Asymmetrical Information - Megan McArdle

01.21.13

The (So Far) Failed Promise of Electronic Medical Records

Were we wrong to believe in them?

Remember how Obamacare was going to "Bend the cost curve" for health care spending?

That was OMB director Peter Orszag, back when Obamacare was being debated.  There were a number of theories about how it would accomplish this.  There were electronic medical records, which had been passed as part of the 2009 stimulus, which would cut down paperwork and medical errors.  And ACOs, which would finally bring America to the promised land of "bundled payments"--i.e., paying a flat fee to keep patients healthy, rather than per-service, which was supposed to radically change treatment incentives.  Medicare pilot projects were going to open up new, more efficient ways to do things.  If all those failed, the legislation contained an Independent Payment Advisory Board which will recommend a panel of automatic cuts unless health care cost inflation stays below a fairly low target level.  

The jury is still out on IPAB, which won't go into effect for a while.  One by one, however, the others have disappointed.  Medicare pilot projects have so far been a near-complete bust; "After accounting for the fees that Medicare paid to the programs . . . Medicare spending was either unchanged or increased in nearly all of the programs", said the CBO.  ACOs may eventually produce the promised savings, but the top-notch facilities who inspired the program have declined to participate.  And now EMRs are also looking a little wobby . . . indeed, they may actually increase costs, according to the New York Times.  It seems that they make it a whole lot easier to bill for every niggling thing you do.  

Mickey Kaus . . . er, well, takes the mickey out of overcredulous supporters of EMR.  I'm not sure this is entirely fair.  I was, as long-time readers know, a great skeptic of Obamacare.  I still am.  But EMRs were the one claim I found convincing. (Their promises about ACOs, on the other hand, were frustratingly vague: they could tell you a lot about how great the Mayo Clinic was, but didn't seem to have much of a blueprint, even a rudimentary one, for how ACOs were going to work.  Any attempt to flesh out the details was met with the explanation that this was all really too complicated to explain in a limited time period but of course it was all going to be worked out.)  

Besides, even if EMRs are having trouble now, it's still possible that in the future, they'll deliver cost savings, along with improvements in the quality of care; early automobiles did not much revolutionize American life, but they provided the platform for iterative improvements that eventually did.

All that said, Kaus's follow-up is well worth reading; it contains notes from readers who have lived through the transition: 

My wife is an overworked Family Doctor and hates electronic medical records.  It was sold as an enormous time saver but turns out to have slowed her down.  She used to be able to talk to a patient, check off boxes on her records form as she spoke to them and say goodbye.  Now instead of checking a box with a pen, she clicks on  a box on her computerscreen, waits for it to open, marks the appropriate checkoff, and then closes it.  Repeat .  Repeat.  Repeat. Then close the patients file.  Maintaining eye contact is a thing of the past.  However. Her billing has gone way up.  Things she used to do but never add to her billing are now added automatically because the nanny program prompts her.  I think the electronic records enthusiasts thought that they were pushing best practices.  “Was patient advised about smoking” “ Was the patient counseled about weight loss and diabetes/ hypertension etc” My wife always did these things but never added them to her bill-now she gets paid for doing it.  It seems like electronic proponents were wrong on BOTH of their selling points

Another reader says: 

My wife is a staff physician [at] a major East coast hospital.

Her employer was one of the first to sign up for federal money to implement a system which hospital management freely acknowledges is “terrible” but there was so much money on offer that they couldn’t say no.

Probably the biggest problem with electronic records is simply that it requires the physician to input all notes and orders, rather than dictate them.

As a result, as my bride puts it, “they’ve taken the highest paid person in the department and turned him/her into a data entry clerk”.

On average, she and her colleagues spend more time per patient wading through drop-down menus, clicking boxes and filling in required but utterly irrelevant information than they do at the bedside, actually treating the patient.

In short, it’s her experience that they see fewer patients per shift than they did previously, and spend less time with each one, now that they are required to sit down at a computer after seeing each patient and jumping through hoops to place orders instead of, as previously, simply telling the nurse what is needed and then moving on to the next patient.

To be sure, these are second-hand, anonymous reports.  But they actually match broader complaints that I have heard hearing from physicians about EMRs--some even before the stimulus money had all been spent.  

Two years ago, I wrote an article that was quite critical about the state of healthcare IT, one in which I tut-tutted doctors for resisting change.  Should I have listened harder?

Not to every complaint, but there's one that I should have taken more seriously: that EMRs make it harder to interact with patients, and thus may interfere with quality of care. This is what docs were saying, and it's echoed by these Kausfile writers.  I probably should have been less skeptical then, but I certainly am now, because of personal experience.  

A few years ago, I got an iPad, and not long afterwards, I got a case that had a keyboard in it.  I was ecstatically prepared to have this setup revolutionize my reporting. Finally--finally!--I could type in my in-person interviews, the way I already did with phone interviews, rather than laboriously transcribing from handwritten notes.

I took this new setup to a three-day reporting trip on a college campus out west.  (The story I was reporting, on college costs, never ran, thanks to a bizarre chain of coincidences too tedious to detail here.)  For three solid days, I typed up all my interviews.  It was great: my notes were all in one place, and the application I was using also did voice recording, so that you could go back and listen to what the subject had been saying while you typed up your notes.  

But while the physical business of reporting was going off with unparalleled ease, the trip was a little disappointing.  I wasn't getting great quotes, or making the kind of personal connections that I often--though by no means always--enjoy with interview subjects.  I completed my last interview.  I closed my iPad case.  Then I made an off-hand remark to my subject.  He brightened, and responded.  Suddenly I was having exactly the kind of candid, insightful conversation I'd been wanting to have for the last three days.  I almost opened my iPad so that I could get this . . . and then thought the better of it.  I left it closed, and recorded the session on my phone.

Something about typing puts a barrier between you--even when you are sitting face to face with the subject and can clearly see them over a very small screen.  Maybe this is an artifact of age, and the younger generation will not experience typing the same way.  But I've experimented a few times since, and discovered the same thing: typing kills an interview.  You can have rapport, or you can have fast, legible notes, but you cannot have both.  I assume it's even worse for doctors, since most of the doctors I've seen using EMR are actually facing away from the patient while they type.  

Doctors were saying that at the time, and I wish I'd listened harder.  We're not trying to cut costs at the expense of care (if we wanted to do that, we could just decide not to cover expensive treatments, whether or not they're effective).  And even if we were, this is probably not the way we'd choose to do it: too much worry that conditions missed by a physician who doesn't have the patient's trust might end up in the ER.  

We made it worse by the way that we executed the program.  The rush to buy any system, even a bad one, reported by the second letter writer is all-too-plausible because of the way the EMR initiative was structured: it was part of the stimulus, which meant that the priority was getting money out quickly.  Systems were supposed to be in "Meaningful use" by 2011, which didn't give much time for establishing interoperability standards and quality benchmarks . . . or even for shopping around to find a better system.  The single most obvious benefit of an EMR--instant record sharing--never even materialized because no one established a rule--or a standard--for making systems talk to each other.  As a result, they don't.

Small wonder, then, that EMRs have disappointed: we put a barrier between the doctor and the patient, lowered one between the doctor and the billing department, and didn't even insist on the very qualities that were supposed to make EMRs so great.  But let me make a partial defense of the reporters who believed the hype.  We should have listened harder to the doctors who were sayng that this was no panacea, and we definitely should have investigated claims about the state of the technology out there, rather than listening to over-optimistic sources who insisted that it would be great.  

But it was hard to tell what the systems were going to look like, until they actually appeared.  In other words, while the problem of inserting a barrier between patient and doctor was real, it was hard to tell that until you'd actually experienced it (or even if you've actually experienced it . . . and happen to be the sort of wonky reporter who is very used to people taking down information without looking at you.)  And it was not easy at all to see how the records would be implemented--that they would, apparently, be more designed for the health of the hospital budget than of the consumer.  It's disappointing that we were wrong, and we should have been more skeptical.  But it was not entirely predictible.  

The lesson I take away from this is that we should hope for improvement, but we should not bank them long before they materialize.  It's a lesson that Obamacare has now taught us over and over.