https://www.linkedin.com/in/cklemke

This great healthcare interview is with Christine Lemke. She’s the president and co-Founder of Evidation Health. Evidation works with healthcare networks to improve health outcomes while saving money for everyone. She leads strategy, research, and special projects at Evidation.

Prior to co-founding Evidation Health, she was the co-founder and Chief Operating Officer of Sense Networks, developers of the first machine learning platform for mobile phone activity data (acquired by YP.com). She has also held roles at 3iGroup (Paris), Microsoft XBOX and co-founded Chicago-based Channel IQ, a product analytics company. Christine has a BA from the University of Washington and an MBA from HEC Paris.

Other questions Christine answers:

-Can you give a case study how you helped a healthcare network?
-How are you using personalized medicine?
-What are your biggest challenges ahead integration personalized medicine and healthcare analytics to improve care and health outcomes?

Transcript
Dave Kruse:  Hey everyone.  Welcome to another episode of Flyover Lab.  This is Dave Kruse from Madison, Wisconsin and today we have Christine Lemke with us.  And Christine has a wonderful background which I’ll probably mess-up, and the reason why she is on the Flyover Labs is her current role as a President and Co-Founder of Evidation, and we are going to learn more about it.  But Evidation works with the healthcare networks to improve health outcomes, while making the hospitals and everyone more efficient.  At least I think that’s it, so we are going to learn more.
And so there she leads strategy and research and special projects and prior to Co-Founding Evidation, she was the Co-Founder and CEO of Sense Networks, which was developers of a machine learning platform for mobile phone activity data, which was acquired by YP.com; that sounds interesting too, so we should talk a little bit about that.  And then she can tell you about – she’s had a lot of other interesting roles with XBOX and Channel IQ and so like I said, she has quite a rich background.  So Christine, we definitely appreciate you coming on the show today.
Christine Lemke:  Thanks for having me.
Dave Kruse:  So maybe your background is a good place to start, since I probably destroyed some of your background.  Maybe just kind of given an overview, so maybe the roles that you really liked and kind of shaped to who you are and then we’ll get more into kind of what you’re working on now.
Christine Lemke:  Great, it sounds good.  Well, my background is at first glance really, really different.  I would say I think the coming thread in all of the things that you mentioned, my work at Sense Networks, my work at Microsoft in the XBOX division, my work at Channel IQ and so forth has the connector of data analytics in all of them.  So at each one of these functions I was either managing products or teams to do data analytics to improve something about the product or services that we offered or in some cases find the right people who are misappropriate for whatever offer and something had to be give to them.
So Sense Networks is a great example of for long time marketing was using what I’ll call static demographic profile or market research driven profile that tends to be fairly static, to figure out how to match people with the right advertisement.  So if I’m the Hilton Hotels, I might hire a market research firm and say, here are all the character qualities that who is a typical Hilton Hotel’s customer.  And so we’ll run a campaign and we’ll target all these people by their demographic qualities.  It could potentially be where they live.  It could potentially be where their income is.  It could potentially be whether they are male or female.  What their employment like character might be, etcetera, and we are going to send them an ad and hopefully pray someday they might click on that ad.
But as things started to move very digital, and as specially as things started to move mobile, there was a whole new set of data that we could take advantage of as marketers in order to find the right people and the right moment, where if you presented them with an offer from Hilton Hotel they might accept that offer more readily.  And so Sense Networks was really trying to take new types of data off the mobile phones and this could be location data, so its data that updates in real-time.  This could be calling patterns or testing patterns or thing like that and match offers to the right people in the right situation.
So not just the right people from the demographic sense, but the right situation where they might be planning travel if you are at Hilton Hotels or the right characteristics of people based on their travel pattern.  So you know things like frequent business travelers might be able to light up in a location data set much easier than just making static or heuristic assumptions of that, the character, the people or the demographics for those people and so that’s really what Sense Networks did and that type of thing comes to work really well, and we thought where else could we apply this behavioral analytics we call it, in order to improve something in the world and instead of maybe improving ad clicks or improving more phones that people buy or more Hilton Hotel rooms that people sell, we thought well, there’s lots of people in healthcare right now trying to understand behavior change.
And it makes sense to us that in order to change behavior you probably need to understand behavior at scale, in order to change behavior and so couldn’t we apply those analytics and all that thinking and all that work to the healthcare sector and you know build a really great company that not only healthcare helped people better, but make sure that patients get matched with the right care and the right situation.  And so that was really the motivation for starting Evidation.  And all through my career you know, just sort of an early fascination with the power of real time streaming data that’s coming from all sorts of sensors on things and so all of this started cutting in supply chain actually.
Supply chain was the first places where sensors were on boxes everywhere, sensors were on trucks everywhere and so there are lots of interesting things that you could view and analyzing that sensor data and always optimize things in the real world, and that was Channel IQ and Microsoft XBOX is analyzing some of the data on the XBOX in order to – in this case like figure out some of the problem areas in the XBOX and fix them.
Dave Kruse:  Got you.  And so with the – that’s quite a good overview, thank you.  And with Sense Networks and with Evidation, but how did you access that data.  Somehow I curious, were you working directly with the mobile app companies or working with ad networks, because…
Christine Lemke:  You know I think it evolved.  So Sense Networks when we started it, it was in 2006 I think, gosh that’s been a long time.  So initially there wasn’t a lot of sense data.  There wasn’t a lot of location data being emitted from cell phones, because if you remember it was pre-iPhones.  Some people had BlackBerry’s and the location wasn’t persistent across all these devices, like it is today.  Most people have Smartphone’s where there is lots of location data available.
And so the early data set that we started at Sense Networks was actually taxicab data from the city of New York believe it or now and from different cities around the country, San Francisco, Chicago, etcetera.  And so those were first like location behavior signals, so we could figure out sort of how the city behaved based on all these location signals and literally that’s what we cut our teeth on, that’s what we cut some of our system on, that’s how we built them the system, that’s how we created our first set of algorithms etcetera until we got to the point where we could lodge partnerships with big telecommunication companies who of course had tons of location data.
Dave Kruse:  Got you.  So with the taxicabs what were you – I’m just curious, what were you analyzing?
Christine Lemke:  Yes, it was a little cookie.  In the early days of taxicabs location data analysis we were looking through signals on, man this is going to sound really cookie now.  We were trying to figure out whether there was more traffic or less traffic basically in and out of different retail areas in the city and literally like working with hedge funds to figure out well, are we bullish about the economy or bearish about the economy based on some of these signals.
So you get like all the location data, the latitude, longitude timestamp and you would also get whether somebody was in the cab or not in the cab, you have cab utilization.  And so you could sort of like put all these signals together and create a map of the city and potentially project if you had some of the major cities, potentially correlate that to some of the signals in the wire market for consumer, sort of consumer attitude about spending, discretionary spend which drives you know over 60% of the United States economy at least.  And you can start to make models like that.  And then if you were real good at it, you could start to make models on individual stores if the stores are large enough or you could make models based on shopping malls or gas stations or you know other things like that.
Dave Kruse:  And what do you mean by models of a retail store or a gas store.
Christine Lemke:  Yes, you could do some fancy analytics and some early incarnations of machine learning to try and predict what their stock market was going to be based off of some of these moving patterns, and what specific retail stocks were going to do.  Yes, you could do some correlations and figure out what specific retail stocks were even going to do based on this leading indicator that you might have which is, a taxicab addresses.
Now of course, when you do things like this there’s loads of what’s called Spurious Correlations that pop-up and so you have to be – you have to apply a lot of rigger to these things, but it turns out if your – if data is so good for that and you’re so good at it, then you should just start a hedge fund.  You shouldn’t sell your data to anyone, you just start a hedge fund.  But in fact we didn’t see a ton of signal in the taxicab data, truth be told, but we made a system that could analyze location behavior patterns and that was a system that was really interesting and attractive to lots of big health communications companies.
Dave Kruse:  Interesting.  So for a project like that, how do you kind of at the beginning wrap your head around like what’s interesting and what’s now.  How do you try to figure out where is this thing versus the noise.  I mean this applies to all projects, but did you use that project or a different one?
Christine Lemke:  Yes I’ll tell you, in marketing there is a different standard of what have been in healthcare.  In marketing the way that you care, you care about Spurious Correlations, but you don’t care to the extent that you care about it in healthcare.  Because in healthcare there is Spurious Correlations and you follow and you perhaps create guideline around it and you end up hurting somebody right if you are wrong.
In marketing it just means you know a 1,000 more people get the ad that weren’t supposed to get the ad.  Well maybe you wasted a few dollars on targeting people that shouldn’t have been targeted in the first place or you missed a few dollars because you weren’t targeting people who should have been targeted.  And so in marketing it’s definitely a little bit looser, experience has a higher tolerance to Spurious Correlations.
And so there are a couple of ways that people do it, at least back then.  There are a couple of ways they do it back then.  I’m sure it’s evolved since then.  I’m definitely not the expert any longer in marketing.  But back then, we started with an intuition, you might start with a few ideas that you give the data people of when you think situations might arise or people might be more likely to click on an ad or be more likely to purchase something, because we were really up to the purchase, lots about the click and more about the purchase.  And you create rules around that and then you test them and see if they worked or not and then you just go test after test after test in some cases.
And then there is the opposite way of doing it where you can stuff everything into a big pot and you let the machine you know do what’s called the semi-supervised learning and so some of the data you know is – some of the outcomes of what you do are known. So you know that in these specific circumstances through this pile of data some people clicked on these ads and some people did not.  If you have labels for some of them so then you kind of let the model like learn.  And then you apply that model and then you vigilantly watch that model to make sure it doesn’t degrade over time.  And then as soon as it hits the point where it starts to degrade, then you are like okay, a new training set.  So I’m sure it’s far more sophisticated now, but that’s how we did it back then.
Dave Kruse:  Well, that’s good overview okay.  And we could talk about that all day, but let’s switch to Evidation and kind of more healthcare analytics and stuff you are working on now.  Can you share what you do at Evidation and when you started the company, yes, if your employees are – any other stats you could provide that would be – it’s always helpful.
Christine Lemke:  Sure.  So I am the President and Co-Founder of Evidation Health and we are 23 people, who are funded by GE Healthcare and Stanford Healthcare.  I’m sorry not GE Healthcare, GE Ventures, excuse me, and Stanford Healthcare.  We are an elegant mix of people from my world I’ll call it, so its technology, data science world, and traditional healthcare background type folks, which turns out to be very difficult to merge these two cultures and these two backgrounds together and so we are actually really proud of the fact that we’ve been able to do it and we think that merging the two cultures has resulted in more innovative solutions for the market and just a really different way to look at problems with healthcare and attempts to solve them.
And so it’s very simply what Evidation Health does, is we are trying to figure out what specific digital solution and these could be digital therapies, these couple be services around the till, they call them in pharma.  They could be disease management programs that has some sort of technology enablement and they could be devices of all sorts, devices that people use in homes for home monitoring, it could be devices that people are wearing, the ware able type devices.  So generally anything that patients are wearing or consuming outside of the hospital environment we are looking at.
And so we are trying to evaluate, out of all these digital solutions and device solutions what works for him, so what works for which patient, when does it work and how much does it work and how much is really interesting, because it’s not just how much clinical utility does it provide, but we are also interested in how much economic utility it provide.  Because in an increasingly outcomes based or performance based world that clinical and economic utilities together is what drives the decision at lots of major healthcare companies to pay or not pay.
Dave Kruse:  Got you.  And are there specific projects you are working on around where it evolves in digital health and what are some of the…
Christine Lemke:  Yes, absolutely.  So some public thing we can talk about is we’ve partnered with a company called CrowdMed in the summer and did a project, we’ve done.  We did – we initiated a study and we worked with them to figure out through CrowdMed, through the businesses you don’t know yet and Jared the CEO and Founder of CrowdMed is an amazing guy.
But CrowdMed does cloud based diagnostics of very difficult cases, that people who have been undiagnosed for years, they present all of their historical records to CrowdMed.  CrowdMed then identifies all the records.  It keeps the data very safe to be identified and then it distributes it to a network of physicians and clinicians who look at the case and then solve the case. And generally speaking, they solve the case, the give the diagnosis back to the patient.  The patient goes back to their physician and their hospital validates it.  And generally speaking the folks, the study that we did sort of measured okay, after they get the diagnosis like what will happen, what is the cost to the system before they get the diagnosis via CrowdMed, what’s the cost and the impact to the system after they get the diagnosis.
In our study we showed that actually getting that diagnosis done in this innovated way actually reduced load on the system, reduced cost and increased patient satisfaction overall.  So that’s one that’s public, but that’s a great example of the way that we work with some of these digital solutions in order to understand whether they are making any impact and how big that impact is on the system.
In other ways we partner with pharma companies who are building services around their drugs in way.  So the market has had great use – invented some therapies but we needed to go one step further which is help us with adherence.  We need you to start going at risk, at performance risk for some of these therapies that you have, especially as they become more and more expenses when you talk about specialized therapy.  And so a lot of these companies are trying to figure out what are the right digital services to put around their therapies, make them most effective in the real world.  And so we help them by providing not only structure to do that, but some of the backend technology to evaluate these things and also a ready population and their data access in order to evaluate whether the thing is working or not working.
Dave Kruse:  Interesting.  And can you share much about the pharma engagements? Like how it works from the – they come to you with certain questions and then you kind of lay out a plan.  Like you know what – how does it work with – are they analyzing variables, or they are analyzing pill?
Christine Lemke:  They are analyzing a bunch of stuff.  I’ll try to be super concrete with that and divulging any special information, which is always like a tight rope.
Dave Kruse:  You can always add that.
Christine Lemke:  Yes.  Let’s say there is a company who has created a therapy for rheumatoid arthritis or perhaps it’s been in market for a long time already.  Rheumatoid arthritis is a chronic condition.  Usually the medication for this progresses as the diseases progresses.  That might start with your physician saying, oh, you do lots of exercise, right, it might be progressive to you.  Now you need the pills, it might then progress into you.  Actually you need to switch to an injectable.  So it’s a specially molecule in order to help you with modality, help you control some of the pain, etcetera, and help prevent further degradation of your body essentially.
And so one problem in rheumatoid arthritis that people come to us with is it’s very difficult to measure the progression of that disease, right.  So besides sort of those patients have reported information when they go to see the physician every six months or a year, whatever it is, there is very little data that anybody has or very few tests that people can really run to understand like how a patient is going through those different phases that I mentioned.
But one obvious way that you could measure that if the patient was willing is you know as the patient can wear a device, right, to measure their steps everyday or accelerator[ph] in their day that you understand their mobility and how its progressing along the shape of the disease and so as people move through this disease state, sometimes they become less mobile.  And so understanding all of that might take giving and outfitting the patient with a special device and the analyzing all that data and creating an algorithm that associates to different patterns of the disease progression.
So that’s the first step that we tend to help companies with is – sometimes we already have the data.  Sometimes we already have lots of patients with that particular condition with lots of behavior and activity data and we can just help them with the analytics to do the correlation.  Once they are through that stuff.
Dave Kruse:  How do you already have that data? Is it from past projects, or do you get data from..
Christine Lemke:  Yes, so we are partnered across the ecosystem with payers and providers who have some of this data, and we know that they have some of this data because often they are the ones powering all of this data into their system, so that’s one way that we know.  So we’ve been able to in partnership with some of our co-partners, combined with variable data with health data specifically and to de-identified in privacy safe manner to figure out, are there some data that are available already to figure out some of these associations.  The second way that we do it is we run an independent panel of consumers who have tracked things for years and who from time to time will opt into giving us access to their data in order to do some of these things, in a very de-identified privacy safe way.
Dave Kruse:  Interesting and so what from the payers perspective, what type of data would they give you, let’s say around this rheumatoid arthritis example that would be helpful.
Christine Lemke:  They might – in some cases there are lots of companies in general, larger payers, the companies in general who allow people to participate in wellness programs where they are giving lots of activity tracker data, lots of diet tracking data, fleet data, etcetera in order for the company to provide like a better wellness program for them.  In some cases they have some of that data in a consumer opt-in way and we can get access to some of that data.
Dave Kruse:  Interesting.  Okay.
Christine Lemke:  Everything in this system is designed to be end patient or end consumer opt-in.  Like we are not the company that just sort of does things behind the scene.  We literally go out and ask everybody.  We run on IOB controlled process for all of these things that we believe very, very strongly that consumers should own and control a lot of their own data.
Dave Kruse:  Got you, okay, that makes sense.  And I think I cut you off, you were going through the typical engagement, I mean about – you go through on the step two.
Christine Lemke:  Right.  The second step that this company that might be targeting rheumatoid arthritis is, okay, we understand certain of the markers of disease progression and we understand sort of the characteristic of some of this, the patients evolve and we have either identified them after the market where we create our own app and we want to test and learn it.  And so literally we drop that application, we handle sort of the backend of it.
So we make sure that we are tracking all the right data on the app, we make sure that the algorithms will start characterizing people as they use the app.  So we do all the backend work and then we are able to just drop the app in a population and watch that population use and consume that app.  And maybe in using and consumer that app this company learns a lot about how patients engage with it, they learn what works, what doesn’t work and we help them optimize that app through the highest levels of thinking that’s possible.
And then we start to get a sense for whether it’s going to drive an outcome or not, whether it’s actually going to increase medication hearing, either primary or secondary mediation hearing.  We actually learn whether people like the app or not, we actually learn whether there are other types of outcomes that the app can improve, if the service can improve.  And once we have a good sense of that, like it has been indicated as it promised then we go to what’s called evidence generation which is literally designing and running a prospective, in most cases randomized control study, where we are literally separating out the exact impact of that service on clinical and economic utilities, so on cost and on health outcome.
Dave Kruse:  So what will be the example of that, that kind of third stuff?
Christine Lemke:  Yes, the third stuff is, so CrowdMed is one example of that, where literally we designed and ran a perfective RCT.
Dave Kruse:  Got you.
Christine Lemke:  So we are working with a few other digital health companies to that are unannounced.  We don’t love to talk about our clients.  If clients want to take about partnering with us, that’s great, otherwise we really want to look best at privacy and confidentiality.  But some examples are diabetes management programs, some examples are medication reminders things like this.  Examples are they are over 600 DPPs in the United States, these Diabetes Prevention Programs.  So there is a whole class of different types of programs, let’s call them that need to generate evidence that they actually impact people’s health.
Dave Kruse:  So you guys really provide kind of from beginning to end from initial kind of questions all the way to the clinical studies, you are involved.
Christine Lemke:  I’d like to call it a pipeline and the comments that they need in each step of this pipeline is access to people and access to data.  And that’s been really the core competency of the company, is to provide access to people and data who are willing to participate in these things, willing participants in their health, willing participants in research and really managing and making sure that yes, there is informed consent this way, that there is compensation when appropriate, that there is data showing back with the patient, etcetera, that’s really our core competency and then people can build solutions on top of that.
Dave Kruse:  Interesting.  So you guys get involved in a variety which is cool.  So do you, is it real more consulting or do you guys kind of have like an analytics platform that you use or what have you built over the years.
Christine Lemke:  Yes so over the years we’ve built a lot of things to facilitate all the different solutions of top of you know the people and data access platform.  And so that includes things like how they send product.  So literally if you want to design and do test and learn as I call it or evidence generation, there is a product that we can spin up a study, get people involved in that study, get access to the different data elements that you need, do what’s called protocol compliance during this phase which is essentially like very smart messaging and then we have all the connectors, all the different data elements that you’d want to include in that study, and so all of that is pretty turnkey.
I think the slowest part of launching the study is actually the work to design the study and make sure that we are designing it properly, and in many cases we’ll partner with a principal investigator.  In some cases, some companies want to use our own health action team, that’s great too.  We are super, super flexible on that, so…
Dave Kruse:  Got you, interesting.  Well, that’s quite a useful tool, because in most of the stages you are involved with, they have some type of digital health component to it.  You are not a CRO, just doing a typical trial, right.  You’ll have some type of variable or would it have a genetic component?
Christine Lemke:  Yes, that’s right, that’s right.  So our focus today is really on things with digital components to them.  It strikes us as our – it strikes us as something a little bit unique and different which you run these projects in a way that provides for – it requests like a different level of analytic and a different level of information from the patient population.  But I think traditional theories should start – are well expected to arrive.
So an example is, it’s one thing if you go, if a CRO might partner with a data aggregator, perhaps potentially just aggregate a bunch a data, then it’s like what do you do with that data.  Well we’ve spent.  I personally have spent over a decade analyzing behavior data.  The data science team at our company has already built a bunch of behavior algorithms that people can use.  The team is already well – has a product that cleans all the data as it brings it in, normalizes it properly for behavior analytics, merges the behavior data with the health data, which is not you know trivial and so there is a lot of pre work done in order to get the outcome that everybody needs.
Dave Kruse:  Wow.  All right, so your platform could take in the data from let’s say the pharma or the healthcare network and then merge it with – and that’s interesting…
Christine Lemke:  Which is ISIS, with location data with all sorts of behavior data or behavior signals from the real world.
Dave Kruse:  Got you, interesting, okay.  And so I got a question.  Last might I was in the urgent care with my daughter and I was sitting there.
Christine Lemke:  Oh! I’m sorry.
Dave Kruse:  It was okay.  And so I was looking around the room, of course a lot of the equipment there looks like it was back 30 years ago when I would have been in the urgent care and it’s pretty much the same.  So you’re doing lots of interesting things and probably a lot of impactful things and actually we should talk about some of the value you have seen that you’ve created or discovered.  But before that, what’s kind of limiting a lot of these new technologies or new ideas or new models from actually like getting out in the real world and what would you do to try to get this new technology out there faster, and technology could just be the analytics your running.
Christine Lemke:  Yes and so we come from – I come from mostly the technology world and it’s interesting in technology.  So there is this mean going on, I’m going to rush just a bit, so my answer might be a little bit scattered.  So there is this mean going on in healthcare that consumers don’t pay and that’s really like holding the industry back in some ways that consumers don’t pay.
And so when consumers don’t pay, which I’ll argue that consumers actually do pay.  We pay our health insurance premium like every single month, like a drum beat.  But maybe the finer grain example, the finer grain way to say that is consumers don’t always decide what programs to use.  Does this mean that consumers don’t pay, and therefore since consumers don’t pay, we have to bend [ph] some of these big healthcare companies and providers to pay for your program, right or to share cost with consumers for your program.  And so the half paid effect of that of course is at least through digital products that are newer in market, even for some medical devices and things that are new in market.
You have to generate evidence, so you have to like walk into a payer, you have to walk into a health insurance employer, you have to walk into some of these decision makers to tend to decide what gets reimburse, what gets partially paid for, what gets partially reimburse, etcetera.  You have to show them that you’re ridged, actually impacts the system in a positive way; reduces costs, improves outcomes.  Actually the bar is even much higher.  So it can’t just improve outcomes or reduce cost, but it also has to do that within a 12 month ROI period, which is kind of insane.
Dave Kruse:  Yes.  Yes.
Christine Lemke:  Especially when it comes to health.  And so that tends to be the mantra, so.  I would say that’s not enough. So it’s great that you have proven that perhaps you have proven some clinical and economical utility and of course we are always happy to help people do that.  But then the other thing that you have to worry about is, how do I get consumers or patients engaged with my programs and if you have the water there, but you can’t bring the horse to the water, then you are just kind of screwed anyway, right.  So the thing that people aren’t paying enough attention to is that consumer engagement team and if you don’t have that, you are just dead in the water already.
Dave Kruse:  Got your, interesting.  Okay.  And I think we are kind of running out of time here, but I am curious if you have an interesting case study that I don’t want your projects where you are – felt like a really interesting result based on your analysis or something to add value or something the insurance company or health network could do to improve health outcomes that you can share.
Christine Lemke:  So we – one example that I think is public as we know.  So we are behavior analytics people.  We look at this type of data in a very different way I think than other companies see it.  So one thing that we are really proud of is we looked at people, patterns of behavior over time, like over a very long period of time and correlated those patterns of behaviors either with people tracking, like their sleep and their diet and their steps in some cases.  In other cases in tracking their weight, their blood pressure, glucose readings, etcetera, they are just tracking things on a regular basis and so we were able to correlate or find a correlation between how habitual people are and not surprisingly their medications hear in.
And so now you have a tool or you have an algorithm as we call it that’s been validated across 300,000 people, across disease areas, [inaudible], hypertension, diabetes as a proxy for medication hearing.  And also we are proving that other things, this habit measure we’ll call it or this routine score that we call it measures and now if you are able to just take how that turns, the behavior patterns of those and identify who is going to need help when and how expensive help they are going to need versus others buyers.
So there is definitely some folks in the system who need a little bit more help adjusting with the routing, need a little bit more support perhaps in live coaching, a little bit more than others, especially when they get discharged from a hospital with specific conditions.  And then there are other people who are what I’ll call set it and forget it type.  Where you give them, you clip them with instructions, you give them reminders, but they can follow a routine like pretty easily.  And so now you have weighted sort segment to your population by their behavior patterns, that you can attach the right behavior program to them and reallocate your resources, your expenses resources to the people we need the most versus the people who would be bothered by them really, don’t need it at all.
Dave Kruse:  Interesting, that’s a good one and how do you choose out kind of habitual pattern.  Do they eat breakfast at the same time every morning, or they go to the store same time each week or what are your looking at?
Christine Lemke:  Yes, you know we published a paper on it.  It’s published on Fox1 actually and so you can go into the details of it if you’d like, anyone really can and we are happy to treat it, there’s a deposit, but it turns out that over time it’s not just the time or data people do things or the manner that it’s the quantity at which they do them.  So how much do they, how many steps do they take a day that can boost that? How much do they track their diet and is it consistent and do they track three mean a day, do they track two means day.  It actually doesn’t matter the number of meals they track day as long as it’s consistent.  So consistency is key.
So sleep is kind of the same way.  It doesn’t matter if you’re a four hours sleeper, some people are versus a 10 hour sleeper; some people are 10 hour sleepers, as long as you are really consistent about it.  The moment that you are inconsistent about it, it impacts your BMI, it impacts I’m sure other productivity measures in your life that we haven’t been able to study yet, but for sure it impacts your BMI.
Dave Kruse:  Interesting, okay.  And where do you want to take Evidation? Like do you want to get more into the generic components and maybe already are, we haven’t talked about it.  Maybe you can kind of call it digital health, but yes, what’s your future look like?
Christine Lemke:  I’m personally obsessed with the generic component of all of this stuff.  I think it will be really interesting.  I don’t know at what pace the company attacks that point yet, but really what we are really after at the end of the day is creating a really interesting connected population of people who want to be more involved in their health and being more involved in their health might be participating in a research that – so it can understand what’s working and what’s not working.
Participating in their own health by providing data that helps us match them with the right programs just for them and their unique situations and then encouraging people to enroll in these programs that will really make a material difference in their health outcomes.  Like that’s really what we are about at the end of the day and then all these solutions on top are just innovative ways for other folks to engage in the system of people who want to be really involved in their health.
Dave Kruse:  Yes, well that’s a great mission.  Must be pretty easy to attract decent talent with that vision.  I mean it’s always hard.  But you got a vision.  It must be nice to walk up each day know that your…
Christine Lemke:  It is.  It is, it’s actually really fun to wake up each day and think about this problem.
Dave Kruse:  Yes, I bet, I bet.  Well I think we are out of time here, unfortunately.  This is fascinating.  Maybe we’ll have you back on in another year to talk more, but I definitely appreciate your time Christine and hear all your insights.
Christine Lemke:  Sounds good.  That’s Dave.
Dave Kruse:  And thanks everyone that are listening to another episode of Flyover Lab and we’ll see you next time.  Bye.