Causeway Solutions
ConnectWithCauseway, Travel Data Insights 2023
April 22, 2024
Healthcare marketers can attract new patients by looking beyond ailments or treatments and leveraging behavior and lifestyle data – while maintaining HIPAA compliance. Listen in as we explore AI and predictive modeling to create Advanced Audiences for more effective marketing strategies - that’s also more budget-friendly. Listen to the full episode.
Thérèse Mulvey, Vice President of Strategy: Welcome to Connect with Causeway. I'm your host, Thérèse Mulvey, Vice President of Strategy and Insights at Causeway Solutions. Joining me today is Lauren Kornick, our Manager of Strategic Partnerships. Hey Lauren, welcome to episode eight! We're excited to welcome back our friend and colleague, Tim Duer. Tim's our Vice President of Healthcare and Strategy, and he's going to talk to us today about patients as consumers and what happens when a consumer becomes a patient, which is our very clever title. So how do healthcare audiences relate to patients and users of healthcare services? And we're going to talk a little bit about what data is available, how do we understand it? There's always so much information out there, and what Tim is going to do today is help us sort through it. So with that, welcome back, Tim.
Tim Duer, Vice President of Healthcare & Strategy: Hey, great to be back. I think this is my third or fourth appearance. If we had an actual studio. This is probably where I get a parking spot or a name badge or something now. Right? I'm officially on the team.
Thérèse: If I could get you a parking spot, I would. I always wanted one of those. I guess I'd have to drive somewhere now, but whatever the details of it, we can work out. Once you get the five-timers club, we'll talk about it.
Tim: Okay. SNL style. Gotcha.
Thérèse: Yeah, I think Lauren and I get spaces first if I'm correct anyway. Okay, as many of our listeners know, we've been doing a lot of work in the healthcare sector as we try to uncover solutions for this important and complicated category. Healthcare data is tricky because of privacy issues, so it's not a surprise that the industry is continually looking for different ways to use data and to understand patients. Tim, I know you're looking at this and what would you say the headlines for data and the healthcare sector are right now?
Tim: Thanks, Thérèse. I think, just like everywhere else right now, AI and machine learning is always coming to the forefront. I think everywhere you look right now, there's talk about AI, whether it's ChatGPT or any other tools that are out there right now. But in the healthcare space, that's been evolving over the last five years, even as I was moving out of my previous clinical role into now the data analytics side of things. But really when people are talking about AI and healthcare, I think appropriately, almost all of those efforts are going towards helping with diagnosis and treatment planning and other things involved in actual healthcare. Very little of it right now is really being pushed towards the marketing side or that stuff that happens before a patient walks in the door. Most of what I've been seeing out there has really been using AI to comb through some of the huge data files that are in online patient records.
Try and look for some of those trends that might help a physician with a diagnosis, something that might not look so obvious on the surface, but AI can connect so many of those dots behind the scenes in a medical record and really help point a real life doctor in the right direction rather than starting from scratch. Other things that kind of ties in with the contraindications or precautions when you're talking about pharmaceuticals, that AI can help look at all those crossovers. And the other one I've seen a lot about lately is really the diagnosis of the radiology department. Radiology departments have long been overworked and overburdened and anything they can use to look very closely at some of these records and some of the scans and images that come through, then let the physician pick it up once the discrepancies. So again, I'd say there's lots going on in the healthcare space, but so much of these efforts really are again, appropriately focused, more on the clinical AI side of things, not so much on the things that happen outside of that space.
Thérèse: Well, and that makes sense, right? I mean that's hopefully what our medical professionals know about is medicine as opposed to marketing, which is probably really good for all of us, and I can totally see and appreciate how this helps in terms of providing better care. But I also know that when you switch over to marketing, it gets very tricky when it comes to privacy.
Lauren Kornick, Manager, Strategic Partnerships: And as someone in marketing looking into medical, I found an interesting stat at the American Medical Association found that 75% of their patients were concerned about their privacy on their personal healthcare information. But then on the other hand, only 20% of them actually knew a scope of outside company access. So as someone who is definitely not in that 20%, I'd like to know about the scope. So everyone talks about HIPAA, but what does it really mean, Tim, as our expert?
Tim: No, I would happily give a little bit of a history lesson here because honestly, HIPAA does get thrown around all the time. But very few people really understand what HIPAA is and I even myself get it confused sometimes and have to look into it. So HIPAA, first of all, it's H-I-P-A-A. I tend to spell it wrong myself all the time, but it stands for the Health Insurance Portability and Accountability Act. And this originally came into play in 1996 and we're not going to ask Lauren where she was in 1996…
Thérèse: No, let's not go there.
Lauren: I was two!
Thérèse: Whatever!
Tim: Really that's when I was kind of getting going in my healthcare career. And that was back in the age of everything was a pen and paper medical record. And so one of the big keys with HIPAA was trying to really focus an idea of preparing for portability. So that was the piece that was very important because it was trying to start paving the way for health systems and move towards electronic medical records, which are now commonplace. But honestly, we still haven't really achieved the portability piece where those records don't necessarily talk to each other. But again, back in 1996, almost exclusively, we were talking about pen and paper medical records that had to be carried around, were within a hospital, and that was the big push. But what really HIPAA stands out for now is of the parts they had in the Title II aspect of it, and this really talked about the privacy and security pieces.
Again, at the time it was a little bit more of the afterthought, but now it's really become the far away concept of HIPAA as people think about it. And so these are really the regulations when we talk about patient privacy and how you can and cannot use anybody's medical information. All of these are more than appropriate. Believe me, as somebody who went through a 20 year career in clinical healthcare, it's absolutely essential to make sure that these privacy aspects are in place. But the downside of it is the regulations and fines are so hefty that basically every health system is scared to death of anything that could potentially be perceived as a marketing based HIPAA risk. And it's really narrowed down what health marketers can do because compliance and legal teams appropriately are really keeping a close eye on what they can do. And then the answer becomes, when in doubt don't, and then you end up with very challenging pieces.
Fast forward to now 2023, we've been talking about the challenges that healthcare systems are facing. They're really struggling to get themselves in the health industry, although there's so much money changing hands, the margins have always been small and now they're basically negative. More and more health marketing teams are being asked to really drive volume growth, get people back into the health system or grow facilities, but their budgets are even tighter. Now when they're facing these challenges of we're not allowed to segment using things because we're worried about HIPAA, it's really come to the forefront of how can we smartly and intelligently look at some of these things and not just say we can't do that.
Thérèse: It's interesting, I remember 1996 and all the things you had to sign and this big change and everything. But I never really thought about all of your records and your information in a manila folder in somebody's office was probably a lot safer than having it be on the worldwide webs. So it makes sense why it's so complicated and also why it's so important in terms of marketing. Then what I'm wondering is what data can they use if they're facing this challenge of slim margins and difficult ways to really make the best of their marketing dollars?
Tim: There's a number of options. I'm not going to pretend to be a lawyer, nor do I play one on tv. This is my version through my knowledge base in my history. The key is any data that is used in marketing for healthcare has to be anonymized. And that also means you can't just anonymously send things but pull from somebody's patient record to figure out who to send to instead.
One big option that's in the field right now is what's called claims-based data. That means you're taking huge amounts of patient data. It's all anonymized with a de-identified patient.
The record doesn't know who it attaches to, where they live, anything about that person when it comes to them as an individual from a residential standpoint. But instead, what it does is keep their demographics, the age of the person, the gender, height, weight. All the other aspects that can be used to really help figure out: (1) with a diagnosis, and (2) starting to look at commonalities that are between people.
A hospital might take this internally and roll all this up using claims – claims being insurance claims. There's so much documentation on there. That is one thing I do not miss from my clinical career! All the things that have to go into a billing cycle for a health system, those pieces, once you take away the patient, now we can start looking at consistent pieces of how do you come up with a targeted segment for a specific service line, say like cardiology. Looking at a consistent basis of age, gender, etc., of who's building the cardiology service line. It can even get specific like a procedure, like an angioplasty or something like that where they can say, okay, who has had a procedure within the service line? And as I said, this could be rolled up just within one hospital or health system. What's becoming more common now is different platforms or different services taking many health systems, rolling all of their data up and using that scale to get even more detailed views of segments of who's likely to use a service or a procedure or things like that.
Thérèse: Huh, that's really interesting. I would not want to be on the team that had to figure out how to do that without breaking laws, but luckily none of us are lawyers or law enforcement, which is probably good. Basically what you're saying is much like we do in research and analytics by depersonalizing personal information, we can use it in a way to help understand a large population without breaking any of those rules or putting anyone at risk, correct.
Tim: Absolutely. I mean, and this is a giant undertaking for any health system that chooses to do this and certainly the vendors that are receiving this. There is a huge amount of security compliance oversight to make sure that dataset that's being transferred out is anonymized to make sure that nobody knows who's on the other end of it. And again, that responsibility falls on both parties, but it's also why a lot of these services when it comes to claim based data, when you're talking multiple health systems do involve long-term subscription services. If you're going to go through the efforts involved through the legal team to make sure you're able to export your data, you're not going to do that on a one-off. Typically claims-based data when it comes to marketing usually involves an extended contract a longer period of time because there's so much work on the front end to get these things ready to be used by others.
Thérèse: So I know at Causeway we've used other behavior and lifestyle data to predict the type of healthcare solutions people might be looking for, and I know it's based more on the things I do outside of healthcare, like shopping and things like that. Can you explain how that's different from what you described and from the claims-based data model?
Tim: Yeah, absolutely. Let's back up one step. With the claims based, you're looking at who has taken an action. And in this instance, the action is either going to be seen by a certain type of physician having a certain diagnosis, having a procedure, whatever it is, but there's an action that actually had occurred with that person, even though the person is then anonymized. Really, I'd equate this to what we're all used to in the retail world. A retailer like Amazon or your grocery store with a frequent shopper card. What they're doing is saying, you made this purchase and now let's look at what others that made that purchase did next. If you buy a dog bed, all of a sudden Amazon's going to start pushing the next thing of, okay, do you need a dog toy, a dog tree, a crate?
Because it's looking at the internal algorithms of one person made this action and they did the next three actions, so let's start pushing people down that certain pathway. And I'm sure we can have an incredibly long conversation on that, but that retail analytics has really been a game changer. It's all about predicting somebody's next step based on somebody's prior step. What's completely lacking there is you don't know the “why.” There is no “why” in that conversation. There's no behavioral aspect, there's no ideology. I bought a dog bed, but is it because I got a dog or because my daughter just bought a toy that I wanted to play with. I have no idea. There's all these different why's that claims-based data or retail analytics throws all the “why” out, and it just says the what. Causeway does a lot more predictive analytics. We focus on the behavioral or ideological aspects to the consumer, what motivated them to do this, not just did they do it, but what's motivating to take that action or not take that action.
Lauren: And I think that “why” is the interesting part of our approach. It's something I saw in our research for this topic, and no hate to your previous industry, Tim, but it's this idea that providers seem to fall into this mindset where people almost become just numbers and files. They're just patients or just their ailments or just their treatments. Where this approach purposefully looks beyond that. It looks to the why, it's who the person is rather than what they need treatment for.
Thérèse: I know as a patient, I often feel like my healthcare provider isn't really looking at me, but at the numbers that define me, which I actually do not appreciate. I doubt any of them are listening, but as someone who's worked both inside and outside of the system, I'm curious what your take is on that.
Tim: This could definitely be a very long conversation on that front, but I mean honestly, if we're going back to the mid-nineties when HIPAA was coming into play, that's where I was kind of deciding. I knew I wanted to work in healthcare, but at the time wasn't sure which path I wanted to go down. And for me, I ended up pursuing physical therapy intentionally because I wanted a career where I could actually spend a little more time with the patients and have a little bit more opportunity to figure out the why, not just what I had had enough exposure to no physicians unfortunately had been forced into this of industry of how fast can we turn patients over? And even though I speak often and still love talking about the idea of value-based medicine of where really reimbursement is contingent on how well you do at the moment and in the recent past at least volume is the key when it comes to healthcare reimbursement is about how many patients you see, not unfortunately about the quality of the care.
Thérèse: That is unfortunate.
Tim: It is, and physicians do their absolute best, but if you read and learn more about the physician burnout, that's the challenge is that the volume has to stay high to support everything around them, but that's not why you got there. And honestly, that's some of the goal of the AI introduction is can that reduce some of that burden of documentation or reduce some of the burden of the more tedious tasks so the physician can get back to the care. I mean, if we look at the silver lining of an AI future in healthcare, hopefully that's it, is that it gets back to that. But with all that said, I mean that's truly the key of when they're in the numbers game. That's the challenge. What I think it connects to all of this conversation is the AI aspects can help you find out what needs to be done.
And when I was a physical therapist coming out what needs to be done or the diagnosis even in many instances isn't actually the hard part. It was figuring how do you match the treatment recommendations with the realities that was the patient in front of you. So if I knew a patient that I was working with, it was easy for me to give the exercises they should do at home, but did I know that it was when I was in pediatrics that I know the child loved a certain sport or a certain activity and so I had to make it connect with them so they're more likely to do it? Could I hear their conversation to figure out, okay, the kid needs to do the stretches, it's the only way it's going to get better. But if I just hand 'em a piece of paper and say, go home and do these, I shouldn't be shocked when they don't do it.
I have to figure out what connects with this patient of, is it another interest that has nothing to do with their injury or how do I speak to them in a way that motivates them to follow through? So if I go full circle of my previous clinical life into the data piece, while the diagnostic criteria or the segmentation based on claims or other background demographics is the starting point, if you don't attach the why, you're really kind of working with one hand behind your back as a marketer, it's easy to say, let's identify a demographic that might fall into a category, but if you can connect with them on an individual level, what's their motivator? Now all of a sudden your job got that much easier. You really connected and move them through. And that's what I honestly loved about my clinical career as well.
Thérèse: Interesting. Given that I have actually experienced going to a physical therapist and don't do the exercises versus going to a physical therapist and doing the exercises, it's amazing what the outcome is. They're really different.
Tim: It is amazing when they give instructions, you actually follow them. But again, I'd say if you just hand somebody instructions and they don't follow 'em, I'm saying it's just as much on the clinicians sometimes that you've got to connect or they're not going to follow through.
Thérèse: So true. So true. All right, so we come down to really two different approaches here. I can see the benefit to both, but if I'm a healthcare marketer and I have a marketing team that is trying to decide the best way to serve my clients and ultimately be successful, what do you see the pros and cons of both? I'm sure both have benefits and both have negatives, but how do you break it out to make it a little bit more simple?
Tim: I don't know if it's the pros and cons. I look at it as the idea that whether it's an rolled up platform of claims-based analytics, or even just internal knowledge of a department knowing who their patients have been in the past, this kind of gets your general targeted patient population. Right now, anybody in this call, you don't need a healthcare background to know who's more likely to use orthopedics or cardiology. When we're talking about a slightly older population with some risk factors or maternity care, it's pretty easy to say, don't send 60-year-old men maternity care pamphlets. There's some parts that are just pretty simple and that's the claims-based piece that the new data lets us get more precise about how we focus on that. But it's still just a matter of the who. I think when you add in some of the behavioral aspects, you start talking about why they might engage.
When you put that spin on it, now you're looking at maybe it's custom work saying, why is somebody favorable of one health system versus the other, or some of the non-custom work that we do of how do you identify with that person? For example, in the non-custom work, don't send somebody a brochure about your fast waiting times and the fact that you now have online scheduling, if we've modeled that patient say, actually what they're really primarily concerned about is provider connection. You may deter them from coming to your facility by talking about how fast you'll be in and out the door because you didn't know that that actually is not at all what motivates them in their choice.
Lauren: Yeah, I think it's kind of interesting how each of these approaches can give you a completely different marketing strategy and audience just within the same category. Say cardiologists, you can approach that stereotypical demo of that audience. And then on one side, the claims-based approach, you can look based on historical data, get another layer of it, the potential need for a cardiologist based on that data layer. Or on the other side, you can have a behavioral model that can give you the fuller picture of who and how to approach these people. So maybe there's another behavioral aspect that you didn't even think of. Say that maternity pamphlet to that 60-year-old man. Obviously he won't use it, but he could say, be a healthcare influencer and he could forward it on to a niece, a daughter or whatever. That could actually be a good strategy.
Tim: Yeah, I think you're right, Lauren. Really honestly, if we say it's not one or the other, because there, as you said earlier, there's pros and cons of both. But when you can combine these things and marketers can recognize there's a larger opportunity here, so stay with the cardiology example, I guess even though my background was in pediatrics and sports medicine. So take nothing. I'm saying here as the same as actual medical advice. It's--
Thérèse: Almost everybody has a heart. Yeah
Tim: If you take the first group we said of claims-based or more demographic level, starting with the demographics first, you can narrow down a cardiology population by saying adults age 40 to 70. And if they have a consumer flag of smoking, that really gets the conversation started, that that is audience segmentation claims-based data might let you go further. And again, we mentioned it earlier like an angioplasty, an elective cardiac procedure where claims-based data might tell you exactly for your health system or those like yours. Can you narrow down that age further? Is it less than 40 and 70? Maybe it's 40 to 60, is who tends to use your health system Or similar, maybe there's an income bracket or race or gender or something else that is much more closely correlated with those that are getting that procedure there. Now you've identified a really good target for your service line in your procedure, but how are you going to speak to them?
That's where behavioral predictive modeling comes into play. Do we want to talk about new online scheduling or phone? Do we want to talk about those drivers of choice that I spoke about earlier? Is it bedside manner and connection or is it quality? Is it convenience? Which one of those messages? And then we've got some new models that we've looking at of who's most likely to seek a second opinion or has higher or low trust in the physician. And these are all sort of who that person is. Phase one and two tell you that they're at risk or they're likely to need a certain procedure. But phase three that we talk about, when I have that conversation, that tells me I can either narrow down one audience knowing we're talking about our fast wait times because that's what they care about. Or maybe I'm going to send everybody in that group, but I'm going to send 'em three different areas. I'm going to talk to one group about speed, one group about the quality and another group about how when you're there you have different convenience aspects. I think that's a big game changer. Again, when you do the hybrid version, you say demographics plus claims-based plus behavioral predictive modeling, you really get smart individualized marketing messages.
Thérèse: Given the changes and options that have come about over the past few years. And everything we've talked about here really shows how quickly this is moving, how many different options there are. Do you see that happening, Tim, where people are going to pull in layers of options in order to get the best solution? Or in all honesty, are they really going to be able to afford that?
Tim: I think so. If healthcare marketing teams are going to survive and grow, this is what it's going to take. They can't go back to the well anymore and just do a basic segmentation and send it out to everybody in a region and hope it connects. Consumers expect personalization now, and that doesn't stop because we're talking about healthcare. The key is, like we've talked about at the beginning of this conversation, you don't want to do anything that crosses the barrier of privacy and protection. But recognizing connects with their consumer is what people expect. They want the message. Again, don't tell the person who cares all about bedside manners, about how fast you'll get them in and out the door. That's a deterrent to people that they now come to assume that individualized messaging connects. And to me, the good healthcare systems, that's what they're going to go with.
Lauren: It's kind of funny how it ends up being pseudo cycle. So there are these potential patients out there that really need to be treated as consumers when marketing to them, then they become patients, and then they're treated by or treated with the assistance of machine learning, ai, all of that. But then they're also consumers with our modeling to try to figure out what preferred style of treatment do they need in the first place. And all of that's packaged together.
Tim: Lauren, I agree a hundred percent. I think what you're describing there reminds me of something I speak about a lot with everyone is that the paradox of standard work, and this was something I worked with when I was in the hospital doing some standard practices with lean and continuous improvement. And the paradox of standard work basically says, the more standard you can put in place, the more time you get back to put to things that shouldn't be standardized. And essentially what you're describing is if we can use AI in the healthcare space, whether that's for a marketer or that's for a physician, but if the AI can accelerate the pace of all the tedious, repetitive type of work that needs to be done or identifying the initial audience, then the creativity can really shine and the time can be spent on whether it's the physician connecting with their patients, or maybe it's the marketing team thinking about creative ways to really connect with a potential patient. But that's what I love about all this, that if we can really promote that paradox and say, let's get all the other things accelerated so we can spend more time on the parts that people really want to be doing and the parts that really can matter.
Thérèse: That's really interesting, Tim. I love thinking about it that way. I do see that, I mean, we all have medical experiences, whether it's with personally or whether it's with family, and obviously the more that doctors can focus on the patient, the part that is, as you said, creative, which is really just a solution for making them feel better, the better. So that is where the automation can be less fearful and more just helping patients have a better experience and hopefully in the long run, better health, which is really what we're all looking for. Well, thank you guys. I learned a lot today, and I know this is not unique to healthcare. We're doing this in a lot of different areas. We really want to help people who have consumers that they want to understand better to go beyond demographics and really understand, to your point, Tim, the why, really understand what they want and how to serve them the best way we possibly can. So I really appreciate your time. It's always great to hang out with both of you.
For our listeners, we hope you enjoyed this episode of Connect with Causeway. Please be sure to subscribe to the podcast and tune in again for our next episode.
To learn more, visit Causeway Solutions to get started!