyone else has looked into this.

28 Likes

I tried to find other peer-reviewed studies of LVGO’s effectiveness but to no avail. I’m still open to buying LVGO (the numbers look great!) but wondering if anyone else has looked into this.

I would suggest that it will take some time (probably years) to determine any effectiveness and monetary value. As I understand it, LVGO’s product is attempting to change health behaviors. And habits don’t change overnight. By the time effectiveness is determined, LVGO’s products may already be widely adapted and hard to displace. Even if its effects are modest, users may be loathe to give up any perceived benefits. Honestly, much of what we do in medicine has only modest therapeutic value. But we do it anyway, despite the cost. So even modest benefits may be much better than most other medical interventions.

4 Likes

Tchalla,
I read Bert’s article and MF interview so I am interested. Does the company report any member retention or churn numbers? Also is there a way to tell what % of the 208K members who have signed up are actively using the system? Do they report average monthly strip use for example? If the churn is low and members are actively using the product it augers well for the business.

2 Likes

“If the churn is low and members are actively using the product it augers well for the business.”

I’ll have to find where I read it, might have even been Bert’s article, but I did read that their churn is very low. The only churn that they have is not people stopping the use of the product, but people no longer employed by the company paying for it.

In other words their retention rate is very high. If I find the article I’ll share the numbers.

TMB

2 Likes

5thhorseman,

I believe the author of the following article does a great job of breaking down the concerns posed by the Wordpress article:

https://www.valueinvestorsclub.com/idea/LIVONGO_HEALTH_INC/1…

Refer to comment 8 left on 10/09/19.

The VIC author seems to have a direct connection to Livongo management and was able to inquire about the article. It seems Livongo is aware of this blogger and has dispelled concerns raised by 3rd party consultants in the past. Furthermore, their direct checks with current clients confirmed that the vast majority saw lowered total healthcare costs among their diabetic population. I do not believe Livongo would be able to acquire the Federal contract nor maintain such strong relationships with their channel partners if their effectiveness could be easily called into question. If you look at their very positive App Store reviews, it seems like it’s really changing people’s lives.

Livongo’s solution has never been done before and there is definitely an uphill climb to mass adoption but the financials thus far show strong execution. A platform like this should not work for everyone, it takes a great deal of self-motivation as Bert mentioned. I think most people are initially drawn to the platform and stay on because of the large cost savings, giving it a chance to collect data and personalize health nudges and for their personal coach to get to know them. Oftentimes these nudges are tied to a reward as well, some clients offer no copays on insulin if members measure their glucose levels when they are asked to. After a while, members really start to realize the health benefits of the platform, something that health plans without these benefits and technology failed to do in the past.

Livongo’s founder, Glen Tullman, also won the Ripple of Hope award and I believe it hard to question his ethics as that honor is not given out lightly. He has a deeply personal reason, his own son who was diagnosed with diabetes, for wanting to create change with Livongo. You can also check out his book: On Our Terms: Empowering the New Health Consumer.

I agree that Livongo’s long term benefits are not yet proven but I do see a first-mover disruptor that is executing well within a very large TAM with no incumbents.

Long LVGO

11 Likes

Not just members churn but % active members is important. Frequently members may sign up if their company offers something for free but do not use it effectively. In time companies may realize that this is not working (i.e. their healthcare costs are not dropping as projected) and may discontinue the program. That is why it is useful to know active member use.

Also anyone know more about their AI+AI program? From Bert’s article:

It is this AI+AI engine, one of the key differentiators in terms of what Livongo does and how it does it. It aggregates data from multiple sources that use currently available devices that can be operated by patients, it interprets that data to provide a context and applies the data by providing users (members) with recommended therapies. The data is then iterated to build improvement based on specific data for individual payments. Simply put, it is this AI+AI paradigm that is the company’s differentiator and moat.

If my blood sugar is high it is going to suggest some predetermined remedy based on the numbers I suppose. What will AI+AI do better? I am curious to get Saul’s take given his background.

1 Like

Very much appreciate 5thhorseman’s post of healthy skepticism. Thank you for for sharing that research.

But i find remmdawg’s point, along with Livongo’s very high retention rates, to be persuasive in Livongo’s favor:

“Even if its effects are modest, users may be loathe to give up any perceived benefits. Honestly, much of what we do in medicine has only modest therapeutic value. But we do it anyway, despite the cost. So even modest benefits may be much better than most other medical interventions.”

One interesting example is the explosive growth of the CPAP machine for sleep apnea. The ResMed stock has been on an impressive run as rapid growth continues for selling the Medicare approved expensive supplies and equipment. Therapies clearly demonstrate reduced breathing events during sleep, but there is not a shred of evidence that my specialist doctor knows about that indicates reduced events for heart attacks and death. Moreover, CPAP retention rates are relatively poor at under 50%, certainly compared to the +90% for LVGO services.

But for early stage companies, there is increased reliance on the integrity of leadership. In that regard, Glen Tullman’s life tells me that LVGO is a good bet.

Disclaimer: i have been burned before when highly capable CEOs turned out to be crooked. Early in this century, i mistakenly believed in the integrity of CEO Dennis Kozlowski, whose excesses sent him to prison, and sent my investment dwindling 80%.

LVGO numbers look outstanding to this point and the prospects look even better to me. It could still go bad, but right now, i rate LVGO up there with AYX and RNG.

GLTA

5 Likes

Hi Deviant, thanks for the additional link. It contains a lot of good information even beyond just post #8. As per that poster, it seems LVGO did, at one time, make claims about their product that were perhaps unjustified and which they have subsequently toned down:

Regarding the questions around reduced inpatient admissions … To be clear, we are still waiting on a more direct answer from the company but as your blogger states, “in all fairness to LVGO, they don’t promote the inpatient admissions result anymore.

We think it is arguable whether the statement “Livongo for Diabetes program delivered an $88 per member monthly reduction” is mis-leading in suggesting causation and, frankly, is a minor semantics issue in light of the savings that clients are reporting. LVGO has only used this language in a handful of press releases, and we haven’t seen any more-aggressive causal language from the company.

That said, I agree with your assessment and have taken an initial position today. I’m wary of the extended lockup period and will by buying on dips as I continue to read up on the company.

Generally, I think it’s worth being skeptical of both sides of any issue and I think in this case the truth probably lies somewhere in the middle.

Thanks also, addedupon, for your comments.

5 Likes

Hi Everyone!

I had some time to take a look at the skeptics view of LVGO. I didn’t find the blog author too credible or the statements to persuasive. Some of what I wrote was echoed in earlier posts and links, but I wanted share my thoughts with you. Having published over 6 peer reviewed articles during grad school, publishing is hard and the authors made it past the peer review process. Sure, relying on any one study is problematic, but I found the design convincing and well executed. Are they biased since they work for LVGO, absolutely. But that bias does not make the study invalid, instead we should review carefully and look for other studies that support this work.

Point 1: Publishing in a journal with 2 as impact factor.
Impact factors measure the average number of citations per article per year. This is a rough proxy for journal quality as higher cited articles are more influential then lower cited articles. However, the blogger has clearly never published. Of course everyone wants to publish in better journal, but not all research makes it in there. As a researcher, you have to balance prestige vs. time. Sure, they could have sent it to JAMA and it would likely have gotten rejected along with hundreds of other articles. Then they would have had to send to NEJH, wait another few months, and done that till the article landed. This could take YEARS! Basically, this is non-sense. My guess is things that are not based on a randomized controlled trial have a very hard time getting into these top journals. There are always exceptions, but that is my guess.

Point 2: Interpretation
The blogger claims that the study does not use causal language. After reading the study, the authors use three techniques (difference in difference, instrumental variables, and matching) to conduct the analyses. Both of these are called selection on observables (SOO) and proxy causal quantities of interest. Yes, this is not a perfect causal design, double blind randomized control trial, but given the application, using the two methods they used is as strong as it could have been. Since its a medical journal, it is probably harder to publish anything outside a randomized controlled trial as causal. Using the staggered rollout of the study is creditable. Of course there are issues with any design and lots of assumptions, but I am confident they have a strong quantitative analysis. They also do a few tests to show that their finding is not the artifact of one test where they just got lucky.

Point 3: Didn’t have the right outcomes
The blogger claims that they would know how to correctly design a diabetes study and the study authors measured the incorrect things. We could sit here all day and fight over this. Would it be good to have a study that looked at insulin use, 100%. Is the current study 100% invalid because it didn’t measure this, NO.

From an investing perspective, this brings up an interesting point. Companies are the one’s paying for this (that’s good for us). Of course they want their employees to be healthy, but one goal of LVGO is to reduce costs for employers. So if that is what they want to show, then their outcome is just fine. Here they showed cost is mostly reduced by lowering office visits, with diabetes costs coming in second. I think the reduced office visits is a key insight here since it is 1) very inconvenient for the employee 2) very expensive and 3) lowers productivity of employees.

Point 4: Authors don’t understand diabetes
I am not a MD and incapable of making any insights here. What I do know, it reading one study and using it to evaluate another study is not really how science works, but hey this guy can do what he wants.

I’ve taken a small starter position in LVGO and will continue to watch.

B

19 Likes

Hi all,

I started lurking about a month ago and figured I’d jump in here because I have an answer for TexMex about the AI-AI. There’s a white paper on Ai-Ai here: https://www.livongo.com/applied-health-signals/whitepaper.ht…

I’ve had LADA/T1 Diabetes for a year and have gotten it pretty well controlled using diet and insulin. Thanks to reading some great books and a Dexcom G6, which is a terrific device…when it works…a PIA when it doesn’t. But I’d rather have it than not; I’ve come to rely on it for peace of mind. When you use this continuous glucose monitor, you don’t need to test often, just in the first couple days until you’re comfortable the cgm is calibrated properly. Despite what Dex says, most people on my 4 FB support boards need to calibrate. Once that’s done though, you don’t really need to stab your fingers for blood, which is great. My point…the fact that Livongo recently announced a partnership with Dex is really good news…the data Dex already collects will be passed along. Data points every 5 minutes, not just several stabs a day. This will build up Livongo’s data in a big way and give their AI more to make better decisions.

People who don’t have a cgm will be thrilled to have access to as many strips and lancets as they need. Doctors’ perscriptions don’t allow for as many as you really need if you’re going for good control. Testing 2-4 times a day is ridiculous; for good control, you’re talking 8-10x. This is why cgms will, hopefully, and soon, be given to all T1s and T2s. Personally, I think cgms are the big game-changer.

I polled my boards to ask if anyone was a Livongo member. So far only 2 responses, but both said the testers were excellent. Having a good one is crucial.

However, they’ll need to do a lot more to tailor their nudges if they want to really succeed with the T1 crowd (small % vs the T2s though). The nudges have been very generalized and both people have ignored them because the advice didn’t apply to their diabetes. One person finds the nudges really annoying; the other doesn’t even get them now…she opted out of their texts. If someone’s already under control, they won’t listen to a company (sometimes not even doctors) for advice. All 3 of us could see how this would be useful for those who are not under control but want to be.

And if the member’s company is offering no copays/free insulin, done deal! That’s HUGE, with the cost of insulin. Out-of-control T1s or those who don’t subscribe to low-carb/keto diets go through a lot insulin! People will comply for free insulin.

I can see how Livongo’s program could help many people, more so those with T2 and hypertension or those who want to lose weight, thereby reducing company healthcare costs.

27 Likes

Fafar… excellent information based on your experience… thanks for sharing…

However, they’ll need to do a lot more to tailor their nudges if they want to really succeed with the T1 crowd (small % vs the T2s though). The nudges have been very generalized and both people have ignored them because the advice didn’t apply to their diabetes. One person finds the nudges really annoying; the other doesn’t even get them now…she opted out of their texts. If someone’s already under control, they won’t listen to a company (sometimes not even doctors) for advice. All 3 of us could see how this would be useful for those who are not under control but want to be.

Yes, my impression is that LVGO offering is well targeted to T2, and thats a larger population that is more prone to drift from a regiment and therefore, quick expert nudges are a useful way to get it in control.

On another note, would appreciate any insight or thoughts you have / share about pumps and among them TNDM vs PODD vs Medtronix products. Whether you use them or learnt about it from others.

And if you prefer to use Dexcom CGM over Medtronix, would TNDM pump become automatic choice?

thanks
nilvest

Hi Fafar, sorry for the format challenge in my previous post… hope the request makes sense to you.

ANd yes, congrats on being able to control T1 condition, please keep it up.

nilvest

Regarding churn

From the S-1
Factors Affecting Our
Performance—Product Intensity and Enrollment Impacts our Performance” below. We also closely measure member retention, and our average monthly member churn for 2018 was approximately 2%.

We calculate our monthly member churn by looking at the members who were with us at the beginning of each monthly period and then subtracting the number of those members still on our solution at the end of each monthly period and dividing that number by the starting member number for that monthly period. To get our average annual monthly member churn, we take the average of all twelve months of churn

but it is touched on on Q2

https://finance.yahoo.com/news/edited-transcript-lvgo-oq-ear…

and Q3 commentary

https://www.fool.com/earnings/call-transcripts/2019/11/07/li…

Further commentary from the President and former CMedOfficer on validation of clinically important outcomes

Further abstracts and publications are in the works

https://www.youtube.com/watch?v=Qpf8fA5DcaQ

https://www.livongo.com/impact.html#Clinical_Outcomes

1 Like

You have confirmed me in my intuitive reaction to the blogger’s writeup. He just did not come across as rigorous in his analysis (such as it was) or believable. I have seen many such writeups, and have had to judge a few.

arnie j. long LVGO a bit more than minimal.

Ok I have bought onto the LVGO bandwagon and taken a 2.5% position. Things that got my attention:

  1. AI/AI framework is not just marketing speak. There is a thought process behind it. In their white paper they even have 3 examples which makes sense.
  2. Client retention is 96%, member retention at 90% which is not bad
  3. You can divide people into 3 types - disciplined ones who will take care of their ailment, the ill disciplined ones who will don’t care, and the people in the middle who need constant reminders. It seems when a company offers LVGO and the strips for free the company probably loses money on the first group because they are anyway committed to taking care of themselves. The other 2 groups may respond to nudges and definitely if the company offers insulin copays which is tied to the member doing regular testing.
  4. Company has done 48 studies and published a lot. They are by far the ones with the greatest #. I found a SA article by Mr. Chu which contrasted them against their competitors. Unfortunately it is locked now. He said one of Google’s companies is a competitor though much smaller. Sundar P. made a reference to that in the last EC it seems. Makes me feel good about the industry.
  5. Lot of details in the presentation below. One recent stat. They have 208k members and 770 clients which is about 270 members/client. They just signed up the feds and NJ state each of which they estimate should bring in 1000s of members.
  6. I did not realize that 60% of the population has a chronic condition (they could address that population) and 90% of health care spending is directed at chronic diseases. Seems like this has the potential to reduce health care expenditure.

https://seekingalpha.com/article/4317070-livongo-health-lvgo…

14 Likes

Hi Nilvest,

On another note, would appreciate any insight or thoughts you have / share about pumps and among them TNDM vs PODD vs Medtronix products. Whether you use them or learnt about it from others.

And if you prefer to use Dexcom CGM over Medtronix, would TNDM pump become automatic choice?

I don’t use a pump yet, so no personal experience here. Every day, though, there are people on the boards asking for help on which way to go and MANY, sometimes over 100, people respond every time. It’s a supportive community. But it’s all opinion-based and there are definite pros/cons to each pump.

I’d say that right now most respondents seem to favor either Insulet’s Omnipod’s tubeless pump or the Tandem’s T-slim right now. Biggest issues: the tubes can occlude (bend/compress,no insulin pass-through, glucose skyrockets) and many people just don’t want a tube hanging off of them. Pods, which are little storage tanks, can leak (and there goes all your insulin). A lot of people moved away from Medtronix, apparently having trouble with it, but there are still fans…When somethings works, you’re grateful! I don’t think anybody on my boards is using the Roche Acu-check pump …I’ve never heard it mentioned and didn’t even know they made a pump til recently. And I use their very-accurate tester.

But … Closed-loop is coming!

https://asweetlife.org/new-diabetes-tech-to-look-out-for-in-…
If the newer systems are more user-flexible and work properly, this will be huge for T1s. Note that the Omnipod solution will also connect to the Dex G6. It would be great if all the pumps could connect to any cgm.

The existing Medtronic 670G closed-loop system has not been a hit so far with my boards. There was one set glucose target of 120 for everybody. I shoot for low-to-mid 90s myself. Dr Bernstein followers shoot for 83. Many would be a-ok with 120 though; for them, it would be a huge improvement. There have been technical problems causing disuse though and that’s talked about in the above article too. I read somewhere that it kept users in-range 10% more than user-directed pumps. <meh, they’ll have to do better than that for me to be impressed>

CGMs
Everyone is waiting for the new Dex G7. It’ll be smaller and hopefully stay in longer than 10 days…and I hope generate less trash. Current applicator is large and non-recyclable. They are not going to, but should, work on their algorithms. The Dex is preferred for their alarms and less need to fingerstab. The article says Dex will refocus use of the G6. Bet they’ll really push to have T2s use them…enormous market…they’ll have to convince more insurance companies and Medicare. Doctors already want their patients to have them so no convincing there. Some insurance companies make you submit several months of dietary logs to show that you’re trying to control. Then turn that around…some board members got denied because their logs showed they were pretty controlled already!!! (It pisses me off just typing that…it’s outrageous…it’s really hard to stay controlled, especially at night, and that’s when people really rely on the alarms. T1s die in their sleep a lot, apparently. Oh boy.)

ABT’s Libre works for a lot of people, lasts 14 days now, is already small, is cheaper, but doesn’t give alarms. There are 3rd-party apps to get around that though and one will even call you or your backup person when you’re having a severe low. Newer versions will alarm, I’m sure. So they’re not out of the running.

If I had to choose between a cgm and a pump, I’d pick the cgm. You can always give yourself shots. Most T1 diabetics lose the ability to tell when their glucose is low. It’s just, bam, you’re incoherent and on the floor all of a sudden. That’s how crucial the cgms are.

PS: Only heard from one more person about LVGO and she too liked the free tester/strips and mentioned that there’s a good report to give your doctor, but didn’t mention whether she liked the nudges. I doubt there will be more responses at this point.

6 Likes

TexMex,

He said one of Google’s companies is a competitor though much smaller. Sundar P. made a reference to that in the last EC it seems. Makes me feel good about the industry.

You’re most likely referring to this: https://medcitynews.com/2019/01/dexcoms-first-gen-cgm-sensor…

LVGO will definitely have competition and Dexcom looks to be one:

“G6 Glucose Program” (only for Samsung J3 phones so far). This app is intended for people who are not at risk for severe hypoglycemia, offering a different version of the currently available G6 app. The G6 Glucose Program will offer in-app chat with a wellness coach and removes the high or low glucose alerts. Dexcom has not said when it will launch, though more pilots of CGM in type 2 diabetes are expected in 2019 if all goes well. from: https://diatribe.org/whats-coming-dexcom-2020-low-cost-slimm…

I wonder if this is the same or different than the app mentioned in the first link?

Nilvest,

As you can see, the T2 market is wide open:
https://www.dexcom.com/sites/dexcom.com/files/professionals/…

Correction: Previously I said some people had to submit food logs to insurance. Typed the wrong thing…should have said glucose logs.

1 Like

Shoot, meant to include this paragraph in cut/paste. So…Dex will be competition to LVGO even though it’s partnering with them:

Insulin dosing advice for users on injections. Dexcom recently acquired TypeZero, a software startup that has insulin dosing advice algorithms for injection users. During the presentation, a slide of a CGM-based mealtime (bolus) calculator, forward-looking hypoglycemia prediction, and even sleep and exercise advice were shown – e.g., “For 60 minutes of medium intensity exercise, we recommend eating 38 grams of carbs” or “Before you sleep, we recommend eating 9 grams of carbs.” These features could make dosing insulin far easier for those on injections, removing guesswork and math. TypeZero’s Control-IQ algorithm for automated insulin pump delivery is currently in its pivotal study using the Tandem t:slim X2 device; a launch is currently expected in summer 2019. https://diatribe.org/whats-coming-dexcom-2020-low-cost-slimm…

So, right now, with the G6 app you can (not all people do) log insulin use, carbs eaten, exercise, and how you feel. They already capture a lot of info besides glucose readings. I don’t know what the G7 software will look like but I imagine it will give the same or more choices.

(I think I’m talking myself into buying more DXCM. lol Haven’t really looked at it for a year. Yeah, one of those people…but trying to change my ways.)

2 Likes

As you can see, the T2 market is wide open:
https://www.dexcom.com/sites/dexcom.com/files/professionals/…

I was checking around to see what this “seven plus” is. My apologies, I thought I had some reference to the new G7 coming out. Apparently they called their cgm from 2005 the Seven Plus. So it’s an old pdf. I couldn’t find a comparable doc with today’s coverage to see how many companies cover T2s.

1 Like

(I think I’m talking myself into buying more DXCM. lol Haven’t really looked at it for a year. Yeah, one of those people…but trying to change my ways.)

And you are talking me into looking at DXCM as well…
honestly, I remember looking at DXCM when I was reviewing TNDM almost a year back… and did not dig into it for some reason… I look back and cant remember why but it certainly looks like a big miss as DXCM has done very well lately…

Fafar, thank you for sharing great insight into many things around these stocks.

1 Like