Too much data, big and small.
Every morning when I log onto our electronic health record, one group of messages are from patients with their self-recorded health information data that they’ve sent to me for review.
Here’s my home blood pressure over the past 24 hours. Here are my fingerstick glucose readings over the past 24 hours. Here’s my weight over the past 24 hours. Here is my depression score over the past 24 hours.
A lot of data — oftentimes a lot of data that I don’t quite know how to react to, what actions I’m supposed to take. Now don’t get me wrong; it’s not the patients’ fault. I asked them to send me all this.
Finding the sweet spot
In the order section of our electronic health record, during an office visit or other times of order entry, I can put in a request for the patient to get a template to send me a lot of biometric health information about themselves directly through the patient portal.
Many patients ignore these requests.
Those motivated patients who take this to heart will sometimes send me too much information, sometimes not enough, and we rarely hit the sweet spot — just the right amount.
I have a patient with Type 2 diabetes who has been stable on the same medication regimen for many years, who refuses to stop checking his glucose level five times a day, despite the fact that all his numbers remain in an incredibly well-controlled range, never any hypoglycemia, no hyperglycemia, and no readings that would result in either acute or chronic changes.
I have told him he is more likely to make himself anemic by pricking his finger for all those tiny drops of blood than uncovering anything dramatic that will improve his life.
The occasional reward
It’s unusual, although rewarding, when these numbers do suggest a trend, highlight something out of whack, something that catches my eye and makes me want to intervene.
Take the recent case of a patient who had been sending me his (normal) blood pressure readings for quite some time, when suddenly I noticed a spike in his numbers.
“That’s curious,” I thought. I sent him a message through the system, anything going on? Any changes, run out of medications, or is there something else going on that you want to tell me about?
Surprisingly, what I got back was eye-opening. The answer: over the holidays he had indulged in a lot of salty smoked salmon, followed by a weekend of watching sports with its associated nachos, chili, alcohol, and other less then advisable food choices.
He took my gentle nudging to heart, and promised he’d be “good starting this week.”
At our departmental Grand Rounds this week, we heard from the head of our new technology innovations center about innovative research being done by high-tech computer and engineering gurus trying to figure out ways to harness the data our patients have within them, both large and small, to improve their health.
They envision a time when every piece of information will be collected, sorted, and analyzed, and that worrisome trends will be highlighted through a series of computer algorithms, hopefully giving us an opportunity to intervene and change clinical outcomes.
The speaker gave examples of decreased mobility, fewer steps taken during the day, altered sleep, changes in dietary patterns, unsteady gait measured by accelerometers, as well as a multitude of other variables that might indicate a changing state of health, either physical or psychological.
Amazing opportunities to harness evolving technologies in clever ways to capture and interpret data that might otherwise escape our detection.
Many of the clinicians in the room, while discussing this afterwards, noted that we are now suddenly going to become responsible for an exponentially larger amount of data that our patients are going to generate, as their phones and smartwatches and implanted devices and responses to queries and data-mining from their computer and shopping history and even their internet searches, will suddenly become fair game that we are responsible for, to somehow say, “Hey, wait a minute, Grandpa’s getting sick; let me leap in here and do something about it.”
Now we are going to be expected to notice that instead of going out for his daily walk, Grandpa is suddenly staying home and watching movies on Netflix. Instead of ordering healthy fresh foods from a grocery delivery service he’s now ordering salty takeout, therefore putting him at risk for yet another congestive heart failure exacerbation.
Or he’s not sleeping enough, or sleeping too much, or not taking his medicines, or taking too many medicines.
Unfortunately, the possibilities are endless. Looking at too much data creates too many possibilities. It could always be something. Sometimes the explanation for why these things happen are totally benign, and sometimes they’re insidious markers of some impending horrible event.
Putting all of this on the providers, expecting us to be 100 percent vigilant and 100 percent insightful, is unrealistic.
This will only work if an entire team is engaged in looking at this data, and putting it in the correct context for Grandpa, utilizing input from his family, his friends, his neighbors, his home care team, and all of the providers working to take care of him.
But until we get to that point where we have smart enough systems and artificial intelligence that can extract the correct conclusion from all this data, it feels like it’s going to be too much, a wave that hits us that will only put us further behind in our work of trying to take care of all of our patients.
Not that I don’t think we should pursue this. I think this holds great potential, and we may be able to develop some really cool tools that’ll help us detect worsening depression, avoidable admissions, heart attacks in the waiting, cancer whose diagnosis is as yet unmade. But until the technology is really smart enough and refined enough, I think we need to swim cautiously into these waters.
I don’t think most of our patients want us snooping into what their online shopping and internet search habits are, I don’t think they want Big Brother constantly looking over their shoulder, even in the name of their health.
As we move forward, I think we may be able to engage our patients to have just the right amount of oversight, relinquish just enough peeking behind the curtain, just the right amount of us being nosy — but probably for each patient we will have to figure out together where that level is, how much privacy they want to give up, how much they want us minding the store.
And how much they want us minding our own business.
If the 21st century completely patient-centered medical home turns them into enormous generators of data, both small that originates from them and informs their clinical state, to larger data whereby they contribute to population trends, we need to be mindful and thoughtful of how to figure out how to have this truly serve each patient in their best interest, how to have it best serve each provider, and only build systems that really help us take the best care possible of our patients.
So let’s keep crunching the numbers until we get it right.
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 MedPage Today’s Building the Patient-Centered Medical Home.
Image credit: Shutterstock.com