Summary:

Speaking at GigaOM’s Structure:Data conference on Wednesday, Aetna’s head of innovation, Michael Palmer, talked about the company’s efforts to use patient data to provide better care.

Structure Data 2013 Michael Palmer Aetna
photo: Albert Chau


Session Name: Improving Your Primary Product With Big Data Insights.

Speakers: S1 Announcer S2 Ki Mae Heussner S3 Michael Palmer

S4 Audience member

ANNOUNCER 00:00

… by Ki Mae Heussner, she’s a staff writer with GigaOM, and she’s going to be talking with Michael Palmer, the head of innovation at Aetna. Please welcome Ki Mae and Michael to the stage.

KI MAE H 00:18

So thank you so much for coming out today.

MICHAEL P 00:19

It’s great to be here.

ANNOUNCER 00:20

Thank you.

KI MAE H 00:21

So about a year ago, Aetna launched Innovation Lab, which you lead. It’s a very cool kind of test kitchen, looking at all the different ways you can use big data. Can you talk a bit about what the goals of that lab are, and the thinking behind it?

MICHAEL P 00:34

We look at clinical innovation, new platforms and patient engagement. As we think about the prevalence of diseases, we look at the cost of diseases and how much they are impacting people’s lives. For example, if you think about cardiovascular disease or cancer care, they are huge problems. We spend five billion dollar a year on cancer care on behalf of our customers, six and a half billion dollars on cardiovascular disease. We look at the places where there is a lot of prevalence, and it really impacts people’s lives.

KI MAE H 01:06

Aetna has 30 million customers – a lot of data from prescriptions to data to lab data. With all of that data, how do you determine where to focus? How do you decide what your priorities are?

MICHAEL P 01:19

You look at big problems. I’ll use cardiovascular disease as an example. For those not familiar, metabolic syndrome is a big problem. It’s a combination of obesity – large waist circumference, high triglycerides, high blood pressure, low HDL cholesterol et cetera. If you have three out of five of those factors, that means you have metabolic syndrome. What does that mean? It means you’re basically two times more likely to get cardiovascular disease and five times more likely to get diabetes. That can end in chronic kidney disease and being on dialysis: bad stuff.

MICHAEL P 01:51

The problem with these diseases is that they take a long time to manifest. But if you manage your blood pressure and your cholesterol et cetera, over the ten, 15, 20 years that you have those outer range factors, you can really improve your health overall long term. We look at these diseases that a lot of people have. They create a lot of cost in the system and they create a lot of challenge in people’s lives.

MICHAEL P 02:12

Another one is cancer care. We spend about five billion dollars on that a year. Any family who has been afflicted by cancer knows that’s a life-changing event for you and your family. So the idea of using some big data to drive the first intervention – being right the first time – is really critical. If you can catch the cancer the first time, and you can catch it early, then many cancers can be staved off, delayed or actually cured. Being able to have that right treatment the first time for patients and their families gets them back to normal life as quickly as possible and creates higher quality outcomes.

KI MAE H 02:49

So I know that this has not even been in existence for a year yet, but to this point, can you talk about any gains you have seen or any outcomes that you have been able to achieve?

MICHAEL P 02:59

I’ll stay with the metabolic syndrome example. A lot of our customers are large-plan sponsors that have 100, 000 members or less, let’s say. Sometimes their costs for healthcare are over a billion dollars. If they look at trying to keep the health of their population high, keeping people at work, and trying to manage that one billion dollars of medical costs, all of those things are combined. In metabolic syndrome, we’re working with our large customers to say, ” What are the right interventions you can do with your workforce, to allow them to improve their health in metabolic syndrome?”

MICHAEL P 03:33

An example is that in the big data, we did some analysis: 18 million members’ claims, their lab results, their pharmacy data, et cetera. We munched all this together and found that we can actually help predict, at the individual level, at what probability someone is to get metabolic syndrome in the coming year. In fact, of the five factors, which ones you are likely to get next. That way, it allows the individuals to say, Wow, it’s a wake-up call. If I don’t do something about this, I am going to end up with metabolic syndrome next year, and I’m going to end up with heart problems and diabetes et cetera.” To the extent that we can, in a way, give people more information about themselves at the individual level, and say Heres your probability of high glucose next year, you really need to work with your doctor on this.” That gives them a sense of not having to work on all five factors, but just the one or two that are going to make a difference in the coming year.

KI MAE H 04:25

How do you think this will change how Aetna does business with its customers? If it’s not changing now, how do you expect it to change in the future?

MICHAEL P 04:38

As you look at cost, quality and convenience as three components of healthcare, we have got lots of new tools – iTriage, CarePass – that are helping make the capabilities of our customers to allow our patients to be more healthy, easier and to manage their own data et cetera. If you take this big data, you can push these capabilities out to the individuals at the individual level, and give them access to this data so they can better manage their own health.

MICHAEL P 05:06

The large customers that we have are also looking to manage, as I said, that billion dollars. Those with cardiovascular disease, people who have metabolic syndrome, are about one point six times as costly as the average population in terms of medical costs for the year. So if they can drive down those individual factors in their population, it also saves them money and keeps them healthier and at work.

KI MAE H 05:27

From the customer perspective, on one hand it’s great and amazing that this data can potentially lower costs and improve care. But the other side of it is that you learn that I have some predisposition to some condition, or you know that I am not as healthy as somebody else. How do I know that I am not going to be denied coverage, or that it is not going to hurt me in some other way?

MICHAEL P 05:56

Part of the provisions in the Affordable Care Act make sure that we can’t use pre-existing conditions to deny you care and that kind of thing. Overall, we are trying to make sure our products and services reach a broad population. We have somewhere on the order of 30 million new members coming into the system in 2014, on the individual exchanges. As we drive that new population, we’re going to have to help manage their health. This is a population that in many cases, haven’t had access to healthcare before in a consistent way. Many times they are looking at, Do I go to the E. R?” et cetera. We’re trying to put tools in people’s hands to allow them to use the data at their fingertips. iTriage is an example. You can do a symptom checker and ask what is the most cost-effective place for me to go get treatment, and what is going to be the best outcomes. Matching that cost and outcomes curve is going to be important for our customers.

KI MAE H 06:43

The other tricky thing from a customer perspective is, what if it’s not even an issue of data that you find, but data I don’t want to give you. There was an example this week about a CVS that was compelling its employees to do a screening to get their blood pressure and glucose checked. If they did not do that, they were going to have an increased premium. How do you balance incentivising people to do the well thing”, and not discriminate against people who might not be as healthy?

MICHAEL P 07:12

It’s an interesting challenge, trying to motivate people to do what is best for their own health. I fight the battle of the bulge, I am sure many people do, trying to manage what you eat, your exercise, your activity and that sort of thing. A lot of our large employers, CVS among them, try to figure out ways to incent people to do the right thing for themselves. We have all sorts of different plan designs for getting people to do the right thing and take their medications et cetera. We had an interesting study for folks who were post- having a myocardial infarction – having a heart attack. We said, Well waive all the co-pays for the drugs you need so you don’t have another heart attack.” Before the study, we had a little higher than 40% of the people actually taking the drugs. We waived all of the monetary fees for taking the drugs, any co-pays et cetera. We only inched it up about five or six points, so still under 50%.

MICHAEL P 08:02

Getting people to take the medications that they know are good for them, that they know will avoid, in that case, a heart attack coming on, is still a challenge. Employers are trying to figure out how to incent people in different ways? For some people it is giving them a carrot, for some people it’s a stick. Some people are trying new techniques. Whether it is medication adherence – how do you take the drugs? We could send you a text message, we could call you, we can give you an iPhone app et cetera. All of those are particular solutions that might work for certain populations. We literally are trying different solutions for different kinds of folks. In the case you gave, they are trying a penalty as opposed to an incentive.

KI MAE H 08:39

With this [inaudible] and looking at this data, and bringing that mindset to the rest of the company, what have been some of the challenges you have experienced so far?

MICHAEL P 08:48

Top of the list: privacy concerns. People are always afraid that their data is going to get used in improper ways. We are very sensitive to that. We have been doing this for years – de-identifying data, using large cohorts of data to drive decision-making and that sort of thing. We still have to be careful with the HIPPO rules so that personal information does not get in the wrong hands. At the same time, we want to be able to provide those individual interventions at the individual level so the information goes between us and the patient and/or us and the provider. That little loop needs to stay highly secure. So that’s always one of the challenges.

MICHAEL P 09:27

I think another one is that we’ve had a lot of great data over the years from claims, from pharmacy and from some lab data. But we don’t really have access to a lot of the clinical data. So as the world moves to these accountable care organizations, where the doctors and the insurance company and the patient are all trying to work together to manage the patient’s health. All the incentives will finally be aligned. We are trying give the providers access to better clinical decision support tools. We bought a company a while ago called Active Health, and they have some great tools to put in providers hands to allow them to drive that shared decision-making with the patient, and get the better outcomes. When that tide turns and people are moving more into these accountable care organizations, the incentives will be aligned to drive better care at lower cost.

KI MAE H 10:12

Privacy is obviously very important in healthcare but applies to other industries also. Are there other things that you have learned that you feel like apply to other industries beyond health?

MICHAEL P 10:23

It occurs to me that most other industries, especially as we see the retailization of healthcare with more and more consumer interaction– We’re seeing the activities that the large retailers are doing. You take Amazon’s website for example, their recommendation engine. People should be able to log on to their healthcare company’s website and get recommendations to improve their health. Well, they get those today. Sometimes they are not as personalized as they could be. My sense is that if we can all begin to emulate some of those capabilities that some of the top retail brands are able to do in healthcare, in banking, in other places, we would all be better off. I think emulating some of the best. Taking some of those big data insights that we have, and trying to drive them down to the individual level, and give people really personalized recommendations.

KI MAE H 11:16

What else are you looking to in terms of helping patients that we haven’t talked about? Things on the horizon and other interesting uses for big data?

MICHAEL P 11:24

I think a couple of them are drug safety. We’ve got a tremendous amount of drug information as well as treatment information. We’re working on some tools called Pharmaco Vigilance, which essentially is drug safety. We’re trying to drive those tools so that we can predict when someone is going to have an adverse drug event. We can see those more in the data than today, and be able to give the docs better information when they are prescribing medications for their patients, to say, Prescribe this not that for this particular patient, for these reasons.” It’s almost a version of comparative effectiveness research, but the data that is out there is so huge right now. We can begin to use some of the tools that are in the room next door, literally – some of the tools that vendors have got here – to be able to slice and dice that data. This will be able to provide that individual insight for the physician as they are giving the prescription to the patient.

KI MAE H 12:10

How far away do you think we are from incorporating genetic data into that combination of data?

MICHAEL P 12:18

There are lots of great laws about making sure that genetic data that you have – the law is called GINA – does not get into the decision-making around your healthcare from an insurance company perspective. We are actually doing a pilot with a few companies right now, allowing the patient to get their own genomic information. They get some genetic counselling along with that, and that, we think, will drive a wellness program, driven by genomics. The sense is, if you find out your ultimate demise is going to be cardiovascular disease, there really are preventable factors that you can put in place in your life to avoid that. So to the extent that we can use that information as a first step to helping drive the individualization, that is, people taking what is going to happen to them next personally. We have a long list of things we want to do with genomics but obviously it is going to take some time. The science is a little bit nascent right now, but once we test the wellness programs out and see if they are going to work– Pharmaco genomics, being able to have docs have information on how best to prescribe medications for you based on your personal genome and eventually what treatments are going to be best for you. The Pharmaco genomics is going to be a little bit easier than some of the other features going forward. It’s probably a five-year plan for us.

KI MAE H 13:33

Do you see a place for using social data in all that? People online commenting about their condition and their reactions to medication. Do you see this combining with that to help tailor treatments down the line?

MICHAEL P 13:47

I certainly can, and there are companies that are doing that. They are taking the social data and the medical data – to the extent that they have it. The challenge is that a lot of the HIPAA laws prohibit the ability to connect those two in ways that would be ultimately useful. We have to – our industry and across the industries – figure out ways to do that meet the regulations and the privacy concerns, but also allow like you said, for the fact that not everyone is the same. I believe there are many ultimate strategies for getting people to take their medications or to drive the wellbeing of themselves and their families. We have to come up with a variety of different strategies that will work for different people and improve those over time.

KI MAE H 14:27

We have talked a lot about patient care, and you have talked a bit about how that can help doctors. Are they other things that you are doing or that you envision doing that can help doctors with their job?

MICHAEL P 14:37

If you look at clinical decision support tools, we’re a big fan of evidence-based medicine. If you look at cancer care as an example, a huge percentage of cancer care is not practiced in an evidence-based way. Our goal is to put as many tools in the hands of providers so when they are making decisions on which treatment protocol is best for the patient; we bring all the patient data that we can. For example, we have a company called Medicity which is the largest health information exchange in the country. We have 250, 000 docs on that platform and all of that clinical data can move to the oncologist’s desktop once the person is diagnozed with oncology. Now the oncologist has access to all the information from your history, as well as the latest and greatest information about the cancer that you are faced with. They can put the staging information et cetera in, then they can choose from one of the evidence-based protocols that will be best for you. The goal is to get you the first treatment right the first time, get you back to your family normalcy, and to try and reduce the amount of time it takes to get your treatment done.

KI MAE H 15:41

You have also started working with a lot of other start-ups like Fitbit. Can you talk about how some of those partnerships, and other companies and startups are using the data that you have?

MICHAEL P 15:54

We have partnered with a number of startups in our last year of existence. An example would be GNS Healthcare. This is a company that helped us with our metabolic syndrome work. They were a very small startup and we worked with them very nicely to get the analytics done to create these histograms for the individual patients and to help our large companies understand the cost implications if they intervene in certain ways. We are working with half a dozen startups in any point in time. We also work with some bigger players, when it’s appropriate. I think what we are trying to do is to drive things at speed, and a lot of times smaller companies can work with us even better than some of the larger companies.

KI MAE H 16:35

Interesting. Can you name a couple of the other one that you are working through this with?

MICHAEL P 16:37

A company called Evity is working with us in cancer care. There’s a number of them. Obviously [inaudible] in the coming year we are looking at things around medication adherence. What in big data will help us understand what works for certain patients and does not work for others? We’ll be looking for additional partners in the coming year to help us with that kind of problem.

KI MAE H 16:56

We are almost at time, but are there any questions from anyone in the audience?

AUDIENCE MEMBER 17:03

In the healthcare market there are a lot of new successful business models, it is a little bit of a different market. For example, Pharma and insurance have plenty of money. They are giving tools to doctors for free so that they can get the data back. Can you talk about which new business models are being successful?

MICHAEL P 17:21

Sure. As we look at our business, it is an eco-system, really, of trying to drive better quality and lower costs for patients. We are seeing a variety of business models. As an example, we have a company we are just launching called CarePass that allows the Fitbits of the world, the weight loss programs et cetera, to combine all your data into one place so that you as the individual can begin to manage your own healthcare even better. That conglomeration of data, being able to bring it all together, mash it up in certain ways, can help individuals, again, see better what they can do to influence their health. I think the personal health revolution is definitely a factor these days. As we have these 30 million new people come into the healthcare world, I think that is going to be a really important part of the eco-system.

KI MAE H 18:10

Great. Thank you so much for your time.

MICHAEL P 18:11

Great, thank you all.

[applause]

You’re subscribed! If you like, you can update your settings

firstpage of 2

Comments have been disabled for this post