How Aetna-CVS is Helping People Navigate the Health Care System

Updated: Nov 30, 2021


Interview between Lygeia Ricciardi of AdaRose and Rahul Kak of Aetna discussing artificial intelligence in behavior change marketing for health care

Who doesn't want to make managing their health and health care easier? But changing unhealthy habits and navigating the fragmented health care system can be challenging at best.


Aetna-CVS, the largest health company in the US, is shaking things up. It also happens to be run by the woman at the top of Fortune's Most Powerful Women of 2021 list, Karen Lynch, who stepped into her role just last February during the pandemic.


Lynch has put forward a bold vision that leverages her company's vast reach. The plan includes transforming hundreds of CVS drugstores into health "super clinics" that offer primary care services and mental health counseling in addition to prescription refills and over-the-counter sales. It's clear that Aetna-CVS is also leveraging data and analytics at every step of the way.


In this video, Rahul Kak, who was recently promoted to VP of Customer Health and Experience at Aetna-CVS joins Lygeia of AdaRose to discuss how the company is using data, analytics, and artificial intelligence (AI) to help its members achieve better health in a more seamless way.


Even if you aren't a member of Aetna or don't shop at a CVS (though 80% of Americans are within 10 miles of one) the way Aetna-CVS is holistically approaching customer care and leveraging technology will undoubtedly have implications for the rest of the country. So check it out and see what's in store for you and your family.



Interview Transcript


Lygeia:

Hello and welcome to the AdaRose digital health and wellness video podcast. I'm your host, Lygeia Ricciardi and our guest today is Rahul Kak. He is the Executive Director of Behavior Change Marketing at Aetna. Welcome Rahul! We're totally excited to have you.


So, some background before we jump into questions with Rahul. Aetna, your employer merged with CVS a couple of years ago. The combined entity is the single largest healthcare entity in the U.S. You provide benefits that cover about 40 million Americans. And some of these insurance plans are private, like employer plans, others are public, more like Medicare and Medicaid.


And I'm pretty sure that we have a number of Aetna customers listening. I am one of them. Happy to be one of them. I've been a member for a number of years. But even if you personally are not an Aetna member, what Rahul is going to tell us about is relevant to you because the kinds of changes that they're making and how they're using behavior change modification really has ramifications for all of us.


So again, whether you're getting [insurance] directly through Aetna right now or whether your plan is one of the many who's watching what Aetna does, the ideas that he talks about today are going to have an impact on you and your family at some time in the near future.


Welcome Rahul. We're very excited to have you.


Rahul:

Nice to be on this. Thank you. It's a pleasure to meet you, Lygeia, and be on your program.


Who is Rahul Kak and what does his role as Executive Director of Behavior Change Marketing mean? (1:31)


Lygeia:

Fabulous. So, given this context of Aetna, can you tell us a little bit about your specific role and what your title, Executive Director of Behavior Change Marketing really means? [Note: Rahul was recently promoted to VP, Customer Health and Experience at Aetna-CVS.]


Rahul:

Yes. It's not a common title you see at a lot of companies but one familiar way it's been described in other settings has been more on population health management. And that's the common way I usually describe what the work I do is.


So this is basically using marketing channels and other digital channels to proactively nudge people on their path to better health. And so this is going to come in the context of educational content and mailers, digital media, phone outreach, pharmacist discussions. And we can talk more about that, but in our case it's oftentimes powered by predictive learning models.


And a bit more detail on what we mean by "population health management." It's actually not just educating people about how to manage their own personal health, which I might traditionally classify as a category called "wellness." So what we're doing is guiding people on how they should be or could be interacting with their healthcare system, which is a lot more complex.


And that will be things like, where do I go for care? Am I getting the right treatments to manage my chronic conditions? Then ultimately, what should I talk to my doctor about?


Nudging consumers to make healthier behavior change decisions and navigating the healthcare system (2:56)


Lygeia:

So, with that kind of dual purpose, I would love to dig into a little bit more of the specifics in each area. So for example, on the wellness side, what might be some examples of the kinds of behavior that you can help to nudge people with?


Rahul:

Yeah. And when I say "wellness," so wellness would be like mind and body, how do you manage it better? Usually these are things you can do on your own without interacting with the health system necessarily. So wellness would be things like eating right, sleeping early, or in a consistent way, or exercising frequently.


Now, when we talk about the focus of my group, which is on this more population health management and interacting with the healthcare system, it will be, I'm going to give you some specific examples now. So, it could be navigating folks to primary care locations or nearby urgent care options instead of using an expensive ER for unnecessary categories of care was a very specific example and like a very common one that any insurance company and providers are thinking about.


Other examples would be, how do you educate people to get certain preventative screenings like a mammogram or colon cancer testing? And then you can get more clinical in nature where it could be, what should you talk to your doctor about? Maybe considering adding a certain specialist to their care team or adding a specific type of drug to your regimen, because maybe we're predicting you to be at higher risk for a certain chronic condition.


So it becomes like a very expansive category that can also get very condition-specific, and it could be either predictive or deterministic based on exactly what kind of diagnoses you have.


Lygeia:

Yeah. You gave the population health analogy, and I know that many consumers may not be as familiar with that as folks in the healthcare world. But sort of in brief, I think in part, just to clarify that for folks, that's partly about being able to look across the population and say, hey, we should really be focusing on these people because they're at risk of something relative to others.


What might be an example of how that would work? So again, I mentioned I'm an Aetna member personally. Let's say I, let's get a real life scenario. I want to get a COVID shot. Or maybe you don't even know that yet, but you might assume that I do because of where we are in the pandemic and all.


So how might you use predictive analytics to maybe drop in and even figure out, I guess you would know whether I've had a shot or not because it would have gone through billing. Do you use that information?


How health insurers can better support their members through population health management (5:41)


Rahul:

Yeah, you can. So, I kind of casually dropped the terms like "deterministic" and "predictive." So let me break that down and to answer your question. I think there's probably a targeting component, and then the educational component, which will be like using channels to reach out to people.


So on the targeting component, we could do simple stuff where it's deterministic. We know, for example, as a health plan, what prescriptions you're on or past diagnosis codes or pre-certs on a procedure, and we could use this to think, okay, what's the next best action based on this specific thing we know with a 100% certainty this person has gotten or is managing?


And then there's a category of more predictive which is to say, we think statistically there's a high probability of determining that you are likely to need certain kinds of care, and then we're going to trigger a message to send to you. And so this is where data science is coming in: they're building machine learning models and training them with big data, all of these data points that sort of surround the history of this person's health behavior, consumer behavior, social determinants of health, engagement with certain channels.


And we'll use these past data points to predict the most relevant message at the best time or best pace. And this is, it's scientific but it's not perfect. So you have to be sort of thoughtful about, as a marketer, you have to be really thoughtful about messaging and acknowledging that you're not going to get it right all of the time. But we're doing this sort of in the best interest of what we think may be relevant for you. And then I can sort of talk about the channels.


What are the communication channels available to reach consumers to encourage healthier behaviors? (7:20)


Lygeia:

Yeah, no, I'd love to ask you about the channels actually, specifically. So you want to target somebody, what kinds of channels do you have at your disposal?


Rahul:

Yeah. And the most traditional example that is probably undoubtedly the best way to help someone manage care, is to just have a live person. And many insurance companies do this, and managed care organizations do this, providers will do this where they'll have a care management staff, and it could be nurses or some sort of clinician.


They don't have to be clinical in nature, they could be some sort of case manager. But someone who can help walk you through your healthcare journey and make sure you.. simple things, maybe talk to you and encourage you to see a coach to make sure you're adhering to drugs and following the directions provided by your doctor.


And so, that is sort of the ultimate thing. But it's not scalable. So where my group comes in is, we're trying to supplement what I just described. And we're sort of adding scalability through different types of channels in automated journeys.


So then the channels we can think about, direct mail, email, text messages, there's phone calls, and that can be both live, say, with a care manager or automated phone calls, there's mobile app notifications, personalized micro-sites, paid digital media. We have an Apple watch program.


And I think sort of a crowning achievement in this sort of portfolio would be activating pharmacists. And so for my company for example, we have a footprint of 10,000 pharmacies, and then we have an ecosystem of minute clinics and health hubs.


And so, those are other tactics we can use if we're sending a message to a pharmacist, that can help reinforce certain recommendations that we can share if someone is in that setting. And really any of these settings I mentioned.


Real life application of using artificial intelligence to reach health plan members effectively (9:24)


Lygeia:

So that's a super-cool way to take advantage of this integration essentially between the role as insurer and the CVS piece. So again, just thinking in totally practical terms, so interestingly enough, so I am an Aetna member and I do often go to CVS. So let's say I go to CVS. Maybe I am taking an antibiotic for a particular infection that I have.


So you guys, your group would be able to ensure that essentially somewhat basic contextual information about me goes and is there sort of like a flag for the pharmacist to like, hey, this person's an Aetna member, while she's here you should probably talk to her about getting a flu shot or whatever else?


Rahul:

That's exactly right. You basically nailed it. There's a system in place where we can empower our pharmacists, and our pharmacy techs, to deliver certain types of messages that are relevant for them if there's a match and we know that's specifically an Aetna member.


And that's beneficial because when you're a pharmacist, if you're a pure play pharmacy, you don't necessarily have all this other ecosystem of relevant background medical information, you just have specifically what you have in the system for their drugs. But when you're an Aetna member, there's all this other stuff.


And these are things by the way, these are still very early stage, our companies have merged in the past two to three years, but that is live and active in some types of behavior changes now. And we're in the process of expanding, testing, rolling out.


And it's tricky because you also, there are certain things that may be less appropriate for a pharmacist to talk about. So, we have to make that balance very clear. There's privacy implications, too. So this is something that's being treaded carefully but it's certainly an opportunity.


Lygeia:

Absolutely. That makes sense. And then would the pharmacist have an opportunity to give feedback to others within the broader Aetna system, too?


Rahul:

Yeah. It's a two-way feedback loop and we're close with our folks in the retail pharmacy organization, and it's a partnership to build a program like this and basically create a proper integrated system for continuity of care.


How is continuity of care ensured across health care providers through technology? (11:31)


Lygeia:

Yeah. I can imagine that the, that's what I was going to ask you about next actually. This continuity and how you manage to pull all these pieces together and keep it all updated. Do you sort of presume that many of the members are going to be using an app primarily, or is that not necessarily even part of the picture?


Rahul:

Yeah, it’s part of the picture. It’s probably the most ideal scalable interface because we don't pay money to put a message there. You don't have to pay postage, you don't have to pay Google or Facebook a fee to put a message out there.


I think that and pharmacists, the pharmacists, are also not necessarily scalable if it's only at a certain point in time when you're physically there, and not everybody actively has a drug with our pharmacy systems.


So absolutely, a mobile app is a great platform for that. It doesn't have to be the mobile app. It could be just logging into our, the web version of our authenticated sites. So all of those things are great. We recognize it's not a large percentage of people yet. So until then, there's a lot of coordination with all these other channels I mentioned.


How personalized are these automated “nudges” for members? (12:48)


Lygeia:

Gotcha. I'm curious about, you've talked about how information and data from numerous sources can help you to kind of target the kinds of information that are appropriate for people. Are you also able to, in terms of the nudging aspect, do you also personalize in terms of like this person prefers to hear things through text or they prefer a particular tone?


How much down that road of personalization are you going, or is the personalization mostly about sort of the medical and predictive qualities of what someone might need? Or does it get into it at all like, well, this person prefers to have messages in a certain way?


Rahul:

Yeah. So there's different layers of personalization here. Let me talk about a couple of them. So, one layer is around targeting. And so I've already talked about that, and we'll use machine learning to suggest maybe there's a higher probability that a message is going to be relevant to you. So that's one level of personalization.


If you're getting the message, you can still personalize to groups. So we're not at one-to-one yet, but let's talk about that and kind of like go down the funnel. So at the group level, for example, let me give you an example with mammograms. If you're trying to reach Medicare women and a stat is in our population, it might be that around 30% of them are not getting mammograms when they should be on some regular cadence.


So you could reach out to a cohort that is maybe not getting it because they're managing too many chronic conditions and they're not thinking about preventative screenings. And that's different than a cohort that is physically too far from an actual screening clinic.


And you can just as easily imagine how those two cohorts, the messaging is automatically different for that. And you can very easily think about as a marketer, or even just as a human, how do you talk to those women differently? So that's like kind of cohort level.


And then there's the level of one-to-one personalization which is going to be based on their specific health condition. And as much as possible, we will try to maximize personalization on all these layers. You have to be limited to privacy compliance channels such as enveloped mail or being logged into an app once you get that specific where you've referenced someone's condition.


And you asked about preferences, preferences are interesting. We're not actively told if someone preferences these channels. The main thing is, are you permissioned into certain channels? And so those will be acknowledged. But what I'll tell you about preferences and what's working is that multi-channel touch points are most effective at successfully driving changes to health behavior. Any marketer would probably tell you that makes sense.


And there's a certain level of predictability you have to say that somebody who is already actively using certain digital channels, whether it's email or mobile app, they're probably already on their own kind of a person who's actively managing their health because they're also likewise choosing to get messages from their health insurer or their provider.


They're intrinsically more engaged with their health than that. So I'm not saying it's causative there, but there's a strong correlation with people who just opted into permissions.


How is behavior change using AI in health care measured? (16:21)