Tuesday, April 20, 2010

Legible Gibberish: This EHR Toxicity Has Got to Stop

My frail elderly mother recently got really sick from not eating due to progressive, severe constipation.

She underwent a colonoscopy 3 months ago that found a rectal stricture. The GI doc was (reasonably) attempting to treat the stricture medically with those wonderful powders the GI people give, this time the angel dust made of polymerized antifreeze (polyethylene glycol) known as "MiraLAX®".

MiraLAX worked too much miracle, unfortunately. She couldn't go nearly at all without it, but using it she had runs that kept her going to the bathroom 24x7.

She fell in the wee hours last week going to the bathroom basically from lack of sleep due to diarrhea from the angel dust, hurt her rib cage, and was "out of it" when I found her on the floor. I called for an ambulance.

She was seen in the ED, and the ED doc also found a rectal stricture.

I watched the ED staff move around and minimize what I considered an information-dense, cognitively oppressive, PICIS ED EHR window to get to a positively gruesome, primitive, block-capital ~ 64x16 1960's terminal emulation of the Eclipsys 7000 system in another window, to capture admission information.

It looked and acted exactly like the very same TDS 7000 system I began my medical informatics training with at Yale
in 1992, that we found presented a horrid user experience even then, and that in 2005 Penn sociologist Ross Koppel found to be 'errorgenic' (Role of Computerized Physician Order Entry Systems in Facilitating Medication Errors. Ross Koppel, PhD, et al, Journal of the American Medical Association, 2005;293:1197-1203, link to JAMA abstract here and my commentary here).

I would provide a sample screen shot of the ancient glass terminal emulation, but cannot seem to find one online. It would probably violate a hospital's EHR contract for them to share one.

The dissonance of watching ED personnel deal with a cognitively oppressive 'overuse of GUI' Picis ED system, and then dealing with a caveman-and-dinosaur era 1960's character-based glass terminal emulation on the same hundred-million-transistor graphics chip-accelerated, multi-GHz CPU, modern Windows OS-based screen was simply painful.

(Not to mention that the hospital floors use a third, entirely different EMR -- one of my former subspecialist mentors had told me some time ago in front of mom that he found the floor EHR highly unsatisfactory and keeps his records on MS-Word on a thumb drive that he proudly pulled out of his pocket. I described other untoward EHR observations at the hospital in this 2007 post.)

She was admitted for observation for two days, 'tuned up' and rehydrated and sent home.

She needed to go back to ED a few days later for abdominal pain from being unable to void, was tuned up again, and referred back to GI.

At the GI office she was seen on an urgent basis by a covering doc due to unavailability of her primary GI specialist. That doc couldn't find in their (local, vendor du jour) EHR any mention of the stricture when I told her that's why my mom was there. She reviewed multiple GI EHR screens and finally found mention of it in the narrative after a few minutes, pointing at it on the screen with her finger, but not in the summary or diagnosis list.

The covering GI doc appropriately referred my mother to a colorectal surgeon. The covering GI doc printed out the ED records and GI records for the colorectal surgeon.

The colorectal surgeon saw mom the next day. I told him she had a rectal stricture.

He reviewed the legible gibberish from these EHR systems (which I found astonishing myself, compared, say, to a written ED chart or procedure note or even the simple written notes I used to keep on 3x5 cards), but he could not find any mention of a stricture.

I couldn't find the diagnosis in this legible gibberish collection, either.

The printouts were disorganized from the perspective of a physician, causing them to hunt for important data while drinking water through a firehose. Legible gibberish is a good descriptor.

There were plenty of details on minutiae (seven sets of vital signs, every order done in ED with time and date -- including placement of IV infusor and administration of normal saline). For just the IV placement alone there was almost an entire page of notes.

However, it read like Tolstoy regarding the important stuff.

I myself designed advanced reports in complex areas such as invasive cardiology that were terse, lucid and informative - almost 15 years ago. Who, I ask, designed these?
They show an ignorance of biomedical informatics and information science principles regarding presentation of information -- as if HIT is just an "inventory system for data" instead of a clinical facilitator -- among other deficits.

The ED notes from from the PICIS system would have caused me to flunk a medical student for poor medical record keeping. Worse, the diagnosis was of "abdominal pain" only, and that was buried seven pages into 10 page "summary" ED report (not at top of page 1 where it belongs). There was no mention of the ED doc's finding the rectal stricture.

I'm just waiting for the insurance denials...

The colorectal surgeon looked at me funny when I told him that two other physicians both told me that's what she had, but it was not on the computer record.

I'd been a resident with his partner, however, so I had good creds. He performed a brief exam and said, "oh yeah! rectal stricture! Pencil diameter! Needs surgery."

I ask:

What happens to elderly patients who don't have their son with them, a former medicine resident at the same hospital, who once fixed their CT scanner's hung computer in the middle of the night when service was unavailable and saved a life, being credited for same deed in a letter from the now-Chief of Staff to help secure a Yale postdoctoral fellowship in Medical Informatics?

This EHR toxicity is simply out of control.

** Disclaimer: The following is a purely personal opinion (my "narrative"), always difficult to write about without generating skepticism, and I admit possible non-objectivity in that regard. You, the reader, are free to ignore my observations:

My personal observations here are far from unique. The physicians at this hospital are generally excellent, but these systems do not appear to be serving physicians or patients (here and elsewhere) well.

Unfortunately, the major remediation required to fix them is probably not possible under current health IT leadership structures and orthodoxies.

Why am I so emphatic about these issues?

My father was injured in 1994
at a major Philadelphia hospital and eventually died - from two years of undiagnosed bilateral renal carcinomas and resultant hydronephrosis due to defective information flows - while I was training in medical informatics at Yale. The diagnosis was only made after I insisted a renal arteriogram be performed after I heard about repeated major GU bleeds and one so major it dropped his HCT several points.

I extended his life for six years by removing him from the big-city hospital and taking him to the hospital that is the subject of the above post (where I had done my residency), then mostly paper based. I'd taken him there to be under the care of my former mentors after the big-city Philly hospital doc had written him off as 'needing to get his affairs in order' and abandoned him to die (I could not get oncology or medicine at the big Philly hospital to see him after the diagnosis was made, the chair of medicine stating he "did not want to get involved").

My father underwent partial renal resections at the latter hospital and lived until mid 2000, avoiding dialysis until the terminal few months.


Now EHR's that cause poor information flows seem to be 'conspiring' to do my mother in at the very same hospital that saved my father.

Unfortunately for my father, medical malpractice was then followed by legal malpractice. My father won his 1995 medical malpractice lawsuit against the Philadelphia hospital, but posthumously, in part because I had the presence of mind to insist his deposition be forensically videotaped. He testified at the trial from his grave.

The lawyer representing him apparently had a substance problem and was in severe depression, not responding to the court's request for documents; nearly a dozen major med mal lawsuits including my father's were summarily dropped due to - lack of information. It took the major Philadelphia med mal law firm several years to get them reinstated, a precedent in Pennsylvania (http://www.aopc.org/OpPosting/disciplinaryboard/dboardopinions/21DB2000-Smith.pdf , page 20, Charge VIII: the Silverstein matter).

I couldn't make any worse stories up if I tried.

Finally, and ironically, the hospital now caring for my mother is the very same hospital that thrice, in 1998, 2000 and 2008 would not invite in the door for an interview or even phone me for conversation about posted informatics positions, even claiming in an odd, buzzword-compliant from the CIO alleging 'great care and forethought' in entirely blowing me off for a third time in ten years.


-- SS

Addendum 1:

Although I am not alleging malfunction, but inadequate designs, a reader sent this Health Canada recall notice about different systems from one of the vendors mentioned above. I thought it of interest with respect to the title of this blog post:

Manufacturer: Picis Inc.
Recall Posting Date: 2010-04-12
A) Caresuite Anesthesia Manager
B) Caresuite PACU (post-anesthesia care unit) Manager
C) Caresuite Critical Care Manager
Recall Start Date: 2010-03-08
Recall Number: 53683
Hazard Classification: Type II
Model or Catalog #
A) 4220
B) 4245
C) 1110
Lot or Serial #
A) Version 5.1 or later
B) Version 5.1 or later
C) Version 5.1 or later
Reason for Recall
During Picis post-market compliance activities it was identified that
there were two customer reports of malfunction within critical care
manager (electronic health record) application, which can potentially
impact the manual record keeping of clinical orders and since correct
handling of medication or infusion orders documentation are potentially
affected, patient safety may be impacted.

This recall involves systems affecting patients in critical care areas.

Addendum 2:

A reader pointed out that there is an automated "gibberish generator" at this URL. It took the first few paragraphs of this post and generated this:

The strical informating dust, block-capital emulance even I began Medicalled exactly from not eation, 2005;293:1197-1203, link to began ambulating withose worked found was seen the found the stricture medically go near on Errors. Ross Koppel dust wee horrib capital strication.

It loor. I found was "MiraLAX worked to primize angel found was seen informating to diarrhea from bathroom the colonoscopy 3 mothe was seen that week go the was "out in 1960's time TDS 7000 systems information.

If this were an EHR-generated summary, that would be about right.

-- SS