Leveraging Three Tiers of Health Data
Terry Fouts, M.D., chief medical officer of Emeryville, Calif.-based analytics vendor MedeAnalytics, talks with Health Data Management about the current state of analytics, including the integration of data from outside the clinical domain.
Terry Fouts, M.D., is the chief medical officer of Emeryville, Calif.-based analytics vendor MedeAnalytics. Prior to joining MedeAnalytics in 2008, Fouts served for 19 years as a medical executive for healthcare providers and major health plans, including Health Net, Great-West Healthcare, Aetna, and Cigna, developing policy, wellness and disease management programs, and quality initiatives, as well guiding product development.
He has been responsible for managing national and international programs to reduce medical costs and strengthen health operations. Fouts' medical executive career follows a 19-year career as a practicing internist. He recently spoke with Health Data Management about the current state of analytics, including the integration of data from outside the clinical domain.
HDM: On the proverbial scale of "1 to 10," where would you say the healthcare industry is in terms of comprehending and really making use of the vast array of communication technologies that can leverage clinical and extra-clinical information to deliver a holistic and worthwhile experience for patients?
TF: Those are two different questions. The first is around communications and the second is on analytics. On communications, I would have to go back to personal experience, and I would say we're maybe at a 5. I don't do a bunch of emailing with my doctors, and I don't receive a lot of information on the Internet--a lot of it is done by phone, and every doctor I go to, I fill out the same handful of handwritten documents that ask the same questions.
Now, with Meaningful Use, you have to be able to email patients and share data among and between other physicians, et cetera. People are getting certified for MU, but if you actually look at the rules they don't say you have to do it 100 percent of the time. They say you have to have an electronic medical record with a problem list on x number of patients, and that keeps escalating over the years. The same is true with with email. I think we're about halfway there. In my experience it's not there, but I know what MU is about and that people are getting certified. I'm surmising from that others are beyond the 5--it's just that my pocket of physicians isn't there yet, so I may have a sampling error.
On the analytics side, which is where I actually live, I'll give you a couple curmudgeonly comments. I gave a panel talk on this in Chicago a couple weeks ago. I opened it by saying there are a lot of constructs you have to think about when you think about big data as it applies to a patient. At some point, somebody has to sit across an examining table or couch from a patient and do something and take care of that patient. Everybody gets enthralled with population management, which I totally get and support, but it's not like rounding up a whole herd of cows and giving them a shot.
There were two other primary care physicians and myself on that panel, and we talked about the fact that there are a lot of data points that are not necessarily reducible to data, like the way a patient walks. Or the way they smell--actually, you can smell renal failure. I've had my hand on some guy examining him and I knew he was going to arrest before his vital signs changed, from his temperature and the turgor of his skin. So there are a lot of things even the most capable computer won't pick up that a practitioner will, imperfect as they are and artistic rather than scientific as they are.
All that said, is then the proposition that if we just get more data from Facebook and from your credit report and your buying history at Target, will we take better care of patients? I think the answer is probably somewhat. For one example, we know people who participate in 401(k) retirement plans are much more coachable when it comes to healthcare issues, much more likely to be effective participants in disease management. They understand that they have a financial as well as a healthcare trajectory, and a lot of that is in their hands. It's tantalizing at the very least--have we reduced it to something actionable? I don't know.
I look at health data in three tiers. The first is the basics--claims data, pharmacy data, lab results--we know it's there, and that data is actually pretty good for a lot of things. The second level is, we are now bringing in the clinical data through electronic health records. The great hope was that they would be flawless and would cut down mistakes. They maybe have cut down some, but brought in others, like the cut-and-paste problem, but nonetheless, that data is there now. And the third level is aspirational, like "Can I find out something about you on Facebook that will help me take care of you? What can I find out from your credit report, your marriage and divorce records, your criminal record, the fact you have license to carry a gun?"
All those things do have health implications. But the other thing coming in is the venue of care is changing, and that's one reason we generate so much more data about ourselves. When I started practice, the only way to get a blood sugar reading on you was you had to come in to the office and we'd draw blood out of your arm and run it through a laboratory. That was around 1970-something. Now you can do it at home with great accuracy, and that's true of lots of biometrics. So, the home has become a venue of care. And that data is easily transferable. That's probably the next level, and I think that's more promising than credit scores and things like that.
HDM: When do you think activity data generated through wearable devices will become a quantifiable part of someone's health record?
TF: It should be now. Data that's generated from FitBits and things like it is all distillable digital data that we should be able to incorporate into the EHR, and it absolutely should play a part. When I see a patient coming in who's a runner, who's focused on their health, I know I'll have a much easier time convincing them to manage behaviors, and behaviors drive 85 percent of our healthcare burden. The other thing that we've faced is, do disease management programs work, and even more difficult to discover, do wellness programs work? I think we have to look at it from a couple different perspectives. I think all of us in our heart of hearts believe these types of programs do improve health, but the other thing they do is, health is not just the presence of normal labs--it is also a sense of well-being, and that's extraordinarily important. We can't ignore that.
One other thing is, at this recent panel, we talked about "death by zip code." One of the important pieces of data that we probably do have on people that will help us predict things is exactly where they live. If you're a single mom with two little kids and your zip code keeps you far from public transport and you don't have a car, your health is going to suffer simply because you can't even get the basics--it's hard to get on a bus with two little kids and do three transfers to get to the doctor. If you live in a zip code with bad healthcare, that will affect your outcomes.
HDM: Will the way we deal with privacy have to change if we plan to truly make the healthcare ecosystem a patient-centric communications infrastructure? A lot of the current regulations inhibit the transfer of information or the way information is transferred, which leaves patients feeling as though they have no say in how their own data is shared.
TF: I think we're all a little paranoid about our information. My son works for Visa and he and I have a chuckle over the fact that people don't want anybody to know they've had an X-ray or seen a doctor, but they will go to Brazil and stick their card into an ATM which potentially opens the world to their entire financial data. I have nothing that I would mind having published, but people with AIDS, mental health conditions, or other issues, probably have every reason they don't want people to have access to their information. It's going to have to be a cultural shift. Culture trumps policy every day of the week. I think all of us are a little sensitive about who gets to know what about us.
HDM: Should patients be given more power in deciding how that information is shared?
TF: That's where it has to go, and it has to have a degree of flexibility like privacy settings on Facebook--this is OK, that isn't. But when you get to health data management and governance, that takes some sophistication. For example, what is and what isn't mental health? Let's say I take a beta blocker for stage fright, which is kind of an emotional/mental health thing. But a beta blocker can also be taken for heart failure, high blood pressure, and all kinds of things. So, there are a lot of gray areas--the blacks and whites are pretty easy, but those gray areas are pretty big.
HDM: What industries will healthcare have to look at to take lessons in how those information-sharing dynamics and policies have to work? Finance? Social platforms? Or is it apples and oranges?
TF: Maybe the answer to that is who do you trust? Being an internist, I've always leaned toward thinking you should trust your primary care doc, and I think you need a trusted adviser--and maybe that's the person you give universal access to. And then you have an agreement with them and your Internet service provider about what will and what will not be revealed. That however, puts the internist in a bad position if he reveals something borderline or graylike "I didn't want you to tell my orthopod i was bipolar."
HDM: You could create a scenario where the orthopod could ask the PC doctor for information, and the PC doctor then asks the patient.
TF: That's right. That's a lot of work. You just covered about five or six calls or emails.
HDM: If you go back to the 1980s and 1990s, doctors hated the first wave of capitated HMO care and called it "denial of care" rather than data-based management of care. How will the present-day ACO, value-based model differ?
TF: I worked in both those environments. I was part of the HMO revolution back in the early '80s. I think--and I'm ashamed of it--but I think I participated in that healthcare denial business back in those days, but things were so far out of whack. Let me give you an example. The average length of stay for gall bladder surgery in 1983 was five to seven days days. You had to essentially be marathon-ready before you left the hospital for anything. Then we realized we can do it with this thing called a laparoscope and that would be one day, and eventually without ever having to be admitted at all. The difference between the capitation and care denial in those days and the ACOs is the ACOs are now oriented towards outcomes--having to measure yourself against quality and outcomes benchmarks is a complete shift.
Back in those days, if you got your mammogram, if you got your flu shot, we put a checkmark in that quality box. Well, that's not anywhere near good enough today. We've changed incentives, we're looking at outcomes, and I think we're also looking at care coordination a little better. And we're starting to say somebody has to be responsible for driving the entire ship--maybe we'll put the patient-centered medical home at the center and they'll coordinate all this and make sure they effectively use specialists, that they get people in on time, that they'll have programs in place to help them identify, triage, and engage people at the proper level. Frankly, that sophistication didn't exist 18 or 20 years ago. I think it's a totally different marketplace and environment now.
HDM: Going back to tracking wellness data, the idea of promoting employee health is nothing new. For example, IBM built country clubs at a lot of its big facilities for its employees after World War II. The company fostered activity like golf, softball, and basketball leagues and swimming in a sort of benevolent paternalism. Will healthcare organizations at local levels have to adopt such practices with local businesses in actively forming wellness activities on a community-wide basis? Will that be a default?
T: I think the next wave is going to be a little different, when you get to the concept of employer paternalism. The next wave is the employer will go to the payer or payers, and want to know the best place for treatment for the whole list of conditions. They'll say: "Show me the data, show me the outcomes so we can decide where they should be getting their care." The next level of paternalism is going to be the employers getting very serious, not just taking somebody's word for it, or just giving people free rein to go and shop and do whatever they want to do--they may still be able to do that, but the employers will want to be the trusted adviser when it comes to finding the right venues of care.
It's a big challenge and I think it's maybe the next challenge. The wellness and disease management ships have already sailed and I think it's doing some good. The intent is certainly good. The next thing is: "Help me find where I should best get my care, and it may not be next door. It may be Houston or Seattle or Cleveland, or wherever. I think we're unwilling to accept the fact every provider is created equally, and as we collect data through these levels we talked about, we'll know who's naughty and who's nice.
He has been responsible for managing national and international programs to reduce medical costs and strengthen health operations. Fouts' medical executive career follows a 19-year career as a practicing internist. He recently spoke with Health Data Management about the current state of analytics, including the integration of data from outside the clinical domain.
HDM: On the proverbial scale of "1 to 10," where would you say the healthcare industry is in terms of comprehending and really making use of the vast array of communication technologies that can leverage clinical and extra-clinical information to deliver a holistic and worthwhile experience for patients?
TF: Those are two different questions. The first is around communications and the second is on analytics. On communications, I would have to go back to personal experience, and I would say we're maybe at a 5. I don't do a bunch of emailing with my doctors, and I don't receive a lot of information on the Internet--a lot of it is done by phone, and every doctor I go to, I fill out the same handful of handwritten documents that ask the same questions.
Now, with Meaningful Use, you have to be able to email patients and share data among and between other physicians, et cetera. People are getting certified for MU, but if you actually look at the rules they don't say you have to do it 100 percent of the time. They say you have to have an electronic medical record with a problem list on x number of patients, and that keeps escalating over the years. The same is true with with email. I think we're about halfway there. In my experience it's not there, but I know what MU is about and that people are getting certified. I'm surmising from that others are beyond the 5--it's just that my pocket of physicians isn't there yet, so I may have a sampling error.
On the analytics side, which is where I actually live, I'll give you a couple curmudgeonly comments. I gave a panel talk on this in Chicago a couple weeks ago. I opened it by saying there are a lot of constructs you have to think about when you think about big data as it applies to a patient. At some point, somebody has to sit across an examining table or couch from a patient and do something and take care of that patient. Everybody gets enthralled with population management, which I totally get and support, but it's not like rounding up a whole herd of cows and giving them a shot.
There were two other primary care physicians and myself on that panel, and we talked about the fact that there are a lot of data points that are not necessarily reducible to data, like the way a patient walks. Or the way they smell--actually, you can smell renal failure. I've had my hand on some guy examining him and I knew he was going to arrest before his vital signs changed, from his temperature and the turgor of his skin. So there are a lot of things even the most capable computer won't pick up that a practitioner will, imperfect as they are and artistic rather than scientific as they are.
All that said, is then the proposition that if we just get more data from Facebook and from your credit report and your buying history at Target, will we take better care of patients? I think the answer is probably somewhat. For one example, we know people who participate in 401(k) retirement plans are much more coachable when it comes to healthcare issues, much more likely to be effective participants in disease management. They understand that they have a financial as well as a healthcare trajectory, and a lot of that is in their hands. It's tantalizing at the very least--have we reduced it to something actionable? I don't know.
I look at health data in three tiers. The first is the basics--claims data, pharmacy data, lab results--we know it's there, and that data is actually pretty good for a lot of things. The second level is, we are now bringing in the clinical data through electronic health records. The great hope was that they would be flawless and would cut down mistakes. They maybe have cut down some, but brought in others, like the cut-and-paste problem, but nonetheless, that data is there now. And the third level is aspirational, like "Can I find out something about you on Facebook that will help me take care of you? What can I find out from your credit report, your marriage and divorce records, your criminal record, the fact you have license to carry a gun?"
All those things do have health implications. But the other thing coming in is the venue of care is changing, and that's one reason we generate so much more data about ourselves. When I started practice, the only way to get a blood sugar reading on you was you had to come in to the office and we'd draw blood out of your arm and run it through a laboratory. That was around 1970-something. Now you can do it at home with great accuracy, and that's true of lots of biometrics. So, the home has become a venue of care. And that data is easily transferable. That's probably the next level, and I think that's more promising than credit scores and things like that.
HDM: When do you think activity data generated through wearable devices will become a quantifiable part of someone's health record?
TF: It should be now. Data that's generated from FitBits and things like it is all distillable digital data that we should be able to incorporate into the EHR, and it absolutely should play a part. When I see a patient coming in who's a runner, who's focused on their health, I know I'll have a much easier time convincing them to manage behaviors, and behaviors drive 85 percent of our healthcare burden. The other thing that we've faced is, do disease management programs work, and even more difficult to discover, do wellness programs work? I think we have to look at it from a couple different perspectives. I think all of us in our heart of hearts believe these types of programs do improve health, but the other thing they do is, health is not just the presence of normal labs--it is also a sense of well-being, and that's extraordinarily important. We can't ignore that.
One other thing is, at this recent panel, we talked about "death by zip code." One of the important pieces of data that we probably do have on people that will help us predict things is exactly where they live. If you're a single mom with two little kids and your zip code keeps you far from public transport and you don't have a car, your health is going to suffer simply because you can't even get the basics--it's hard to get on a bus with two little kids and do three transfers to get to the doctor. If you live in a zip code with bad healthcare, that will affect your outcomes.
HDM: Will the way we deal with privacy have to change if we plan to truly make the healthcare ecosystem a patient-centric communications infrastructure? A lot of the current regulations inhibit the transfer of information or the way information is transferred, which leaves patients feeling as though they have no say in how their own data is shared.
TF: I think we're all a little paranoid about our information. My son works for Visa and he and I have a chuckle over the fact that people don't want anybody to know they've had an X-ray or seen a doctor, but they will go to Brazil and stick their card into an ATM which potentially opens the world to their entire financial data. I have nothing that I would mind having published, but people with AIDS, mental health conditions, or other issues, probably have every reason they don't want people to have access to their information. It's going to have to be a cultural shift. Culture trumps policy every day of the week. I think all of us are a little sensitive about who gets to know what about us.
HDM: Should patients be given more power in deciding how that information is shared?
TF: That's where it has to go, and it has to have a degree of flexibility like privacy settings on Facebook--this is OK, that isn't. But when you get to health data management and governance, that takes some sophistication. For example, what is and what isn't mental health? Let's say I take a beta blocker for stage fright, which is kind of an emotional/mental health thing. But a beta blocker can also be taken for heart failure, high blood pressure, and all kinds of things. So, there are a lot of gray areas--the blacks and whites are pretty easy, but those gray areas are pretty big.
HDM: What industries will healthcare have to look at to take lessons in how those information-sharing dynamics and policies have to work? Finance? Social platforms? Or is it apples and oranges?
TF: Maybe the answer to that is who do you trust? Being an internist, I've always leaned toward thinking you should trust your primary care doc, and I think you need a trusted adviser--and maybe that's the person you give universal access to. And then you have an agreement with them and your Internet service provider about what will and what will not be revealed. That however, puts the internist in a bad position if he reveals something borderline or graylike "I didn't want you to tell my orthopod i was bipolar."
HDM: You could create a scenario where the orthopod could ask the PC doctor for information, and the PC doctor then asks the patient.
TF: That's right. That's a lot of work. You just covered about five or six calls or emails.
HDM: If you go back to the 1980s and 1990s, doctors hated the first wave of capitated HMO care and called it "denial of care" rather than data-based management of care. How will the present-day ACO, value-based model differ?
TF: I worked in both those environments. I was part of the HMO revolution back in the early '80s. I think--and I'm ashamed of it--but I think I participated in that healthcare denial business back in those days, but things were so far out of whack. Let me give you an example. The average length of stay for gall bladder surgery in 1983 was five to seven days days. You had to essentially be marathon-ready before you left the hospital for anything. Then we realized we can do it with this thing called a laparoscope and that would be one day, and eventually without ever having to be admitted at all. The difference between the capitation and care denial in those days and the ACOs is the ACOs are now oriented towards outcomes--having to measure yourself against quality and outcomes benchmarks is a complete shift.
Back in those days, if you got your mammogram, if you got your flu shot, we put a checkmark in that quality box. Well, that's not anywhere near good enough today. We've changed incentives, we're looking at outcomes, and I think we're also looking at care coordination a little better. And we're starting to say somebody has to be responsible for driving the entire ship--maybe we'll put the patient-centered medical home at the center and they'll coordinate all this and make sure they effectively use specialists, that they get people in on time, that they'll have programs in place to help them identify, triage, and engage people at the proper level. Frankly, that sophistication didn't exist 18 or 20 years ago. I think it's a totally different marketplace and environment now.
HDM: Going back to tracking wellness data, the idea of promoting employee health is nothing new. For example, IBM built country clubs at a lot of its big facilities for its employees after World War II. The company fostered activity like golf, softball, and basketball leagues and swimming in a sort of benevolent paternalism. Will healthcare organizations at local levels have to adopt such practices with local businesses in actively forming wellness activities on a community-wide basis? Will that be a default?
T: I think the next wave is going to be a little different, when you get to the concept of employer paternalism. The next wave is the employer will go to the payer or payers, and want to know the best place for treatment for the whole list of conditions. They'll say: "Show me the data, show me the outcomes so we can decide where they should be getting their care." The next level of paternalism is going to be the employers getting very serious, not just taking somebody's word for it, or just giving people free rein to go and shop and do whatever they want to do--they may still be able to do that, but the employers will want to be the trusted adviser when it comes to finding the right venues of care.
It's a big challenge and I think it's maybe the next challenge. The wellness and disease management ships have already sailed and I think it's doing some good. The intent is certainly good. The next thing is: "Help me find where I should best get my care, and it may not be next door. It may be Houston or Seattle or Cleveland, or wherever. I think we're unwilling to accept the fact every provider is created equally, and as we collect data through these levels we talked about, we'll know who's naughty and who's nice.
More for you
Loading data for hdm_tax_topic #better-outcomes...