EHR Event of the Month: "You Say Pento, I Say Pheno..."
- Michael Victoroff, M.D., Editor-in-Chief
We play a lot of “post office” in healthcare, sending, receiving and re-sending information at rapid speed across a wide bandwidth. Message loss is not a rare surprise – it’s a law of physics. This month, two cases show that sometimes the problem is in the telegraph, and sometimes the telegrapher.
From a nephrology resident:
Admitted an unconscious epilepsy patient. Supposedly on phenobarbital and other meds. Lab reported zero phenobarb level, so we did workup for other causes. Next day lab called: Analysis was for pentobarbital – not phenobarbital. Phenobarb was actually high.
Lab order and report both show “phenobarbital.” No human transcription was involved. Luckily the tech checked printout from analyzer against report. Lab not able to determine why this occurred. If we got “zero” potassium, we would just re-check, but zero phenobarb was credible. What do we trust?
Sometimes elves come in at night and mess with your system. Without a root cause analysis, we can speculate about a type of error that occurs when systems exchange data. Commonly, this happens after an update, when the index of a list of records (like names, supplies, tests, etc.) that is supposed to synchronize across two databases becomes corrupted. The notorious “mail-merge-offset” error is an example, where each message meant for one recipient gets addressed to the person just below on the list. This can occur from human error on a dropdown list. But, in this case, it looks like the machines did it by themselves.
To report a suspected EHR Safety Event, visit www.EHRevent.org.
Expert Commentary: "What's the Lesson of Genesis Burkett?"
- Michael Victoroff, M.D., Editor-in-Chief
The media has drawn our attention to the case of Genesis Burkett, a newborn who recently died in a Chicago hospital when an order for intravenous fluids was incorrectly transcribed by a pharmacy technician. Some journalists have characterized this a “computer error” because the ordering physician entered the order (correctly) into an electronic system, and the pharmacy tech re-entered the order (incorrectly) into an automated compounding machine. Moreover, a decision support system that should have alerted at the mistake had been disabled.
This case has been invoked as a warning about hazards of automation. Calls have been made for “mandatory reporting of adverse events related to HIT to a national database.” Uh, hello?
www.EHRevent.org is exactly the “national database” being called for. It is a fully accredited, federal Patient Safety Organization and part of an AHRQ network of error-reporting entities. Reports to EHRevent.org are fully confidential, and data that is collected is reportable only in de-identified or collective form.
Computerized Physician Order Entry (CPOE) systems have been shown both to reduce errors, and also to induce some. On balance, like most medical devices, they produce both good and harm, with the balance heavily on the side of benefit. But, Baby Burkett is not really a good poster child for CPOE error.
Playing Post Office
Mistakes in transmitting physician orders were around long before computers. Was this particular event an instance of human apraxia (performance error), memory lapse (transcription error), a comprehension gap (cognitive error), poor training, flawed interface or workflow design, production pressure, distraction/inattention, misperception, or – as usual – some combination? The phenobarb case sounds like a software malfunction. But, the Burkett error could easily have occurred (and has, millions of times) in a paper environment.
The question for error taxonomists is not just how to classify it, but how to learn from it. There is a tendency to think of “performance errors” belonging to the domain of surgery, with “cognitive errors” happening in information management. Yet, operating information machines – even if just screens, keyboards and mice – involves dexterity, hand-eye coordination and other sub-cortical skills. Likewise, ligating the ureter is not necessarily a problem of clumsiness.
EHRs and RECs
As part of the economic stimulus package of 2009, the Health Information Technology for Economic and Clinical Health (HITECH) Act included funding for Regional Extension Centers
. These localized "RECs" support providers with guidance and expertise on adopting EHRs, and help ensure that the providers qualify for incentive payments of up to $44,000/per physician for the "meaningful use" of these sytems.
, a component of The Texas A&M Health Science Center, is one of more than 60 RECs nationwide, and offers technical assistance and information on best practices to 1000 providers in Texas. CentrEast has created an EHR Support Center
to help healthcare providers to:
Easily report Patient Safety Events involving EHRs to the EHRevent Patient Safety Organization (PSO)
Easily report usability issues and best practices to help with your transition to EHRs
Track, update and document usability
For more information on the EHR Support Center, which links directly to www.EHRevent.org
, please visit the CentrEast site