In-Depth: How patient generated health data is evolving into one of healthcare's biggest trends

By Brian Dolan
Share

What patient generated data used to be and why it is increasingly important

StevenSteinhubl Dr. Steven Steinhubl

Up until very recently most people have had only one source for the information they could share with their doctors during a visit: How they felt. Observable symptoms, self-reported healthy and unhealthy behaviors, medical conditions or issues that other physicians had diagnosed them with previously -- those were among the few types of patient generated data before the advent and wide availability of consumer medical devices. While thermometers, weight scales, blood pressure monitors, and glucose meters are on the shortlist of medical devices patients have used at home for decades, today home diagnostic devices, medical sensors and other biometric monitors have proliferated.

At a recent healthcare-focused media event called Voice of the Body, Samsung announced a handful of projects that aimed to help consumers better track and manage their own health through wristworn, multi-parameter vital sign tracking devices. The tech company's (typically hyped up) event kicked off with an introductory video that framed the current transition underway for patient generated health data. The narrator asked:

"What if your body had a voice? A voice that could speak to you in a clear way that you could understand. What if you could ask questions and hear answers from your heart, your lungs, your muscles, your skin? What if this conversation happened every moment of every day? You would learn a lot and you would live a better life. You could adjust your habits according to your body's advice. You could share information with your doctor. Imagine the possibilities. The mysteries unlocked. The insights gained. It would change your life and the life of everyone you care about. What if your body had a voice?"

Samsung and many other personal health device companies are working to create tools that help patient's better explain what is going on with their bodies to themselves and, perhaps in some cases, to their care providers. 

"When I was a medical student back in the late 1980s, patient generated medical data was basically, 'How are you feeling?'" Scripps Health's Director of Digital Medicine Dr. Steven Steinhubl told MobiHealthNews. "And other than the subjective input, that was basically what you got. The only real home monitoring tool at that time was a glucometer, which were developed and first available in the late 1970s. Other than a glucometer for diabetic patients, there really was nothing else other than their subjective input. Gradually, after that, there were home blood pressure monitors became available. They took a while to become more accurate and more accepted... Back then you might ask your hypertensive to stop by a Walgreens and pick up a blood pressure cuff, bring it home, and keep track of it. Of the patients I asked to do that, maybe one-tenth would do it and the next time I saw them, they might have remembered to bring in a list of blood pressures."

Steinhubl says that there really wasn't much patient generated health data in the past, both because the technology wasn't there to gather it and perhaps as a result of that, there wasn't a desire to attempt to collect it.

"You didn't expect patients to come with anything but their medications list and their symptoms," Steinhubl said. "That was basically it."

Dr. Joseph Kvedar, the founder and director of Partners HealthCare's Center for Connected Health told MobiHealthNews that healthcare systems need to figure out how to make better use of patient generated health data.

Dr. Joseph Kvedar Dr. Joseph Kvedar

"If we don’t, we’re leaving so much useful information about our patients on the table," Kvedar said. "And if we’re really serious about providing healthcare, not just sick care, we have to get it right. We really don't have a choice. I say that because for almost two millenia now or more, healthcare has been practiced as if the patient was some sort of physiologic specimen. You sort of do things to the patient, and they respond because they’re physiologic beings. We now know that that’s very antiquated. We have to engage people, and the best way to engage people is to collect information from them when they’re not in front of you. Healthcare is no longer about when you're in the facility, it’s about your life."

One of the oft-discussed episodes in patient generated and controlled health data was the rise of the untethered personal health record (PHR) about eight years ago. Google's since shuttered PHR platform Google Health and Microsoft's still kicking but relatively quiet Microsoft HealthVault. Ernst & Young Principal Dr. Bill Fera told MobiHealthNews that had those two initiatives been more successful, patient generated health data might be further along today than it is.

"I wish the Google and Microsoft forays into personal health records had been more successful," Fera explained. "I think they would have been if we had been able to automate some of the information that was being input. Then I think we would have had much greater patient adoption, but the manual entry made them unattractive in my opinion. But, if we had untethered PHRs that would allow the patient to be the steward of moving their data from office to office, physician to physician , or physician to hospital, I think that would be a great opportunity. I think [untethered PHRs like those] will come back around as we get better at automating data entry."

In managing his own health, PricewaterhouseCoopers' Managing Director of Healthcare Strategy Chris Wasden, said he chomped at the bit for connected personal health devices years ago.

"When you look at the journey of getting to wide acceptance of [patient generated health data], it has to start with having the technology and tools to enable it first," Wasden told MobiHealthNews. "I remember five years ago I wanted a wireless, electronic scale in my home. It didn’t exist. And I had an old analog scale with a dial and a number. I wanted something that would digitally collect information and store it. Well, now there’s a number of those. I wanted a blood pressure cuff that would be able to send my information to the cloud and store it there as well... In the last five years, we’ve gone from a world in which you couldn’t do any of this stuff on your own… to a world in which we have abundant choices to be able to do this. Step number one is get the tech in place."

Wasden said while it's getting better, the early digital health devices were not easy to use. Only early adopters who were willing to endure inconsistent results used them, but we've come a long way in five years, he said. 

Right now: The present state of patient-generated health data in healthcare

What's the current state of patient-reported data in the healthcare system? It's hard to say. There is already a device to help patients gather almost every imaginable vital sign or healthcare data point, and these devices can already connect to a smartphone and store the information in the cloud. Early adopter patients and doctors are already tracking many health metrics and in some cases even inputting that tracking into the EHR. But, as the oft-quoted sci-fi author William Gibson famously said, "the future is here, it's just not evenly distributed."

Many consumers are tracking health

Widespread adoption physician acceptance of patient-generated data is not the norm in 2014. Chris Wasden, Managing Director in the US Healthcare Strategy and Innovation Practice at PricewaterhouseCoopers, thinks that's a problem.

"If I’m only going to the doctor a couple of times a year, then there’s only going to be a couple of data points about me within the clinical environment," he told MobiHealthNews. "Yet I step on my electronic weight scale almost every day. I have my activity measures that I'm measuring with my [tracker device] and my other apps every day. I’ve got all the sleep data I gather from a device on my bed. There are these blood pressure cuffs I use periodically. I’ve got an app where I log every meal. So you start to see that the amount of data I collect about myself is exponentially greater than the data my clinician will ever collect on me. And that’s true for everybody. Even if you’re going to your doctor every month, you still have the ability to gather 30 times the information yourself at home."

Nonetheless, all of the devices Wasden uses are innovations of the past five years. And he thinks the use of those devices is on the rise. Wasden's back-of-the-envelope estimate is that about 20 percent of the population will probably never adopt mobile health tools.

"Then there’s 20 percent of people who will adopt it on their own without any prompting, and half of them are already using them," he said. "They need to adopt and experiment and do more of that. That leaves the big 60 percent group in the middle. And that’s where we’re really going to see the unfolding of the larger adoption that will have a bigger impact. We’re going to see that number increase and grow over a 10 year time frame."

Dr. Steven Steinhubl, Director of Digital Medicine at Scripps Health in La Jolla, California, agrees that the adoption of home, consumer-driven tracking is small but growing.

"You have more capability, not dramatically more right now, but you do have more," he said. "Activity trackers, devices that quantify your sleep, so you have that information, and that's a small percentage of the population that has them -- but growing. That population -- do they have an outlet for it? So, of that group, there is an even smaller percentage who has a provider who is open to receiving that information and then acting upon that information."

Right now, challenge is getting data into healthcare system

What's holding them back is interoperability with EHRs: currently, Wasden points out, a person's information exists on his or her apps, and there's not even a great way to aggregate all that information on a single smartphone.

"We’ve still got a lot of pain points around interoperability, the ability to aggregate information so we don’t have half a dozen silos of data, which is part of what we struggle with a bit," he said. "Increasingly we have apps that will incorporate my [activity] data, [weight scale] data, my [running] data and things like that so that in one place I can see it all. We’re still not there and, there’s a lot of us that would like to have all my data sent to one repository in the cloud. And then I can figure out which apps have access. Because I may want to change my apps over time and if all my data’s in one place, it’s not that big of a deal to change my apps."

There are several things keeping device data out of the clinical sphere right now. Roeen Roashan, an analyst at IHS Technology, shared his views with MobiHealthNews.

"There are a lot of people out there today with activity monitors and consumer medical devices, but that data is not currently being used in the clinical context," he told MobiHealthNews. "Why is that? That's because there is no structured reporting of that data. Physicians don't want to take a lot of data in without using it properly. They see a lot of it as an extra workload. As of right now there is no solution to that." 

He says movement on this gap is coming not from patients and providers but from health IT companies, especially EHR vendors, that see the value in integration.

"The vendors that I have interviewed for our reports tell me they are working on patient portals," he said. "Right now those patient portals are only being used to schedule appointments and to only put in some very shallow, superficial data -- like name and address. In a year or two we will have a system in place where patients can monitor their vital signs at home and transmit that data to that patient portal through their phone and that will go straight to their electronic medical record. That kind of mechanism is very valuable. That is coming from the health IT companies."

Even when there's a seamless way to move patient-generated data into the electronic medical record, it's important that doctors can work with it -- that the data is somehow filtered or curated so it makes less work for a physician, rather than more. That's the biggest hurdle right now, according to Dr. Bill Fera, a Principal with EY's medical advisory practice.

"If we look at wearables, when you have a stream of data that’s coming in, in terms of blood sugar or blood pressure, how do I set up a system by which I’m getting alerts only when those are out of whack?" he said. "If any of them are normal, or normal for that patient, do I need to know every day what those four blood sugars are for all 20 patients on my panel? Or do I just need to be notified whenever a dangerous trend starts to surface, similar to how we monitor weight with a patient with congested heart failure? And I think we’re starting to put those technologies into place, I think that they're just waiting to go mainstream."

One provider organization where this work is already being done is Partners Healthcare in Boston, which runs the Center for Connected Health. Dr. Joseph Kvedar, founder and director of the center, said that they have several patient-generated data initiatives underway. One, using weight scales and blood pressure cuffs with heart patients, succeeded in reducing cardiac readmissions by 50 percent. Another, working with diabetes and hypertension, is close to scaling.

Kvedar said that what holds them back from using consumer devices connected to smartphones is often the age and technology level of participants. He used the example of blood pressure cuffs. Since some of the newer smartphone-enabled blood pressure cuffs use Low Energy Bluetooth, only newer smartphone models work with them. Using these newer blood pressure cuffs would then limit the number of patients that would be able to participate in the programs.

"While 65 percent of our patients are now using [some kind of] smartphone, if you insist that they have later generations that really shrinks the sample size. So we haven't gotten to that program where we say 'You have high blood pressure, we’re going to give you a coupon, here’s a list of four or five, you choose one you like, and we expect you to start updating next week. If you have technical trouble we’ll help you sort it out, but you’re going to buy the technology, you’re going to own the data charger, you’re going to use your mobile device as the hub and that will tremendously lower our cost of getting you on the program.' Again, we haven’t gotten there yet but we see that as a relatively near-term future for how we do this."

Patient generated data is not just sensors

It's important to remember today that patient generated data includes not just sensor data, but manually-entered patient data too. This is still something that's collected on a clipboard before an office visit, over and over, and whenever a patient switches providers. That is starting to move to the patient portal or to iPads instead of paper clipboards in some practices, again at a slow pace.

Kvedar said Partners is working on a number of initiatives with self-reported data, as well -- having patients respond to text or email surveys about their health, or even take their blood pressure or glucose with a non-connected device and report it by text or email. These systems have the advantage of having a much lower barrier to entry, technologically, but the disadvantage of dubious accuracy. Two psychological biases affect the answers -- people subconsciously want to make themselves look better, and they act differently knowing they're being observed -- what's called the Hawthorne effect or sentinel effect. Both of these are alleviated when sensors are added to the mix.

"We see a lot of potential because the strength of sensor data is the mirror of the weakness of self-reported data," Kvedar said. "That is, it’s objective. If I am taking my weight every day and it’s uploading automatically and displayed for my provider or my provider’s agent, then we can use that sentinel effect to our advantage. I’m going to stick to my diet and stick to my regimen because I want my numbers to look good. The same psychology of social desirability applies and I know that I can’t lie to the sensors, and by the way if I don’t upload, my nurse or my doctor’s going to be asking 'Hey I haven’t seen a weight from you in four or five days, step on the scale.' So there’s nowhere to hide. You have to be more adherent."

Doctors' attitudes are another barrier

Fera and Roashan both think the other barrier, besides the technological ones, is in the attitudes of physicians who fear relinquishing control of part of the care process to patients.

"That's sort of a mixed experience. Based on physicians that I've spoken with, the ones who have good experiences with it say they like the fact that they get more data about the patient, but in the other extreme we also have the ones who say, well, I'm not going to base my diagnosis or my profession on data being generated in an environment that I can't control," Roashan said. "...Are patients using the device properly? Is the device itself calibrated correctly?"

He thinks on the one hand, new generations of physicians will be more likely to trust patient data. Fera also thinks that steps can be taken to make the whole process more palatable to physicians.

"I think figuring out the workflow by which it becomes easy for that information [to be verified] as it’s moved into the chart as part of the permanent record by physicians [is important]," Fera added. "As we figure that out, and we get better adoption of those types of technology, I think the cultural barriers will start to break down. And the sooner we can do that the better because I think the flip side of this is a tremendous ability to engage patients. If they’re actively involved with putting data into the chart and they’re responsible for it — whether it’s their medication list, whether it’s review sysmptoms, whether it’s allergies -- being an active participant in care engenders engagement. So the faster we can start to break down these barriers, the better." 

The near and distant future of patient-generated health data

New kinds of care from patient generated data

The analysts and providers MobiHealthNews spoke to had different views on the future of patient-generated health data, but all agreed that it was coming once the data hurdles were cleared.

Steinhubl, of Scripps Health, thinks that patient generated health data won't just improve the kind of care we already do now -- it will enable totally new forms of care.

"I think in the very near future, we will all be able to easily, seamlessly, non-obtrusively, be able to continuously monitor our blood pressure, pulse, activity, stress level, quality of sleep, oxygen levels, carbon dioxide levels, and probably our glucose level," he said. "This will help us to much better understand ourselves and how all of these parameters interact and find a lot of the uniqueness about ourselves. Right now in medicine a lot of our answers are population-based, and often based on the population that the particular provider taking care of that population is familiar with. Instead, we will get to therapies -- and with therapies too often we think medicines -- but these will include dietary changes, sleep changes, better behaviors for stress reduction. Through these kinds of therapies we will be able to much better find what works for ourselves for any problems we see: elevated blood pressure, poor sleep quality, elevated stress -- all of those we will be able to track and then be able to act upon."

Indeed, by tracking health information at home, patients can see a whole side of themselves, medically, that was never available to them before.

"We know what a very normal blood pressure is when you sit in your doctor's office for five minutes and relax, but I have no idea, no one has any idea, what a normal blood pressure should be when you are stressed, running late, stuck in traffic, and angry," Steinhubl said. "All of these tools that have the ability to measure important biometrics at home are going to help us to improve all of healthcare. They are going to change diagnostic capability and take diseases that -- we are going to be able to identify conditions that right now we don't know exist. It's going to be more personalized and more refined and really dramatically end up improving our ability to take care of the individual patient."

Kvedar echoed that sentiment. "We have to engage people, and the best way to engage people is to collect information from them when they’re not in front of you," he said. "Healthcare is no longer about when you're in the facility, it’s about your life. So, we really have to get this right, it’s very very important."

The future: consumer or physician-led?

Wasden, from PwC, who says patients can collect dramatically more data on themselves than their doctor can gather on them, also believes patients, not doctors, will lead the charge in working with that data.

"I don’t think any of this is physician-led," he said. "I think it’s consumer-led. I think the clinician becomes one of the last players to be engaged in that process. Once the consumers start pulling all this together, the consumers have more data now about themselves than the clinicians will ever have, then the consumers start to engage the clinicians. They’re going to start to compare their information with the clinical information and determine what’s the value of the insight. You'll have some early adopter physicians who start to prescribe apps, prescribe devices, encourage their patients to be involved in this type of activity. Those types of physicians are the exception and not the rule."

Mostly, he thinks much of healthcare will be replaced by positive, consumer-led feedback loops, with the doctor's role relegated to crisis cases.

"The only way we’re going to fix data in healthcare is by creating a new feedback loop that engages the consumer in their own health and leads to changes in behavior," he said. "That’s the only reason we should even be involved in digital health. Collecting this data and information is critical to that feedback loop. I think that this is the only way for us to really address the problems in our system. This is our future, in that we’ve got to empower, engage the consumer to take charge of their health and not have this be a paternalistic health system where the doctor calls the shots, makes all the rules and patients are very passive and very unengaged." 

The changing role of the physician

Whether it's that extreme or not, the role of the doctor will certainly change as patient-generated data becomes more prevalent. Steinhubl thinks this is a source of some of the initial physician pushback, but he thinks ultimately they will come to see it as an improvement.

"I think a lot of the knowledge and self-monitoring technologies that we have will allow a change in the healthcare system where very simple things like blood pressure and glucose, which honestly don't require a physician, can be handled in other ways and often in automated ways. Physicians can then take on a role of -- and this is really what their training was designed for -- of being more educators and diagnosticians."

Fera thinks a lot of routine healthcare visits will be eliminated entirely once most people take it upon themselves to keep tabs on the basic markers of their health. Once "visits" are more frequently done by email, app, or video consultation, doctor's records of visits will automatically become more complete.

"I think most care in 10 years will be done virtually," he said. "Information will be sent to physician offices automatically through wireless devices or questions will be coming in via email. More information will be provided or algorithms that guide patients down care protocols or appropriate questions will be used to generate the history of present illness. And I think we’ll see most of that data come to the physicians as part of a virtual encounter and then the physician moving that data into the permanent record. And I think there’s some irony there, because in that workflow you would have perfectly documented what the patient told you, so there’s no chance of missing information, there’s not something the patient said that you didn’t put in the record. You move that to the record and you have a perfect accounting of what was said."

Roashan, from IHS, is more skeptical about how much tracking will ultimately be brought into the home.

"I definitely see a limit to how far this can be pushed," he said. "Again, the intention of all this isn't to substitute for physicians or the hospitals. It's complementary. Lab results and all that -- I do think we are going to see more and more smaller clinics, including employer ones or kiosks like Healthspot that offer telemedicine, lab results -- we might see more of that. I think there is a limit to what we can do at home and I think there should be."

That limit won't necessarily be technological, but it will be a limit to what's useful and practical for people to do at home. He thinks lab work -- which a few companies like Cue and InsideTracker are already seeking to bring to the home -- will not be widespread in home adoption.

"The value of remote patient monitoring is in the vital sign," he said. "It's something you can do on your own and it's more efficient for the healthcare sector as well. It is not necessarily more efficient if you do -- let's choose an extreme -- if you do your own bloodwork. That will never happen because of -- on a compliance level -- I don't see that ever being efficient because it doesn't make anything easier. It has no value in terms of the quality of care. I think the reason why we will see home health monitoring become much more used and applied is that there is a value to that data, that patient generated data. That's something you can do at home -- at all times."

Device convergence and passive monitoring

Most analysts agreed that to get people to really adopt the self-monitoring technology in the future, it needs to be really easy to use. Part of that will mean a continued focus on building easy-to-use user interfaces. Part of it might be putting all of the existing vitals-gathering technology into a single device. This is essentially the challenge that's been issued by the $10 million Qualcomm Tricorder X Prize, which chooses finalists this August. And one frontrunner competitor, Scanadu, demonstrates what's possible here.

"There are certain devices today that can do more than one vital sign," Roashan said. "A device like Scanadu Scout -- which isn't out yet -- it's not FDA cleared. A device like that will be the type of device that I believe a lot of people will have in their homes. It measures I believe six or eight vital signs -- all of them important ones. That's a lot of easier than having seven different devices measuring seven different things."

Kvedar, from Partners, looks to other industries like banking and advertising to see where healthcare will be in the future. And he thinks patient-generated data will consist not just in self-reported data and sensor data, but in all kinds of passively-collected data your smartphone can collect right now.

"So I fully expect in 10 years your healthcare experience to be very personalized based on you allowing us to capture a lot of data from your device, whether it be like Ginger.io, the idea that we can capture your mood based on the number of texts you send and outgoing communications," he said. "We’ll know your GPS tracking, all your mobile purchasing data and we can kind of figure out if you’re eating healthy. We’ll know if you’re drinking too much caffeine or too much alcohol. We’ll know a lot of things about you that we can start to serve up really compelling content to you that should be engaging enough — games, rewards, incentives, what have you in the background — that you will want to stay healthy."

"I think that’s where we’re going and it will be hard, if not impossible, to do without patient-generated data."