The Promises and Perils of EHR Research
Electronic health records — or EHRs — are THE record-keeping system for most American hospitals. And they’re under fire. Congress is holding hearings on how they’re so hard to use they can even cause fatal medical errors. But at one hospital in New York City, researchers are trying to work around the flaws of EHRs. Camille Petersen has more on the promises — and limits — of EHR research.
PETERSEN 1: Noemie Elhadad is a computer scientist at Columbia University Medical Center.
At her desk, she opens a sample electronic health record for a patient. And searches for doctors’ notes. On the screen, there’s a long list of dates without any information about why a patient came in or what they were treated for.
We see, you know, a list that they keep saying operative note, operative note. And unless you click on each note and read you wouldn’t know what the operation is about or anything like that. (0:11)
PETERSEN 2: Elhadad focuses on biomedical informatics. That’s the intersection of tech and medicine.
And she says this is typical for EHRs. To find something like a patient’s treatment history, a doctor has to click on dozens of links. And go to lots of different screens.
Physicians and clinicians as a whole are spending sometimes more time dealing with the software than actually interacting with the patient. (0:08)
PETERSEN 3: That’s why Elhadad built software that organizes all the information buried within EHRs into a single patient narrative. Elhadad shows me how her program works.
We have a test patient called Carmen San Diego. And our test patient here we have about five years of data…(0:08)
PETERSEN 4: The top of the screen is an interactive timeline of doctors’ visits and notes. Below that, the software generates a word cloud of the patient’s biggest medical issues. So if a doctor wants to look at her history of leukemia, they just click on the word leukemia.
We can actually access the notes that are mentioning this problem. And we can navigate easily to these particular notes. (0:07)
PETERSEN 5: Thousands of doctors and nurses at New York Presbyterian Hospital are now using Elhadad’s software. She says she gets great feedback on it.
Ross Koppel is a sociologist at the University of Pennsylvania and SUNY Buffalo. He’s researched EHRs for nearly two decades. And he says this kind of EHR innovation is hard to achieve at most hospitals. That’s partly because just a handful of companies controls the EHR industry.
Those companies have tremendous amount of power and there’s not a hell of a lot of alternatives. (0:06)
PETERSEN 6: Or competition. Which means there isn’t much incentive for companies to innovate. Also, EHR systems cost a lot of money.
If I buy a toaster and my wife hates it I’ll go blow another 30 bucks and buy another toaster. But if you spend a billion or 2 billion, it’s not like you can say golly there’s a better system out there. You are married to the mob. (0:14)
PETERSEN 7: So hospitals are pretty much stuck with the EHR systems they have. But researchers, like Noemie Elhadad, are creating new capabilities for EHRs.
Across the hall from Elhadad, Nicholas Tatonetti uses data from thousands of electronic health records to study medication effects.
He trained an algorithm to recognize drugs that cause bad outcomes like an irregular heartbeat.
And when we show it data for a combination of drugs it can guess…ooo this one has patterns that look like a drug that causes arrhythmia. (0:08)
PETERSEN 8: Using the algorithm, Tatonetti discovered that one type of antibiotic combined with one type of heartburn medication causes the dangerous heart condition.
But he says because EHRs are hard to use, doctors sometimes don’t fill in important data. And that’s a problem for researchers.
All of that just means there’s less data for us to work with. And so that affects the type of questions that we can answer. (0:06)
PETERSEN 9: Ross Koppel, the sociologist, agrees with the data limitations of EHRs.
EHRs have an ocean of wonderful data that we should be able to utilize. The impediments are terrible, because they stop us from jumpstarting healthcare dramatically. (0:14)
PETERSEN 10: Koppel says fixing EHRs in hospitals across the country would require big changes to the EHR industry and how it’s regulated.
Still, Noemie Elhadad is pitching her software to hospitals beyond New York Presbyterian. She’s trying to sell other New York City hospitals on the benefits of revising their EHR systems.
Camille Petersen, Columbia Radio News.