Wednesday, December 16, 2015

Oh, And It Is Also An EHR

You wouldn't -- I hope -- still drive your car while trying to read a paper map.  Hopefully you're not holding up your phone to follow directions on its screen either.  Chances are if you need directions while you are driving, you'll be listening to them via Bluetooth, glancing at an embedded screen on your dashboard, or maybe looking at a heads-up display on your windshield that doesn't even make you take your eyes from the road.  Or maybe you're just riding in a self-driving car.

But when it comes to your doctor examining you, he's usually pretty much trying to do so while fumbling with a map, namely, your health record.  And we don't like it.

A study in JAMA Internal Medicine found that patients were much more likely to rate their care as excellent when their physician didn't spend much time looking at their EHR while with them; 83% rated it as excellent, versus only 48% for patients whose doctors spent more time looking at their device's screen.  The study's authors speculate that patients may feel slighted when their doctor looks too much at the screen, or that the doctors may actually be missing important visual cues.

Indeed, a 2014 study found that physicians using EHRs during exams spent about a third of the time during patient exams looking at their screen instead of at the patient.  It is a dilemma; the records hold important information, and inputting new information is generally thought to be more accurate when done at point-of-care rather than at some point after the exam, so doctors are damned if they do and damned if they don't.

As one physician told the WSJ, "I have a love-hate relationship with the computer, with the hate maybe being stronger than the love." 

No wonder that the president of the American Academy of Family Physicians says: "We've taken this technology and we've embraced it, but I think a lot of us don't believe it's ready for prime-time. We've got this interloper in the exam room, but it's not there to help with the medical side as much as it's there to check boxes for insurers."

I might quibble that the familiar physician shibboleth about EHRs being there to serve insurers' purposes rather than to improve care perhaps is one reason why they are not ready for prime time, but I certainly don't dispute the fact that they are not.  After we've spent the past several years and over $30b of federal incentives to persuade physicians to adopt EHRs, physician satisfaction with them appears to be declining.

Know any health care professionals who rave about their EHR like they do their iPhone?

The problem is that we forget that the record is not the point.  It wasn't the point when it was on paper, and putting it in an electronic format doesn't make it the point.  The information in it is a tool -- just a tool.  It is supposed to help the physician diagnose the patient, and record what happens to the patient, so he/she can be better diagnosed in the future.  Figuring out what is wrong with a patient and what to do about it is the point. 

Paper records were siloed and made the physician draw his/her own conclusions without providing any assistance.  EHRs have the potential to draw data from larger patient populations, even if they don't yet do so very effectively, and can also give some assistance to physicians, like warning about drug interactions.   But working with them still involves looking at too many screens and having to populate too many boxes.  No wonder physicians are employing scribes

Don't get me started on medical scribes.

Let's picture a different approach, one that doesn't start with paper records as its premise.  Let's start with the premise that we're trying to help the physician improve patient care by giving him/her the information they need at point of care, when they need it, but without getting in the way of the physician/patient interaction.

Let's talk virtual reality.

Picture the physician walking into the office not holding a clipboard or a computer or even a tablet.  Instead, the physician might be wearing something that looks like Google Glass or OrCam -- not a conspicuous headset like Oculus but something unobtrusive (a concept that investors are already pouring money into developing).  There might be an earbud.  And there will be the health version of Siri, Cortana or OK Google, AI assistants that can pull up information based on oral requests or self-generated algorithms, transcribe oral inputs, and present information either orally or visually.

When the physician looks at the patient, he/she sees a summary of key information -- such as diabetic, pacemaker, recent knee surgery -- overlaid on the corresponding portion of the patient's body.  Any significant changes in blood pressure, weight, and other vitals are highlighted.  The physician can call up more information by making an oral request to the AI or by using a hand gesture over a particular body part.  List of meds?  Date of that last surgery?  Immunization record?  No problem.

The physician can indicate, via voice command or hand gesture, what should be recorded.  It shouldn't take too long before an AI can recognize on its own what needs to be captured; the advances in AI learning capabilities -- like now recognizing handwriting -- are coming so quickly that this is surely feasible.  Keeping an EHR up-to-date should be child's play compared to, say, beating Ken Jennings at Jeopardy! or Gary Kasparov at chess.

In short, the AI would act as the medical scribe, without the patient even realizing it or the physician having to worry about it.

More importantly, the AI could quickly pull up/synthesize any pertinent literature, or assist the physician in coming up with a diagnosis and/or treatment plan -- as Watson is already doing for cancer.  Maintaining and presenting the EHR are the finger exercises, if you will; helping the physician deliver better care is the main function.  And without intruding on the physician/patient relationship.

Building better EHRs is certainly possible.  Improving how physicians use them, especially when with patients, is also possible.  But it's a little like trying to make a map you can fold better while driving.  It misses the point. 

We need a whole different technology that subsumes what EHRs do while getting to the real goal: helping deliver better care to patients.


  1. Absolutely agree with your pushing the need for better technology. It's surprising that no vendor has launched new technology in the EHR world. Instead, we are asking physicians to be clerical staff also.

    I wonder if the federal incentives drove so much new demand that the incentive for a vendor to introduce innovative software became relatively unimportant to the vendors.

    1. Yes, I fear that the federal incentives and standards had the unintended effect of freezing the product as it was...why innovate if it could mean not qualifying for the incentives? Now the established vendors and their existing approaches are more entrenched than ever.

  2. I was a member of the ASIM for years and had great hopes. I have now quit teaching as the EHR makes it almost impossible. I pity the medical students now struggling with this mess. I am so glad I am retired. My own physician retired after telling me that he spent three hours every day after seeing patients entering information in the EHR. He quit.