Skip to content

Patient Generated Data: Insights or Nuisance?

February 11, 2016

Attitudes1For about 10 years now, we’ve been collecting data from patients, mostly from sensors in their environment (activity trackers, blood pressure cuffs, weight scales, etc.), trying to break free from less-than-reliable self reported data.

We’ve shown that, properly used, patient-generated data is:

  • Central to improved quality of care for chronically ill patients (i.e., patients with heart failure in a telemonitoring program had significant improvement in mortality compared to controls).
  • Useful in engaging patients around lifestyle choices to improve the care of both diabetes and hypertension.
  • Useful in engaging patients in interventions that encourage them to improve their own care.

Despite these studies and a bevy of evidence from other investigators in the field, I still get routinely reminded, when I am out presenting on connected health, that “physicians don’t want this data.”  This is a resounding theme that ranks right up there with concerns about liability and reimbursement.

On the one hand this response is predictable.  Physicians’ time is squeezed more than ever these days. The thought of having to pour over reams of normal values seems daunting.  I take this objection as evidence that we aren’t clear in communicating our vision — or the true benefits of having real-time, real-world patient data available.  In so many other industries, software has been designed to troll through complex data sets and cull out the meaningful relationships or events, and then presenting them succinctly as actionable information or valuable insight rather than just raw data.  We can do this for patient generated data too.

Then there is the objection that it is complex and costly to collect such data.  Things are getting better on this front.  We’ve shown that for diseases such as congestive heart failure, remote monitoring programs that use patient generated data lead to a significant drop in hospital readmissions and, more importantly, a decrease in total medical expenses. These major cost savings quickly add up, while improving the quality of care delivered. It is hard to object to that combination.

Validic, a forward-thinking company, is one example of an organization providing software that enables easy collection and normalization of all sorts of patient-generated data. The company just came up with an exciting innovation that allows a patient to take a picture of any data-collecting device (imagine the screen on a home blood pressure monitoring device) with her mobile phone camera and have that data flow into the electronic record.  This is analogous to taking a picture of your check and having it deposited in your checking account.  This will lower costs significantly.




The objection that irks me the most, though, is the one that basically admits to striking out before even getting up to bat: “I don’t know what I’d do with all of those readings anyway.”  This approach disappoints me in many ways. First as physicians, we’re trained in the scientific method.  We’re supposed to be curious, to think about novel ways of providing care and above all, ways to improve care.  Yet, some health care providers feel completely at ease that a disease as significant as hypertension can be adequately assessed by taking a few readings in the doctor’s office once or twice a year.  Maybe that made sense when the sphygmomanometer (aka blood pressure cuff) was first invented. But now that these devices are ubiquitous and can be purchased in every drug store, why would we not want to take advantage of the extra data points from home readings?  We made the leap for glucometers decades ago.




Chapter 5 in my new book, The Internet of Healthy Things, outlines what I call ‘the new white coat anxiety’ — the myriad concerns and changes taking place in health care today that is causing providers angst. Chief among them is apprehension to change. Change is coming from all sides:  shifting payment models, changing from face-to-face interactions to virtual care, a lack of awareness about personal connected health trackers and real anxiety over what to do with all of this data.

My friend Dr. Steve Ommen, Associate Dean at the Center for Connected Care at the Mayo Clinic in Rochester, MN, told me when I interviewed him for my book: “Providers will come to see that incorporating connected care into clinical practice will be of benefit to both their practice and their patients, but it’s not the way we’ve all been trained to practice medicine for 150 years.”

Many don’t realize that when Laennec introduced the stethoscope it was ridiculed.  In fact, it took 20 years for it to be accepted as a standard tool for clinical practice.  The analogy may be apt, but we don’t have 20 years to wait for health care providers to embrace patient generated data.  As we embrace the world of value-based provider compensation, it is imperative that we look at collecting information from and about our patients differently.




How do you think we will be able to help providers embrace patient generated data and accelerate the adoption of connected health?

13 Comments leave one →
  1. Nirav Desai permalink
    February 11, 2016 10:33 am

    Hi Joe,
    Great points. I think it will help adoption as more stories & examples come to light of how physicians are managing to incorporate all the new data into their workflow. This article shares examples from UPMC and Duke.

    The more we can show the exact examples of how reams of data were made meaningful and easier to process within a clinician’s workflow, the easier it will be for the skeptics to visualize. How about a day-in-the-life video?

  2. February 15, 2016 12:55 pm

    good thoughts and appreciate the reference.

  3. Jeff Phillips, MD permalink
    February 15, 2016 9:36 pm

    Hi Joe. I agree that there is a significant barrier to physicians looking to incorporate this data in their workflow. And that’s too bad. Outcomes can be improved if people open their minds to new ways of doing things. Of course, the devil’s in the details- it needs to be in the EHR, presented unobtrusively, and at the right intervals. The last thing we need is a meaningless data point. BUT, if done correctly, patient generated data could keep people out of hospitals and healthier at home. Now if only our system would pay us for this…We’re still about 90% fee for service with no reimbursement for phone calls, emails, secure messaging with patients, etc. Once that changes, you’ll see heads on a swivel.

  4. February 15, 2016 9:46 pm

    moving in the right direction, i think

  5. February 25, 2016 9:45 am

    Good post and a very interesting discussion.

    I think a good part of the anxiety physicians have about receiving a stream of biometric values is a lack of context. I am an emergency medicine physician. I am most accurate when I have a more complete picture of the patient. Vital signs only represent part of the story. Heck, I have admitted more patients with normal vitals than abnormal ones. They are only part of the story…

    PCPs are frustrated when their staff leaves them a post-it note that says: “Mary doesn’t feel good. BP is 178/90.” What are we supposed to do with this? Ultimately, we call the patient, perform a history, collect data, and formulate a plan. The missing information here is the medical history and we are appropriately TRAINED TO REQUIRE this.

    Don’t get me wrong, I love patient generated data. What we need is something that goes beyond a patient sending me a stream of vital signs. Clearly disease processes such as diabetes and HTN can be reduced to numerical values, but bad stuff happens to these two cohorts and to really understand what is going on you need more than numbers. Maybe IT needs to focus on how we might create intelligent systems that guide patients to providing more complete information (relevant HPI) about themselves when they are sick. If that information also contains biometric data, it will be even more actionable.

    • February 25, 2016 11:06 am

      well stated. the vision is indeed to provide that complete picture for proper decision making. it will be done virtually, eventually.

  6. L. Tan, MD, MS permalink
    February 25, 2016 11:27 am

    I agree that the potential for data to improve care is enormous. However, as physicians we should be able to determine pre-test probabilities of test outcomes, and also know what these results mean to the patient clinically. For years we have been inappropriately doing PSA testing for screening even though we have always known it to be statistically and clinically poor for this purpose. Your example of the stethescope is another perpetual medical myth that ausculation is a reliable diagnostic tool. The point is that everything we as physicians do is based on (limited) statistics and probabilities. The “limited” part means that “art” has a large place in what we do. I am happy to incorporate your data but show me how to use it appropriately based on good and validated evidence. Thank you.

  7. Jim Reid PA, PMP permalink
    February 25, 2016 12:58 pm

    For those that haven’t read it, Joe’s book is full of insights on this and related subjects. There is a whole chapter dedicated to making data actionable on the “big data” scale, but first we have to get past what I consider to be the “little data” problem described above. What to do with all the data points on any individual patient? I think physicians probably project a more intimidating picture than is really necessary. No one wants to see every glucose reading, BP, or weight on every one of their patients every day, any more than they want to see all of the engine temperature, oil or tire pressure measurements coming from sensors in their car every minute as they drive down the road. But if the oil or tire pressure falls or the temperature climbs beyond preset thresholds they sure want to see the red light illuminate on their dashboard, so corrective action can be taken. Most people don’t realize that there is an on-board computer in their car monitoring hundreds of sensor metrics every second, filtering them, trending them, and then only sounding an alarm when something is out of spec. Or that there are alarm tiers guiding the urgency of an intervention – ranging from take your vehicle to a shop soon (check engine light) all the way to stop the car immediately (oil pressure light). The same algorithmic logic applied to managing auto sensor outputs can be used to filter, trend, and alarm on data coming from the IoHT and tiered responses by care managers, CDEs, nurses and physicians can be programmed. We deal with “dashboards” in many aspects of our business and personal lives. Wouldn’t the analogy of a IoHT data dashboard for our patients be helpful in alleviating some of the new white coat anxiety?

  8. Catherine M. permalink
    February 25, 2016 1:37 pm

    My recent experience as a patient, being asked to complete an online health assessment prior to my appointment, leaves me with little motivation to provide my doctor with self-reported data for three reasons: 1) the assessment was poorly designed; 2) when I asked my physician if she had received my responses, she had no idea what I was talking about; and 3) once we clarified the confusion and she found my assessment responses, she explained that the problem was that the data was not connected to my EHR.

    The underlying problems however, were bigger than the technical difficulties we identified together. First, the assessment was a modified version of the HAQ, a standard assessment focusing on functional categories that patients rate on a scale of 0-3. This particular version however, presented me with questions so poorly modified that some of the questions were unanswerable. Furthermore, someone decided to replace the 4-point scale with a 10-point scale that did not correspond with the questions. I answered what I could – especially since I had just experienced a severe flare-up and I thought my responses may be worthy of discussion with my doctor.

    But the second problem – my doctor’s unawareness of the request to complete the assessment at all was more troubling. If she didn’t know her patients were being asked to complete the questionnaire, who was collecting the data? More importantly, why was I being asked to take the time to complete a questionnaire that my doctor clearly didn’t need me to complete? She seemed baffled herself, and mentioned that maybe someone else in the department had put it into the system; but the point is, she did not know! And she is Director of the Department.

    Now, thankfully, my chronic illness is not life-threatening and I am not talking about sensor-driven data collection that could trigger alarms. I am willing to believe there is value in collecting patient-generated data for more serious conditions. But not so fast. This is not the first experience I have had when I’ve been asked to keep a log, monitor my symptoms, or answer questionnaires, all of which I have complied to do. It is also not the first time I learned that the data had little, if any, impact on my care. Imagine a future scenario in which I am asked to provide my doctor with data to monitor my health for a more serious condition? My belief system is rooted in the experiences of what my doctors have not done with data they have collected over the years – reviewed it, shown interest, or made any recommendations based on the interpretation of it. I’m not sure I will be willing to be bothered.

    Providers need to identify what they want to learn about their patients that they cannot learn from diagnostic tests and 10-minute visits. What information will help them better understand their patients lives and how to best treat them? What is the best way to collect this data? They need to show their patients what they plan to do with the data, share what they hope to learn, and explain why they care about the data. I believe most patients, like myself, want to contribute to the greater body of medical knowledge if they can. When doctors and patients are working together to collect meaningful data that moves toward a greater understanding, and purpose perhaps, for all patients, it’s more likely that both sides will become active participants in adopting the behaviors of connected health.

    • February 25, 2016 9:37 pm

      thank you for such a well-written, thoughtful expose. so many nuggets of wisdom in there.

  9. February 25, 2016 2:10 pm

    I’m a manager for a baseball team. Picking up on your theme about striking out before getting up to bat, may I remind you that I don’t want to look at a sea of data in order to figure out where my second baseman should stand with Big Papi at the plate. Oh…what…you mean you have actionable information down to the square inch on where to position all of my fielders, whether there is one strike or two? Then count me in.

    Your point, Joe, about OUR collective lack of framing the opportunity of PRO data is right on. Doctors, like old school baseball managers are genetically resistant to being told how to practice medicine or coach baseball. But the managers have let the data hounds into baseball and the doctors are still fighting it.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: