Garbage in, garbage out: The EMR in action

29th February, 2012

When can we rely on the information in the electronic medical record?  Ever?

My first patient of the day was a gentleman in his 50s, who’d had previous surgery for prostate cancer.  He was scheduled for replacement of a defective penile prosthesis.  His history and physical examination (H & P) form was computerized and legible, as opposed to the handwritten scrawls we always encountered in the past.  However, it wasn’t much help.  It clearly stated that the patient had no prior surgery, which was silly.  The truth was obvious–we were sabotaged once again by the pernicious ease of entering wrong data into an electronic medical record.

How much of the other information in that patient’s H & P was true?  It’s impossible to know.

I think we can all agree that we expect more of some H & P’s than others.  If the gynecologist or the orthopedist has done the physical exam, especially if it’s a “short form” for outpatient surgery, I don’t pay a lot of attention to the documentation of heart sounds.  I’m pleasantly surprised if the existence of the heart is acknowledged at all.  But when a consulting internist performs a preoperative H & P, especially if the patient is genuinely ill, we have every reason to hope for better.  We may be disappointed.

With handwritten H & P’s, if parts of the exam were omitted (like the rectal exam on a patient for eye surgery), they would be left blank or “deferred”.  Now, what we see is compulsive documentation that may or not be true, because the computer systems won’t let you leave any field unpopulated.

Recently I’ve seen a normal cardiac auscultation documented in the case of a patient who actually had the loud, harsh murmur of aortic stenosis.  If I can hear it, I can assure you it wasn’t subtle.  False negatives are worse than no information at all, because they can lead you to think your patient is healthier than he really is.

Seemingly harmless inaccuracies can lead to serious consequences.  One night when I was on call, a very happy patient was admitted to have a kidney transplant.  A match had been found for her from a cadaver donor.  But when I examined her, I was startled to feel a firm, three-centimeter mass in the right side of her neck.  Both her primary physician and the renal specialist had recorded her neck exam as normal, with no palpable masses, on the computerized record.  Was this a new finding, or had the mass been there before?  Had anyone actually examined her neck?  No one knew.

The surgeon had no choice but to cancel the transplant and send the kidney to another recipient.  The mass in the patient’s neck had to be evaluated to rule out cancer first.  The patient was devastated, and left the hospital in tears.

The mere presence of legible documentation doesn’t make it thorough or accurate.  On a computer, it’s perilously quick and easy to check off a list of negatives, especially if they’re all mandatory fields.

I’ve developed the following practical guidelines for critical interpretation of the electronic H & P.

If a positive history or physical finding is described, it’s probably true.

If a negative history or normal finding is documented, one of the following is true:

The question was never asked, or the exam never done.
It was done in a hurry.
It was done by someone in training who gets most clinical information from Wikipedia.
It was done properly and is really negative.

A long list of negatives should be viewed with suspicion unless the patient is a healthy athlete under the age of 25, and perhaps not even then.

Back to my patient who needed a new penile prosthesis.  I established that I needed to do my own H & P in order to have information I could trust.  That done, I took a longer look at the internist’s recommendations for perioperative care.  I append them verbatim:

“Pt is at low risk for surgery. Please avoid shifts in Blood Pressure and
Volume. As is true with all surgery the anesthesiologist should mind
the blood pressure as this will reduce any unknown cardiac risk the
patient may have. A profound anemia would add further risk, which
this patient has no evidence of. Should heavier than expected bleeding
occur, please keep Hct over 30 for further cardiac risk reduction.”

Although I can’t prove it, I would bet money that this internist had created a standard note on his computer with sage advice for the anesthesiologist, appropriate to any situation.  Thank goodness.  How would I have managed without it?

Is this the quality information we can expect from a completely paperless system?  Computers, after all, don’t generate content; they only store it and make it available for retrieval.   At the end of the day, if you put garbage in, you’ll get garbage out.  Whatever time we thought we saved will be spent sorting through the trash.

2 COMMENTS