Welcome to the big unlock where we discuss data analytics and emerging technologies in health care. Here’s some of the most innovative thinkers in health care information technology. Talk about the digital transformation of health care and how they are driving change in their organizations.

Paddy: Hello everyone and welcome back to my podcast – The big unlock. This is Paddy and it is my great privilege and honor to have as my special guest today Colt Courtright, Director of Corporate Data and Analytics at Premera Blue Cross. Colt welcome to the show.

Colt: Paddy, thank you for having me.

Paddy: You’re most welcome. Now for those of us among our listeners who may not know who Premera Blue Cross is. Would you like to share a little bit about the company?

Colt: Sure. So, we are a Blue Cross plan. We’re headquartered in Washington state just out of Seattle. We cover a little over two million lives and our primary markets are Washington and Alaska. For folks who may not be as familiar with the insurance market we predominantly insure commercial, individuals so these are working age people not retired Medicare or Medicaid populations. We are an insurer of choice for some larger name companies that people may know like Starbucks, Amazon, Expedia, Warehouse or Microsoft. So, we do tend to insure a lot of well-known companies that have a national presence. In addition to local employers in Washington state and Alaska.

Paddy: That’s great that’s I think a wonderful introduction to Premera. And of course, I love the Pacific Northwest so I’m looking forward to my next visit to Premera sometime soon.

Colt: Actually, give it a few more weeks. I’m literally looking at snow so it’s unusual for us at this time of year, but we actually have snow coming down.

Paddy: Alright! So, Colt you are heading up the analytics function of Premera. So, tell us about how the analytics function is set up at Premera. What are some of your primary focus areas? Where does a function sit within the organization who reports it? Can you just talk to us a little bit about that?

Colt: Sure. So, I lead the corporate data and analytics that’s really an umbrella organization that provides services throughout. Insurers tend to be heavily data driven and so under this umbrella includes the large production data assets like data warehouse, data science platform, data lake environments. The large containers of information that analysts and actuaries and underwriters use as part of normal insurance business. I have a consolidated responsibility for sales analytics and reporting, marketing and analytics and reporting, clinical and operational analytics reporting. So essentially everything except for underwriters and actuaries although they are very large customers of the data assets and analytic capabilities that I and my team oversee.

Paddy: Right, that’s interesting. I’ll come back to the comments you made about marketing analytics versus all the other types of analytics related to your clinical data and member data. So, let me start with this. There’s a big focus on population health management today and there’s obviously a need to understand patients or members holistically by using data from multiple sources. Right. So, in addition to your own claims data what kind of data are you now using. And how are you using them. Can you give us a broad overview of that?

Colt: Sure. You know the way that I might answer your question in the beginning is we think about really addressing three primary issues with regard to population health. And one has to do with the fragmentation of care. So typically, the more ill someone is, the more doctors they see and the need to coordinate care across a large number of provider or ancillary organizations like labs and pharmacies becomes very important. So, fragmentation of care quickly leads into a fragmentation of data issue. And we spent considerable time trying to get our arms around that and maybe I will talk about that in a minute. And then the third area around population health is really the fact that incentives still are not aligned across the system. So, you know payer, provider and employers oftentimes have competing incentives and so Premera look for ways to address one two or three of those issues. And I’m probably most equipped to deal with a data piece and I’ll say many years ago we moved down the path of building a longitudinal record of members across provider groups and employer contracts. So, we could really get a more holistic picture. And then we moved down the path of working to share that data with our own clinicians, case managers, mental health professionals, pharmacists, medical directors to be able to better engage with members in supporting their own personal health within the population health. And so those are longstanding efforts that we pursued and those have moved into new capabilities around interoperability where traditional claims data has been really added to significantly by bringing an electronic medical record information and we’ve been able to do that across a number of states. I can talk more about that in a minute. But having access to lab results, pharmacy information, detailed clinical, CDA, ADT type information is really a significant addition to providing a better longitudinal member who record for Premera and then ultimately are provider partners.

Paddy: Right. So, let’s dig into that a little bit and maybe maybe we do a rapid fire of a handful of different data types that are not emerging as important sources of insights about patients in your pursuit of this meaningful longitudinal patient record and a 360-degree view of the patient as well. Or the member as you might. So, let’s start with the rapid fire. Let’s start with EMR data. How are you going about integrating clinical data specifically EMR data with your claims data?

Colt: So today we have two primary efforts underway to capture that data. We connect to an HIE in Alaska. They invested in that state with the capability that in real time consolidates pushes and pulls medical data across provider systems in Alaska. And we likewise receive that data in real time in Premera. So, you can imagine that if you are. A member that goes to the ER and you hand over a Premera card and your ID. The system identifies you as a Premera member and we know that you’re in the ER in under a minute. So that’s one example of how we’re integrating to EMR data. Another example is we have a commercial partner that operates across a number of states that focuses specifically on ADT data feeds. And so, it’s not the entire record in a similar fashion to what we’re able to get through HIE. But we can accomplish the same thing I just described now across eleven states and over 400 EMR is because of those two partnerships.

Paddy: That’s interesting. So is that therefore the optimal or the easiest way to get access to EMR data. In other words, work with somebody who has already put in the work of aggregating EMR data from multiple providers and multiple EMR systems. Is that the best way to obtain that data in this context?

Colt: So it is certainly an accelerator right. So that’s why healthcare information exchanges I think are getting more traction than older ideas that people were pursuing that were a little similar in the 90s and early 2000s. Instead the Health Information Exchanges are making traction. And they are a huge accelerator to be able to get access to this information for the benefit of the member and other providers.

Paddy: So, let me talk about the patient matching you’ve made a reference to patient matching especially in the context of clinical care. You had also mentioned that part of your responsibility was marketing analytics. So, for your marketing function you’re targeting your health care consumers, your existing members or potential members as the case may be. What kind of data do you use for that function and how do you go about the patient matching in that context?

Colt: So, we have I would say we have the typical payer approach to market in an environment that’s heavily employer driven. Right. So, we do have you know individual lines of business. But the bulk of our work is really in the employer space and oftentimes in the large employer’s space. And so, we do engage in and I would say a more traditional type analytics and in the marketing space it’s really once we have clients that we provide differentiated in data science like capabilities that are continuing to expand year over year in insurance market.

Paddy: Right. Do you ever use a commercially available consumer data such as for instance from the credit bureaus? Do you find that to be a valuable source of information on your members or do you not feel the need for it?

Colt: So, we do have some data we have not found it to be valuable and I would say that even more strongly when we do our own machine learning work to try to understand who might engage with us around their personal health journey. Right. So, if someone is newly diagnosed with a condition of have a program to better support their needs and help make them an informed person around the condition, they have an informed health care consumer. We have actually on the data science side attempted to work with that information to see if that had any predictive power in understanding who might be more willing to engage with our clinical programs. And again, we have not found that to be, it does not provide a large lift and you know one can imagine why but we don’t know for certain. Health care is very different and people’s you know idea about how they make decisions around their health seems to be quite a lot different than if they were to buy a book on Amazon or a type of coffee at Starbucks.

Paddy: Right. Right. And of course, your member base is a lot different from someone who let’s say it deals with Medicare or Medicaid populations. The reason I mentioned that is that this whole notion of social determinants of health that there seems to be a lot of interest especially among providers who are addressing a large Medicaid population because they want to know where they live what their needs are. But as food or transportation or any of that and that’s you know that was kind of where it was going with that question, but your customer base doesn’t include any of those segments. And so, I imagine that has something to do with why the consumer data doesn’t meaningfully add value to you.

Colt: Yeah. Where we eat where we tend to see some of the social determinants impact. And to your point yes, you’re right there is because these are you know employed we tend to have a younger population. And just given her our employer base and we’re fairly active in Northwest our demographics do look a little different than some other plans. And what I would say is that the closest connection we found is having awareness of support in the home other family members or other things that can be a meaningful predictor around for instance readmission to a hospital or return to an AR are repeatedly. That’s where we tend to see value in our population.

Paddy: All right. So, let’s move on in the rapid fire. Let’s go to the next data type IOT data wearables.

Colt: We experimented early on. We don’t capture a lot of that anymore. I would say we didn’t see at the time a lot of willingness of the provider community to entertain that. And similarly, we did not find a lot of I will say financial value for the plan in that space. To the extent that those things are becoming more motivational and they create healthier improvements for the individual I think they’re fantastic. And certainly, we have plans to kind of work to help people engage with those type of tools to help them on their healthcare journey. I won’t throw out names, but we actually work with a vendor organization around healthy lifestyle. And another one around pain management and IOT devices are part of that and helpful from a pure financial perspective. We didn’t see the value for our core work and the provider community. When we were more heavily involved in that area it was not receptive to use it in that information. Now that will likely evolve over time and we may end up coming back to that at some point in the future.

Paddy: Right. And that’s actually as I said earlier to my last data type that I wanted to ask about genomics data.

Colt: We deliberately we’ve deliberately stayed away from that partly because there were some fairly restrictive regulations in our Alaska market around accessing and using that information. So, at this point we do not leverage that as a core part of our work.

Paddy: Do you see do you see a future, or do you see potential for genomics data to be an important part of understanding patients and intervening. Do see that to be something in the near future. Or do you think that’s got a longer timeframe to play out.

Colt: So, in those two areas that I kind of have a lot of passion around this one is interoperability the other data science. When I put all of my data science, I can certainly say that there is tremendous promise in that space. Right. Everything from trying to better understand clinical trial data and even ultimately bypass the lengthy delays and RCTs are things that the pharma arena is examining the opportunity for tailored medications and treatments or even-tempered benefit designs. Really get kind of exciting to talk about. But I think we’re still very early in trying to even know where and when we should be doing that work. And then of course you know there’s gonna be a public concern that can be used to disadvantage certain individuals. So, I think you know there also needs to be a closer examination to make sure that when this is used I think it is a win not an effort. When it is that it is used appropriately across the health care system from the research community through the insurance industry.

Paddy: Right. Right now, that is well said. Well you mentioned two things that are part of your core role. One is that data sciences hat that you wear and the other is interoperability. So let’s talk about data sciences. So how do you go about really leveraging all the advanced analytics tools and platforms and capabilities do you build it all in house? Do you the partner with someone you know what kind of investments have you made in let’s say artificial intelligence which is all the rage. Can you talk a little bit about the advanced analytics capability that you’ve built for yourself?

Colt: Yes. So, I’d say that we we’ve been doing both right. So, a lot of my time and attention has been placed around how do we take traditional analytic activity in a payer system and move that into the world of data science. How do you move from structured columns and rows to unstructured data and then you know frankly because it is new and it’s advancing, and retraining retooling takes time? We also have partnerships and maybe if I were to talk for a second about what we do in each of those two areas it would be helpful. So internally for instance we apply data science to recorded and transcribed and tagged phone calls with our members. So, when members call in, we have a record right. And if you think about why a member calls is generally going to center around something that they didn’t know before and in many other industry the phone call is as a deep act in the prospect. Right. So, they didn’t understand something we may not have described it in the right way. We may not have made the information available in an easy way to access. And so, we take that call data and identify themes know reasons for members to call us. We’ve had to build our own natural language processing capability because the language of the insurance market is quite a bit different. There’s relatively advanced NLP in the retail space of the banking space and even in the clinical care space but our language in the insurance industry is quite a bit different. So, we had to invest in an NLP, construction and training and then obviously with the EMR data that provides a whole new area of focus around nursing notes, doctor notes, misdiagnoses that can be tied back to commercial risk, can be tied back to quality of care, care gaps. Those are areas where we’re focused internally and externally, we do have partnerships we have partnerships with some venture backed startup companies. We have an affiliation with Stanford University for instance and we’re trying to leverage the best and brightest thinking in that space and continuing to move forward. We have predictive capabilities that sometimes require either unique talents or additional data partnering with another organization can improvise.

Paddy: Right. Right now, you know you mentioned interoperability which is the other piece. And from what I understood from your earlier comments by virtue of the fact that you have this relationship with a fairly robust HIE. You are able to get all of the information aggregated and presented to you possibly through one interface one API and that can drastically simplify things from a technical point. But that still leaves the question of semantic interoperability which is what I think you alluded to when you talk about the language of the insurance industry and how that you know overlays on the language of the providers use or any of the other participants in the new data ecosystem may use. Can you talk a little bit about you know how you address that?

Colt: Today we are we’re at an earlier stage and I don’t think that’s unique to Premera. I think as an industry we’re still trying to wrap our arms around pure data even raw data interoperability. We’re leveraging the standards that are out there to the maximum extent and then from the semantic standpoint we’re really still using human beings right. If we can bring the data in and we can interpret it sufficiently not standardize but if we can interpret it sufficiently and push it to the right human being to do something with that information that today is the win. So, we don’t have true semantic interoperability and it’s challenging. As templates and values are designed uniquely sometimes, they’re not even equivalent within the same large provider system. So that is the challenge. So, it’s a win for us to simply be able to ingest data and where we have moved to more recently is to not only be able to receive the data but to be able to push the data. So, where we are able to provide more context is in our own data and they’d be able to put our data into the message typed and make it visible to the provider as part of their workflow and to do that in a seamless way as we can. And and so that I think is just kind of the current state of where we are today.

Paddy: Right. Right. It’s still a work in process and it’s still early days yet. As far as that is concerned. So, let me switch gears and let’s talk a little bit about digital transformation. You know the entire industry in fact most industries are in the throes of a digital transformation. And some may argue that you know other industries are probably further ahead in that journey than health care is generally put. So, can you talk a little bit about what you know what is Premera’s digital program look like and what kind of focus areas are important at this time. Premera as it move towards a digital future and engaging with your members or with your provider network or any other part of your organization.

Colt: Yeah. So, I would say certainly arming the providers and the frontline capability to engage our members is first and foremost. Right. So being able to deal with the fragmentation of care, fragmentation of data, and then trying to align and system incentives actually are a huge part of our digital transformation. It doesn’t sound like what people think of when they hear digital transformation but a lot of it’s around the receptivity and the willingness and to receive the data and act on it. The other party that is a very important force is the member patient right. So, the two people who need to have this information to make a decision are the doctor and the patient. And so, where we’ve made progress on the member side is that we are using interoperability so when we know that you make an appointment with a doctor, we have that capability to do mobile scheduling through a partnership we have its one of the other venture funded organizations I was mentioning earlier. That we will provide you when you have your appointment with an updated record for instance of your pharmacy. So, you don’t have to remember that when you meet with a doctor and the doctor can have insight into. Are you filling the medications as they prescribed so if thats the data we have, and the provider doesn’t they know if they prescribed the medication? We know if the member fills it and oftentimes the member doesn’t remember to tell the doctor that they actually have this other medication that they are taking and that can have a profound impact if the doctor makes a care decision in a new appointment. So, we’re empowering the member with data where and when they need it and then we are also working through transparency. And that’s been a challenging topic in healthcare for a couple of decades. But we do through this other partnership have mobile capability. So, if you as a member choose to accept participation and use of the app we will help you find a high value physician whether it’s a primary care doctor or or a specialist.

Paddy: Right that’s very interesting. It’s fascinating actually. So, I will touch on something else more recent development. You know I was at HIMSS a couple of weeks back. Oh gosh. That was last week. That sounds like a hundred years ago. One of the big things that dominated the conference was the new proposed ruling by the Health and Human Services and the CMS which was around data interoperability. But it was also about a number of other things. Essentially the focus was on providing a data transparency to patients and putting the power back in the hands of the patients in terms of having access to their medical records without having to either pay for it or go through a lot of effort just to obtain it. And so, I’m told it’s an 800 page ruling. So, do you have any initial thoughts on what you know what that means for health plans such as Premera.

Colt: So, you know I need to stay away from specific responses to the ruling because we’re going to have that comes through the association as you know that we are a participant in. I will though comment in maybe a more general way and this may still sound a little heretical or new, but I think it’s going to be an increasingly common observation as this type of ruling moves forward. And that is you know healthcare I think has been let down by IT investments over the last 20 years. The concepts of interoperability to my knowledge begin with this realization in the late 90s that led to various human report and this realization that not having information at the time needed was hugely impactful to patient safety and an impact on medical errors. More recently we know it also influences care gaps produces waste and other things and you know the unfortunate thing is that our industry payers and providers both have spent a lot of money. We spend billions of dollars every year on technology and we’ve done that for a long time in good faith. But when I walk down the halls whether it’s with the provider partner or in Premera we still see fax machines and people scanning documents. And So, I think you know again we’ve been let down by a lot of IT investments. And so, what I like about this ruling is there’s new pressure right. I would say it’s a culmination of nearly two decades of observations around the impact of not having interoperability. And it is topical now and I think it is one of the most topical things that we talk about in the health care system over the next five years. So, I think this puts pressure on future investments to actually accomplish the mission. Stop pushing paper around. Stop having people call each other two to share data. Let’s do this in a modern way. We don’t carry our bank record when we visit a bank across town to get our money. We’re expecting that their system their ATM is going to be able to know you know about us and the ability to access that information and the dollar amounts are going to be correct. So, from a computer standpoint it’s a data problem and data liquidity and data exchange problem. So, I think it’s good. The two most important things that I’ve seen are the one you mentioned member access to their own record. I think that opens up a ton of new opportunities. See you mentioned for instance genomics based on the members interest there could well be other industries that crop up that help advise people and support them based on their care needs and care utilization. And the second really important thing is no data blocking right. So, we need to make sure that systems are communicating and that there’s not any one party because of how they might be incentivized today to stop data from being exchanged appropriately and freely to support health care delivery.

Paddy: Right. Data liquidity I think that is probably the big takeaway from all of this that if anything this is an improved data liquidity and that is all for the good. So now thank you for those comments. So, we’re at the end of our time and I really want to thank you for taking the time out to speak with me and for sharing your insights. It’s been a fascinating conversation and I look forward to speaking again very soon.

Colt: Thank you. I’ve enjoyed it Paddy and I wish you and all your listeners well.

We hope you enjoyed this podcast subscribe to our podcast series at www.thebigunlock.com and write to us at info@thebigunlock.com