Sunday, October 12, 2014

Epic Fails And Deadly IT Cultures (An OPINION Piece)




The Ebola Virus...Image courtesy of scienceblogs.com

It's bad enough that a fellow from Liberia by the name of Thomas Eric Duncan through hubris, stupidity, or simply bad luck brought Ebola to our shores. He did ultimately seek medical attention in the Emergency Room of Texas Health Presbyterian Hospital in Dallas when he became symptomatic with the characteristic fever and pain of an Ebola infection. In fact, he presented twice to the Dallas ER. In between his two visits, Mr. Duncan was set loose on a city of well over a million souls while his disease was at its most infectious level. (He has since died of the disease, and sadly, one of the nurses who cared for him now has it. Let us pray for her speedy recovery.) How could this breach of public health have happened? It seems to have something to do with IT, specifically, the configuration of the hospital's EHR.

CNBC quotes Jonathan Bush (as it turns out, Jonathan is the nephew and cousin of the former Presidents...hat tip to Ranjan), head of Athenahealth:
The failure of a Dallas hospital's electronic medical record system to flag a man who turned out to be infected with the Ebola virus underscores how clunky, outdated and inefficient health information systems typically are in the U.S., a medical IT CEO charged Friday.

"The worst supply chain in our society is the health information supply chain," said Bush. . . "It's just a wonderfully poignant example, reminder of how disconnected our health-care system is."

"It's just a very Stone-Age sector, because it's very conservative," Bush said. "Hospital health care is still in the era of pre-Internet software."

"The hyperbole should not be directed at Epic or those guys at Health Texas," Bush said. "The hyperbole has to be directed at the fact that health care is islands of information trying to separately manage a massively complex network . . . People trying to recreate their own micro-Internet inside their own little biosphere . . . that'll never, never, never be excellent," Bush said. "There's no 'network effect' in health care today."
How does this apply to Mr. Duncan unleashing Ebola in the heart of Texas?
The hospital Thursday night said when Duncan was first examined Sept. 25 by a nurse, he was asked a series of questions, including whether he had traveled outside of the U.S. in the prior month.

"He said that he had been in Africa," the hospital said in a statement. "The nurse entered that information in the nursing portion of the electronic medical record."

But it turns out that answer—which could have alerted doctors of the possibility Duncan had Ebola—was not relayed electronically to them because of "a flaw" in the way doctors' workflow portions of the electronic health records interacts with the nursing portions of the EHR.

"In our electronic health records, there are separate physician and nursing workflows," the hospital said. "The documentation of the travel history was located in the nursing workflow portion of the EHR, and was designed to provide a high reliability nursing process to allow for the administration of influenza vaccine under a physician-delegated standing order. As designed, the travel history would not automatically appear in the physician's standard workflow."
Of course, that particular problem at that particular hospital is now fixed. But . . .
"We have made this change to increase the visibility and documentation of the travel question in order to alert all providers," Texas Health said. " We feel that this change will improve the early identification of patients who may be at risk for communicable diseases, including Ebola."

Bush noted that typically when problems like the flaw in Texas Health's EHR system are fixed, "they're fixed only at the place where they appeared."

"Those mistakes are happening constantly," Bush said.

But, "philosophically I think hospitals should get out of the business of trying to program computer systems, and expand in the business of treating patients. But that's a standard thing that goes wrong with millions of configurations" of EHRs, he said.
Mr. Bush was quite tactful, but the implication of his statement is truly astounding. He is saying, perhaps not quite in so many words, that the IT department of the Texas Health hospital in Dallas, by poorly implementing (my opinion, not necessarily his) poorly designed (again, my opinion, not necessarily his) software, could be responsible for a disaster. This glitch has potentially allowed Ebola to spread further than it would have had Mr. Duncan been put immediately into confinement upon his first presentation. To be fair, the patient had been in contact with others before his first ER trip; still, we can assume he had more interaction with more people than he might have otherwise. We can only wait and see how many of his family members and acquaintences come down with the often-fatal disease. I should also mention that the ER physician should probably have thought to ask about foreign travel when presented with a feverish African national presumably speaking with an accent.

There is much online about Epic, Presby's EHR provider. Google will supply link after link after link if you so desire. There are several take-away messages: Epic has severe interconnectivity / interoperability problems, and it is a HUGE political player, with its founder Judith Faulkner being quite the Obama supporter. Faulkner, and Epic employees, have given millions to Mr. Obama and other Democratic causes. Epic has received significant federal subsidy money, and it is up for an $11 Billion government contract. Michelle Malkin also reports that:
Faulkner, an influential Obama campaign finance bundler, served as an adviser to David Blumenthal. He’s the White House health information technology guru in charge of dispensing the federal electronic medical records subsidies that Faulkner pushed President Obama to adopt. Faulkner also served on the same committee Blumenthal chaired.

Cozy arrangement, that.
I'm straying a little off-topic here, but I think it is unlikely in the extreme that Epic will shoulder even the slightest blame for Mr. Duncan's Dallas destruction. After all, as we say in the trade, PBKAC, Problem (was) Between Keyboard and Chair. In other words, it wasn't Epic's fault that whatever IT employee or committee failed to connect the dots and the map the critical foreign travel field from the  nurses' intake screen to the doctors' review screen. Oops. So sorry.

Personally, for what little it's worth, I do NOT let Epic, or any other software company, off the hook quite so easily, nor do I bow to the IT departments which often control such software but don't grasp the criticality of the workflow they are now governing, let alone the workflow itself.

I've ranted for pages and pages about image sharing, and how it is malpractice for patient images to be essentially held hostage by the IT and other administrative types who are adamant that the competing hospital across town (or across the street) will NEVER EVER be allowed to touch their precious data. And I've yowled and whined about PACS software that was clearly NOT written for use by any practicing radiologist I've ever met.

Please indulge me while I add to these rants.

I had the occasion to accompany Mrs. Dalai to her annual (8 years postponed) internist visit. Her doc showed me how much fun it is NOT to order something as simple as a PA and Lateral CXR in our illustrious EMR's bilious CPOE (Computerized Physician Order Entry) system. It is a complete miracle that any order at all is entered correctly in this absolute abortion of an interface, and I'm not at all surprised when the wrong order comes through for the wrong indication. The electronic chart function isn't any better. Finding a particular lab value can be an exercise in agony (akin to using some PACS I can name) and it just goes downhill from there. When I asked around to find out who OK'd this particular piece of garbage, I was met with shrugs and silence.

Do you sense a familiar refrain? (Lawyers please note...THIS IS ALL MY VERY OWN HUMBLE OPINION, as is every other word that I have ever written or ever will write, unless quoted from someone else, and worth every cent my dear readers paid for it.) Once again, here in the Health Care Field of Dreams, we have badly written, badly designed software, created with minimal input from those who have to use it, selected and then implemented by IT types who also don't have to use it and don't understand enough about those who do to get it done right. This has to stop. Right. Bloody. Now. Hit CNTL-ALT-Delete and start over.

With Epic and the government having their hands deep inside each others' panties, we may well be stuck with these unusable systems for the foreseeable future. (And as an aside, if you deconstruct the Meaningful Use rewards and penalties, doctors are being bribed to buy EHR's that have the certified and confirmed ability to transmit data to Washington, D.C., so again, we won't expect the government to do anything about anything.) But, the demise of Mr. Duncan, and no doubt dozens if not hundreds more that he inadvertently infected between his two ER visits may level the playing field.

It is clear that Epic's epic Dallas fail (which might not really be totally attributable to Epic per se, but rather to the way the product was set up in the field, not passing that one lil' bitty critical entry to where it should go), contributed to Mr. Duncan's being released when he should have been locked up in the local version of Wildfire. It is possible, just barely possible, that this tragic episode will awaken the public to the dangers inherent in the IT-controlled medical software industry and its acronymbysmal spawn, EHR's, CPOE's, and the occasional unruly PACS. Get enough people upset about this, and they will call their congressmen, and more importantly their lawyers. (I would submit that more gets done by class-action suit in this country than by Congress.)

I realize that replacing these huge legacy systems which were outdated before they were even conceived would cost somewhere in the trillions of dollars, and so I'm not holding my breath that this will ever happen. But maybe a few million and billion dollar suits and fines would get the attention of the Epics, the Cerners, McKessons, and all the others who create these nightmares. Or maybe, just maybe, the execs will read this, and the other rebellious propaganda we are starting to see online, and realize that they are causing damage rather than progress, and be inspired to turn it all around. I'm a staunch believer in the electronic record, PACS, computers, iPhones, Apple Watches, and anything else technical. This is the future, without question. But it has to be done right, and so far that hasn't happened.

We can hope that the late Mr. Duncan can accomplish in death what no one has yet been able to manage while alive.  We can hope, anyway...

3 comments :

Macha said...

"Once again, here in the Health Care Field of Dreams, we have badly written, badly designed software, created with minimal input from those who have to use it, selected and then implemented by IT types who also don't have to use it and don't understand enough about those who do to get it done right."

I'm a doc who did IT in a previous life. You have hit the nail on the head. The software we have to use every day is a disgrace. Thanks for your column.

Unknown said...

what you say about the software design and implementation may all be true, however, given the wide publicity over the past several months about the Ebola epidemic, should not the nurse who first interviewed Mr. Duncan have immediately gone out and informed the ER medical staff about the possible problem? We seem to be giving common sense a pass while focusing on technology.

stacey said...

IT types think they account for every bit of possible behavior. Until they leave something out. But then, Software is never a solution or a cure for bad behavior. Even IF the fact that he was in Africa was noted in the EMR, don't you think, given the current level of "awareness" that she might actually SAY SOMETHING to the doctor? Hmmm Let's see, patient, has fever, mentions they have been to "Africa" which, by the way is NOT specific enough... Should differentiate at least West Africa or individual countries, since Ebola is not occurring in many of them... but I digress...Fever and locale should have triggered alarms in the Nurse's head and she should have escalated the information verbally.

When you remove humans from the equation and presume that the computers will do the thinking and the work, you discover that the systems do not actually think nor do they work. They merely make entries in data bases. And to be fair to the developers on this one, you can't expect them to anticipate every possible scenario, which is why things should be more open. But don't tell them that.
But then we agree on things...