The New Tricks

(Note:  This is another bit of musings from 2007 when I was making the transition from conventional clinic work to free lancer.  One of my great anxieties was learning computerized medical record systems.  As it turns out I have subsequently learned six different systems, two of them well enough to trick them into efficiency.) 

Can you teach an old dog new tricks?  I don’t know that the question has ever really been resolved.  But at least it looked as if I had come to the right place to try.  There I was, at the Organizational Learning and Development (OLD) office.  And for the first time in my quarter century of practicing medicine I was going to learn an electronic health record system.

I confess, I liked the old ways.  Paper charts had a sense of authenticity to them.  Like the patients themselves they slowly increased in girth, and even acquired wrinkles and blemishes on their once smooth manila surfaces. 

The paper chart was a personal artifact, like a scrapbook.  As you paged further and further back you traveled in time.  Here is the pre-employment physical.  A little deeper in, there’s the broken wrist from the Little League game.  Go back to the earliest pages and you find the runny nose, the childhood immunizations, and the Apgar scores of a newborn.

But we must change with the times.  Electronic health records are the next big thing to contain costs, improve quality, and increase patient satisfaction.  So here I am, with a new practice, a new employer, and trying to learn the new tricks.

As it happens, there are an awful lot of passwords involved.  You need a different one for the network, the EHR and the radiology program.  They each have to have at least one capital letter, one number, and no less than seven total characters.  Its somewhat like playing Scrabble.  I generally base it on something easy to type, and that has sentimental value.  Anti-machine mottos like Luddite and Landru are my personal favorites.

Along with a handful of other neophytes I did my training in a classroom full of monitors, presided over by an energetic trainer who could make the cursor fly across the screen like an angry gnat.  Click and this happens.  Right click and link to another menu.  Hold the shift bar down and a towering column of orders waiting to be signed go down with a gentle tap of a key.  It’s all pretty impressive until the clinician part of your brain starts to ask nagging questions.

“Um, how can I confirm that the instructions to arrange a test have actually been received?  It just vanishes off the system once I send it.”

The trainer ponders this a while, and to give her credit it does seem as if her faith is not entirely dogmatic.  Who, I wonder, would be sued if something important fails to happen?  The nurse who never got the message?  Perhaps the software writers?  I doubt it.

Now I am starting to wonder about other things; generally a dangerous occupation.  The electronic prescription system is legitimately nice.  No more phone calls and faxes, just line ‘em up on the screen and click.  Of course it puts little roadblocks in your way in the form of warnings of potential drug interactions.  Sometimes these are worth considering.  But while safely in non-patient care training mode I typed in scripts for birth control pills for an 80-year-old man, and thalidomide for a 25-year-old woman.  The computer cheerfully put them both through.

Now my fellow novices are starting to get into the spirit of things, and we start feeding diagnoses into the problem list module.  Ehrlichiosis, pyelonephritis, quaternary malaria, otitis media.  I must give the machine its due, it handled the common ones at breakneck (odontoid fracture) speed, could sluggishly handle the oddballs with the help of an auxiliary index, and it taught me the correct spelling of ehrlichiosis.

After a while you get a sense for what the EHR does well and what it does poorly.  This is the key to getting along with it, because to effectively work an EHR what is necessary is this:  you have to stop thinking like a clinician.

Physicians who fail to understand this generally think that the computer is out to get them, to make their lives difficult.  Untrue.  The computer is not there to make your life easier or harder.  Frankly, it seems indifferent to us.  What is important to the computer is order.

Every time you open a patient chart you must take some action.  Every action, be it a test ordered, a procedure done or a medication prescribed must be documented.  Every action must be linked with a diagnosis code.  And everything must be associated with a charge.

Line up the patient’s open chart, an action, a diagnosis and a charge, and the computer, its quest for neatness and order appeased, will agree to anything.

Once you stop arguing with this implacable sense of tidiness the EHR will actually help you out in some ways.  I am for instance quite fond of a letter-writing feature.  Click on the patient’s name, call up a template and you instantly have a nice letter outlining all the lab work from the most recent visit, with normal values and a section for comments and instructions.  Hit a button and it prints off a copy to mail and simultaneously plugs it into the patient’s record. 

With the radiology program I have the patient’s x-rays on the screen before they make it back to the exam room.  The ability to zoom in and out on areas, and to call up old films for comparison is a striking improvement.

I also like the feature that allows you to “copy” a patient’s chart to their primary physician.  In a fragmented medical system this is a godsend. 

But I still dictate most of my charts, because the “templates” that allow you to document patient visits with a click of the mouse are still atrocious.

In theory you would do this while still in the room with the patient.  But at least in the exam rooms I work in this would require you to turn your back on a patient seated on the exam table, and that seems rude.  I also wonder to what extent patients will resent our taking their problems; their deep felt fears and reducing them to a few seconds of data entry. 

So I work with the computer, not through it.  A quick look at the patient’s history before entering the room, then the traditional face to face conversation with a few jotted notes for later reference.  If a question comes up, say on past lab results, a quick log on to the record and a quick, efficient answer.

My other objections to template based patient encounters are that they generate a large amount of data, but it is dry, dull, and of dubious specific content.

You could for instance click General Appearance: ill.  It might well be true.  But how much more of the truth is revealed by the description of the patient as “cachectic” or “gaunt”, or even “in extremis”.

Often the templates fail us by being too inclusive.  Given a required box to check for BUS Normal (Bartholin, Urethra, Skene’s glands) most everyone is going to click on it, never mind that the memory of just what Skene’s glands are supposed to look like is filed away with other distant memories of medical school lectures.  Did you really check the pupils for reaction to light AND accommodation?  On that routine physical did you really confirm that the nasal septum was midline?

This shotgun approach to documentation does ensure up coding and higher payment rates, for such is its purpose.  But it worries me.  If you ever had to take this stuff into a courtroom, would the admission that you might have just glossed over an area and clicked normal bring your entire defense toppling down?  To the extent that these templates encourage us to be more complete they are a good thing.  But this quest for better reimbursement by broader documentation has its hazards.  At a minimum it diverts you from areas where your attention may do the patient more good.

The computer also ruthlessly homogenizes our record.  Any “foreign” influences are expunged.  The Germanic K in EKG is converted to ECG.  Various Britishisms such as haemoglobin are tidied up.  And Latin is declared a Dead Language.

I resent that a little.  I was taught to write prescriptions by my father, and he learned it from Old School types who took their Latin seriously.  I know that Sig, Rx, QOD and all the other archaic forms are holdovers from a distant past, a legacy handed down from the Medieval Barber-Surgeons.  To them it was both the international language of the scholar and a way to preserve the mystique of the healing arts from the common man.  But it has been a long time since we have been able to hide behind that particular curtain, so perhaps the computer refusing to speak anything but plain English is just due recognition of the new era.

I guess I should make passing mention of a feature that makes me smile every time I use it.  When I order something that the EHR questions I have the ability to hit a button labeled “Override and Accept”.  In effect this is telling the computer to “Just do it!”  To which it replies, “Master, it shall be so.” while no doubt muttering in some deep subroutine “At least until version 6.6 comes out”.

To other senior physicians wondering if they can make the change to EHR I would offer reassurance.  Of course you can do it.  You are trained to recognize complex patterns, you are tough enough to persevere, and your work ethic is at least the equal of the younger, technophile physicians of the next generation.  The few exceptions in my experience are physicians either too close to retirement to muster the energy or too deeply mired in work habits that are woefully inefficient whether you do them on paper or electronically.

Will EHR systems deliver on the promises of their backers?  I’m uncertain.  With regards to costs there are some efficiencies in potentially averted duplication of work, and in reduced clerical help.  But the emphasis on better documentation as a means to “up coding”, or “right coding” if you prefer, is going to result in increased demands on the financing of health care.

I do see the potential for fewer errors and better patient satisfaction with electronic systems, particularly if coupled with effective patient education materials.  I really like having several quick and good medical search engines on my patient room terminals.  There is nothing like showing a picture of what the rash of Lyme disease really looks like to convince a panicking tourist that the microscopic red dot on their arm is not going to be fatal.

But as I said, you have to change your way of thinking.  The rigid, organizational “mind” of the EHR is not going to bend to accommodate the more intuitive thinking of a good clinician, who often “knows things” without being quite sure how he/she knows them, and then uses technology to confirm or rule out these impressions.

Computers and their software continue to grow more sophisticated.  We are in effect driving the Model T in the year 2007.  The systems are going to get better.  And in a few years the tension between clinical thinking and machine thinking should be less, if for no other reason than we will by then be training physicians who have tinkered with computers since they were in diapers.

Hopefully they will be kind to us old timers.  Figuratively we grew up in the horse drawn era.  Yes, we can learn to drive that Model T, and even to appreciate its virtues.  But we still occasionally get the urge to pull out a buggy whip and smack it across the hood a little.

The computer is more or less indifferent to that too, but it makes me feel better.