When is too much data a bad thing?

“Here you go, doctor.”

My patient with incredibly well-controlled type 2 diabetes hands me his fingerstick log at his regularly scheduled office visit. Despite his multiple medical problems (congestive heart failure, coronary artery disease, chronic renal insufficiency, and gout, among others) his sugars have been incredibly well-controlled over the past several years.

Page after page of scrawled numbers, tiny smears of his blood on the pages, his fingersticks range from 90 to 140, never much higher, never much lower. Day after day, week after week, month after month, most days four times a day, fasting and after each meal, and occasionally some evening readings as well.

His hemoglobin A1c has never drifted much above 7.0, and despite my recommendations, he does not want to check any less. I’ve asked him why he needs to check so much, and for him there’s a certain comfort in knowing. He initially presented to the MICU in HONK with a sugar of 939.

I worry he’s going to get anemic from all this testing (I know it’s only a few tiny drops of blood, but I’m trying to make a point here).

When is too much data a bad thing?

As we move along in the 21st-century, in this wired generation, the landscape of healthcare information and technology is changing. Not only have we seen the transitioning of doctors from paper charts to electronic medical records (EMRs), but it is also enabling us to acquire and deal with infinitely more data about our patients.

In the patient-centered medical home, open lines of communication and efforts to empower our patients and bring them back to the center of care will no doubt lead to more efforts by patients at self-monitoring and sending us the results.

On the simplest side, my patients can send me their home blood pressure readings and fingerstick readings through widgets on my EMR that allow them to manually input data to send to me.

As we move up in complexity, we have meters that will automatically sync to the patient’s EMR when patients check their weight, blood pressure, or glucose readings at home.

Wearable technology is starting to disseminate to collect pulse rate, EKG, EEG, sleep, steps, calories burned, and more; the possibilities are nearly endless.

And the potential for even more and closer monitoring with a greater level of detail exists, as patients start wearing these monitors and sending data to us and expecting us to do something about it. Taken to the extreme we will likely have patients embedding sensors within their bodies, and have tiny nanotechnology sensors floating in their bloodstream continuously monitoring every measurable aspect of their health.

But the question remains, what do we do with all of this data?

A fingerstick today can tell me whether the symptoms my patient is feeling are from high or low sugars, helps me fine-tune a dose of insulin, helps patients learn how their diet directly impacts their sugar, or helps spot issues with compliance. An HbA1c reading will help me make global changes in their regimen, since it gives the snapshot over time. And even intermediate glycosylated protein markers can be useful in fine-tuning changes over a short interim period.

But every data point?

The deadpan comedian Steven Wright once said, “You can’t have everything — where would you put it?” And a friend of mine once said to me, “If you videotape more than half of your life, you won’t have time to watch it.”

As this fascination with collecting more and more medical data from each individual human we care for goes to further extremes, we may be soon receiving a continuous data dump of massive amounts of biophysiological data on our patients’ every move and the fluctuating levels and markers of each and every one of their health conditions. Unless we carefully think about what we’re going to do with this data, it ends up being excessive and might not be very useful for the patient’s health.

Care coordinators and other members of the team may be used to review data and help spot concerning trends or dangerous levels. One can foresee a future with large workstations with monitors showing endless streams of data on our patients and automatic thresholds triggering an alarm. Giving patients timely feedback and other useful interventions and resources could help patients interpret what is going on and better manage their overall health.

But we need to make sure we don’t miss the forest for the trees. And the branches, roots, bark, and leaves.

Fred N. Pelzman is an associate professor of medicine, New York Presbyterian Hospital and associate director, Weill Cornell Internal Medicine Associates, New York City, NY. He blogs at Building the Patient-Centered Medical Home

Comments are moderated before they are published. Please read the comment policy.

  • Close Call

    We shouldn’t be gathering data we don’t know how to evaluate.

    That’s why we don’t order CBC’s on every patient we see in clinic. No good evidence of benefit.

    We’ve let Jawbone and Nike market these wearables as “Health Devices”. Until there’s good studies for when to use, and how to evaluate the data, doctors should politely decline to spend a significant amount of discussing the wearable other than a cheerful “Good for you!”

  • JR

    I read something the other day: When a two year old has no control in their life, they’ll control the one thing they can: their bathroom habits.

    I get a feeling that the finger sticking is giving the patient a feeling of control, so is the psychological benefit to the patient worth more than the negatives?

    I have read that continuous fetal heart monitors were introduced with the assumption they’d improve outcomes, but they’ve increased the c-section rate without improving outcomes. We introduced a slew of data without knowing what “normal” is to compare it to, and assumed the “abnormal” meant “more dangerous/worse outcomes” when it actually doesn’t.

    For the fitness trackers, I think that long term they may have some merit, but I don’t think they are ready for medical use. They may motivate positive behavior in some patients, and in that case I’m all for it. But before we put them into medical practice, and start making medical decisions, we need to have data on whats “normal” and what’s “outside of normal but safe”.

    I’m interested in getting a fitness monitor eventually, but I’m waiting for the “fitness” war to narrow down the choices. I jumped in the personal music player/individual coffee makers markets too soon.

  • SteveCaley

    In the way that words have no meaning without a grammar, data has no meaning without a process to bring it to intelligibility. The pretense is that sufficient data will replace the need to think; that quantity becomes quality. An old Marxist saw – ask any Soviet leader (oops, they’re gone.)

    • querywoman

      He’s still gonna die, and this doctor mentions his other severe health problems, but doesn’t say how bad they are.
      I can’t handle waking up with less than a 110 blood sugar.

  • querywoman

    When I became diabetic, I refused to be a total nut about it.
    Some of the things that the modern world make easier cause overuse and paranoia.
    Someone in church complained once that all the text messages ever sent are stored somewhere.

  • SarahJ89

    For someone with OCD there’s never enough data.

  • http://strangelydiabetic.com Scott Strange

    When you are looking at each data point (i.e. a finger stick) the sheer amount of data can make it seem worthless, too much to manage. I would suggest, however, that you are looking at the data in the wrong manner. What matters in a situation such as this is the trends within the numbers.

    As a diabetic, I use a device called a continuous glucose monitoring system (CGMS).The CGMS provides me with values every 5 minutes during the day. That is a lot of data to look at, true, but what I am most interested in is the lines between the data points. Do I tend to go high at night or drop low in the mornings? These questions about the trends presented in the data can give me the clues I need to properly adjust the amount of insulin I take.

    A single finger stick can show you what is happening right now. And that is all. I can make decisions based on that number, but only decisions about the current situation. Do I need to eat or do I need more insulin or do I know something the finger stick doesn’t and I’m not going to take any action?

    You really can’t make adjustments to overall insulin dosing based on a single number, you can only make them when you have enough data to sopt the trends.

Most Popular