(And blessings on your Mom for teaching you the most important part of how to write.)
If the patient told someone who was entering data that he was travelling from Liberia then:
a) I find it hard to see how that got 'edited' to a different value prior to his being admitted/treated in ER. After all, what motive could there be on the part of the person doing data entry to do that?
b) if it did go in as Liberia and he was discharged anyway then I'd suspect someone changing it later to cover their ass as the source of the error. :-(
c) if it did go in as Liberia even for a short time, I'd find it interesting that the hospital admission software didn't flag it on to a higher authority. If, as I would imagine is the case, the software is one that is common to many hospitals rather than being individual to that one then I would have thought that the company maintaining would have issued an update that essentially said if travel="dangerous places" and symptoms="any of these" then TELL SOMEONE or DO SOMETHING TO DRAW ATTENTION TO THIS..
Given the above I'd begin to wonder if the simpler explanation might be the old one that "everyone lies" and that the patient perhaps did not tell them he'd travelled from Liberia at all!
a) Apparently--and according to a hospital source--the travel data as presented in the record only indicated significant foreign travel if the patient arrived from locations where the flu season is different, so "foreign" flu would be recognized. Who authorized that isn't public knowledge at this point, but I can see how some programmer unfamiliar with medicine or a hospital administrator concerned about their most common disease-of-concern would trying to "simplify" things for ER personnel by giving them a binary answer when a more complex one was needed. Which would be a strong signal that the hospital is not used to receiving patients who have done a lot of foreign travel and may thus appear with diseases acquired abroad. (Most Texas hospitals that size--and many smaller--do see travel-related illnesses, but many affluent patients get to the hospital by way of specialists who've already diagnosed it.)
b) Never underestimate the incompetence of any system in a second-rate operation of any kind (not just hospitals.) It could be that the data were changed later, but I think it more likely that--whether by a systems-design error unknown to or approved by administration--the information was entered on the first visit and did not make it to the persons actually evaluating the patient. Or--worst case--that it did make it to them and was ignored because it was not understood. The details should come out in the ongoing investigation (as nurses' complaints from many quarters indicate that they have not received training they feel is adequate to deal with Ebola in their own hospital setting.)
c) Again...never underestimate incompetence. Not all hospitals use the same software. Did the person choosing the software for this hospital (possibly the entire Texas Health group) have the ability to check deeply into the software and also have a solid knowledge of what was needed for diagnosis of unusual (for this region) diseases? Did anyone programming the software know enough about hospital practice to go beyond providing records required to keep track of patients, their medication, and accounting? Were there locally customizable options, such as 'what do you want flagged in travel history?' with a set of options. Back when I was still involved in medicine, and recently involved in programming, there was some startlingly bad medical-related software on the market...but you had to have some grasp of both programming and medicine to spot it. Once in use, was it modified locally--did someone on that hospital's payroll, at the request of someone else, tinker with the software? (Every user would like custom-tailored software that's exactly right for them--and that includes me, though I don't tinker with the stuff I use now. I just cuss at the programmers in absentia.) The hospital claims it's already modified the software so it won't do that again (if it "did that" in the first place), so if I were one of the investigators I'd make them show my code-savvy assistants the actual code, and I'd try to reproduce the original situation, then reset it to something better.
(c) continued (separate because LJ caps replies at an arbitrary length and then doesn't provide a counter so you can tell when you get close. Speaking of lousy programming decisions.)
I think it really unlikely that the patient lied about having told someone he come from Liberia, for the simple reason that the patient knew Ebola was a possibility--even a probability--and knew that receiving treatment as early as possible was his only hope to survive. He was urban--worked as a limo driver, not a country job. Thus he both saw Ebola and knew the facts about it that many Americans still haven't grasped. It was in his best interests to tell medical personnel as soon as possible--it would be inconceivable to him that here--where supposedly there's "the best medical care in the world" as some of our politicians have it (wrongly)--a hospital would be ignorant of the connection. It's possible (slightly) that his accent was against comprehension (there are varieties of English I find hard to understand because the accents fall on different syllables and are tonally different; vowel sounds also vary. It takes me a little time to reset my brain for New Zealand's English, for instance, but also Indian and Pakistani accents and some African ones. In this country, I found Eastern Shore to be difficult at first.) But once he was sick, he would know he wasn't going to be deported unless he lived through the disease in isolation, because that would put more people at risk...and fear of deportation would be his only reason not to tell the hospital where he'd come from.
Significant detail from watching NBC news Friday evening.
Apparently the software doesn't display notes entered by nurses to doctors. Whether that is *all* notes, or just *some* notes wasn't specified (I suspect the latter, but wouldn't be too surprised if it was the former)
From knowing someone here on LJ who is not only a nurse, but *teaches* nursing, I'm sadly not surprised at this.
It's part and parcel of the "nurses are third class citizens" attitude in medicine. They can't *possibly* know anything significant to doctors. :-(
Mind you, there can be legit reasons for nursing notes to be kept away from doctors, as sometimes they contain stuff that probably should be kept from them (including veiled comments about how stupid the doctor is being).
But those notes should not have been classified as "doctor doesn't need to know"
It could be a design error (nurse notes never get seen by doctors). Or it might "merely" be a training error (say, there being a checkbox to make sure the doctor sees the notes and it wasn't checked, or there being a seperate place to enter notes you *want* the doctor to see).
Mind you, even with software written correctly, the training errors can be a royal pain. (I recall when I was getting called in several times a week at 3 am because users couldn't follow instructions on the screen to get data copied to a floppy. And they'd supposedly been walked thru it be their lead before)
I'm also reminded of a long-running discussion on the comp.risks newsgroup/mailing list. The medical field is *very* resistant to having and using (using *properly* anyway) checklists. But far too many errors, *serious* errors could be prevented by using them.
I'd say this is such a situation. Badly implemented "list" in software or failure to use it properly.
I think you're probably right. NEJM has had articles on the difficulty of getting a group or a hospital to change to better practices. And I personally understand the difficulty of learning data entry when systems change and nothing is in the familiar place, sometimes not even called the same thing (looking at you, Microsoft and also Mozilla, which for some insane reason decided to change wording in Thunderbird. If I'm in the newsreader, "Reply" now means "email to sender of post you're reading" and "Followup" means "reply to newsgroup.") Every medical office I've been in in the past ten years I've heard office staff complaining that "the new system" makes their life harder, largely because there's no standardization of the order in which data should be entered (so on older patient records it's here, but in the new system it's there...) or even the format (is Date of Birth M/D/Y or D/M/Y? Is the patient name field long enough? Does it call for Last Name, First Name, Middle Initial, or First Name, Middle Name, Last Name? Or something else? Will the system retrieve a record by patient's name, or do you have to know the patient's file number, or does it retrieve by SSN?
Clerical staff in both doctors' offices and hospitals are overloaded with paperwork, all of it requiring data entry in multiple formats--learning new ones every so often just adds to the load.
I used to write software for the QA dept at the company I worked for. And I made a point of talking to the users about the interface. Still didn't help when leads were too busy to train newbies. :-(
also didn't help when they demanded software do stuff the wrong way (produce reports of properties of a shipment by combining the results of sampling the several production lots that had been used to make it up).
And invariably you train the leads and not only do they not train the newbs, they don't use it themselves.
The last training that I did was to teach basic HTML/CSS to people who would update the public-facing web site. I taught the intro, then the web admin taught the advanced stuff of actually using the CMS system. We had high-level managers and department heads come to the training when there was no way they would ever be using the system.
2014-10-04 02:41 pm (UTC)
And now the good news.
The EMS crew who picked the patient up the second time. They were either told (by family or patient) or were alert enough to pick up on the symptoms (then more obvious) and the patient's origin, but they informed the hospital that the patient probably had Ebola. And they were willing to transport him.
I'm sure someone will question why they took him back to the hospital where he had been misdiagnosed before--and I'd like to know the thinking behind that decision myself. Probably the family told them to go to THPH, and typically EMS will transport to the patient/family-specified hospital. At first thought, he should have been taken to one of the best hospitals--a teaching hospital, almost certainly--whose protocols were more up-to-date. But if that hospital was full and had no space for a patient who would require a separate suite (for the prep room and the patient room) that's one reason not to go there. Another is that THPH was already contaminated by his earlier visit--full of people who were going to need to be closely watched and tested--and taking him back there would ensure that attention (life-saving attention) would focus on those people--something that might slip through if the patient himself were elsewhere. Confining the risk is a good idea. But I'm wondering if Dallas EMS already had a protocol in place for this. I'll bet they have one now. If it were MY EMS, I'd specify a particular unit in the metro area as potential high-risk transport, equip the station it's assigned to with the protective gear needed, have CDC conduct life-action training with the crews who ride out on it. Coordination with the best hospital for the case would have been set up, both crew and receiving hospital would know in advance where to go and what to do when the ambulance arrived. That way, when a call came in, the relevant information could be gathered, and this unit sent. Protocols for its subsequent decontamination and the necessary surveillance of its crew would be written out, checked with CDC for completeness, to minimize the chance for exposure and an outbreak.
But a big Yay for the crew, who undoubtedly had insufficient protective gear but didn't run off in a panic.
2014-10-04 02:52 pm (UTC)
Re: And now the good news.
I don't do anything that qualifies me as not being "anonymous," but my name is Karen.
Quote: "But a big Yay for the crew, who undoubtedly had insufficient protective gear but didn't run off in a panic."
2014-10-04 02:58 pm (UTC)
Re: And now the good news.
I'm sure someone will question why they took him back to the hospital where he had been misdiagnosed before--and I'd like to know the thinking behind that decision myself.
I can think of 3 reasons without even trying:
1) It was the closest hospital.
2) It was requested as the religious affiliation of choice.
3) It was requested as one his insurance would cover.
In other news, I still cannot comment on the Paksworld blog - I try again periodically. but they still don't show up.
2014-10-04 04:25 pm (UTC)
Re: And now the good news.
As a recently arrived foreign national, I doubt he had any insurance. I don't know that for a fact.
Closest hospital may be a reason, but once you have the penny drop that it's probably Ebola, that's a choice point. It's one that could even override religious preference. (How many Presbyterians are there in Liberia anyway? I dunno.)
I read an article that said his nephew had called the CDC to seek help and they'd walked him through getting the right help that time.
2014-10-06 03:57 am (UTC)
Re: And now the good news.
My husband read a different article that said it was an in-law who saw the petechia (signs of bleeding) in his eyes and realized what it might be, and that she called EMS and told them when the arrived she thought he might have Ebola--he'd been in a country with it. They put on protective gear before loading him into the unit.
I suspect, with interviews with different people, several stories will float around, and more than one will be true, but seen from another POV.
On probable factor in the first wrong decision, to send the patient away in that first visit, is one that will be hard to prove though many of us are fairly sure it exists. And that's race. This is a racist society. How one is treated in an emergency room does vary with race and with obvious ability--or inability--to pay for treatment, as well as with the ER's perception of the seriousness of the problem. This patient is African, dark-skinned, speaks with an accent no doubt (though I haven't heard him, of course, I have heard recent African immigrants and students) and is foreign. In an ER focused on the bottom line, this is a liability--someone they would want to get out of their ER, someone they would not risk admitting lest he be a deadbeat. Proving that this had any part in the decision will be next to impossible unless someone lower on the staff plays whistleblower--the hospital's lawyers will have told the spokespersons what to say and how to say it. They know--probably everyone on staff knows--that racism will be suspected and if proven will bring the feds down on them even harder. But Dallas is a very right-wing part of Texas, and I strongly suspect that racism and classism played a part in the decision (and they are now resenting the heck out of what happened next.)
Thank you for this. Very well-written and informative, and you address issues that absolutely need to be looked at.
According to this mornings NPR news a doctor is admitting that the information made it up to him!
Wow. Another failure point. I'll have to look into that. What did HE do with the information? When did he get the information (while the patient was in house, after the patient had been discharged from the ER, a day and a half later?) What did he do upon receiving the information? (A perfectly natural first reaction would be to think "Oh, SH*T! We're in for it now!" But AFTER that what did he do, and why?) Did he see the patient?
Your tire story was really weird to read, as I'd heard it before. Obviously a lot of people have been killed by that particular type of failure, but I knew I'd heard that story from my wife, and one of my wife's astronomer friends is named Sally, and her parents are both dead. Turns out that it happened to her husband, Ralph. His dad was killed by this failure mode, the passenger was able to steer the car to hit a tree on the driver's side, Ralph was in the back seat.
It would have been super freaky if it had been the same Sally in both our stories.
Re: Texas hospitals, I overheard at the Alamogordo pharmacy yesterday about a worker being tested positive for something and possibly continuing working. Being a mild hypochondriac, something that comes with missing a significant chunk of your immune system, I asked the pharmacist what they were talking about, concerned that it might be local. It wasn't, the story was about a nurse who tested positive for TB. She worked in the baby ward. In Dallas. Possibly in the same hospital as the Ebola patient. Yes, there are lots of hospitals in Dallas, I couldn't get first-hand info on it.
Stuff coming off of traffic going the other way kills a surprising number. One of the ambulance calls I was on, the driver was killed by a small sign trailer (a sign mounted on a small two-wheel trailer) which had been carried on a larger flatbed trailer with other similar signs). It fell off the back (why its restraint came loose I don't know) on a curve and of course kept going in a straight line, which put it crossing the center line...wheels behind, long trailer tongue in front...which bounced off something in the road, entered the driver side window--and by the laws of physics, pivoted on the fulcrum of the window-post, smashing the drive in the back of the head and into the steering column...then fell free as the car was heavier and kept rolling forward. Again, a front-seat passenger (the driver's son) grabbed the wheel and controlled the car until it stopped.
I've had 18-wheeler tires disintegrate, blow out, in front of me, as they passed and then pulled in ahead of me. Big chunks of tire flying back. The big highway I drive on oftenest has put in a divider between lanes that's supposed to prevent tires from getting into oncoming lanes, but it's not high enough (I'm not sure how high would be enough, but I've seen loose tires bounce higher than that.)
Oh--and re: health care workers working while sick/contagious. Many people can't afford to take time off work when they're sick. Which is why everyone needs paid sick leave. If a nurse, however, fails to report a positive TB test and continues to work with patients...the nurse is likely to be found out (not as soon as one could wish) and will face stiff penalties.
2014-10-04 07:46 pm (UTC)
Additional software data
According to local news yesterday, several of the hospitals in my city (not Dallas; nowhere even near Texas) have turned up a similar type of software problem with their electronic ER intake forms - there is no place other than "notes" to enter information like "I just got back from Liberia" and no way to flag the notes as "Vitally important; Read This Now" for the next person up the chain. Needless to say, there has been a scramble to find programmers who may be able to fix the problem, as well as additional emergency protocols being put in place (we are assured) to handle such a potential patient when one walks into an ER.
So the computer thing is not just a Dallas hospital problem; it's a problem that may well be endemic to ER hospital software. National news seems still to be concentrating on Dallas, but at least here they're trying to learn from the problem and avoid it.
2014-10-04 08:13 pm (UTC)
Re: Additional software data
It's great that they're trying to fix it. In the meantime, there's a simple, inexpensive, non-digital solution. Train the ER intake staff on it.
Have yellow or orange 3x5 cards for them to use, stamped PRIORITY E. When someone comes in admitting a history of recent (within 21 days) presence in Guinea, Sierra Leone, or Liberia, hand the patient one of those cards, attach a second one to a clipboard writing in the , pick up the phone and call the hospital number the hospital has chosen to use, and say "I have a Priority One possible, travel history positive."
The advantage here is that it's fast (every office supply store has yellow 3x5 cards, some have orange as well), cheap, and the color is bright--it will be noticed, if it is used ONLY for serious infectious diseases (and, right now, only for Ebola.) It can be accomplished in however long it takes to get hold of those yellow cards and have someone stamp them with "priority" (a stamp I've seen ready-made in some office supply stores--an probably quickly made anyway) so the clerk can add the E. It does not require physical contact with the patient. (a hand-stamp would be a positive way of IDing the patient but a) is less visible on dark skinned persons and b) requires contact or very close near contact with the patient.) The criteria--within the past 21 days has been in one of those three countries and is sick--can be printed in large letters on a sheet of paper and taped to the desk as a reminder.
Quick, inexpensive, requires little training time or money.
Programmers digging into the code to change it...longer.
Also cost more. (So would I, if I were still writing code.)
Thanks for the in-depth analysis.
One thing that keeps surprising me is how few people are able to recognise potential systemic failures.
I've worked in a number of environments where every problem was addressed individually, and that particular hole fixes, usually after something bad had happened (and all of those incidences were minor) - but any time I've raised a flag of 'hey, we've got several problems all showing the same pattern, can we fix the underlying issue' I've been told 'no'. Too expensive, no need. And so organisations go on to carry systemic problems around, because so far they haven't been catastrophic.
The lesson for computer types (and managers!) is that bad interface design can kill. It's not just about prettyness or convenience; nurses not being able to flag up or identify vital information IS a matter of life or death.
Bad interface design includes frequent changes of interface offered the user as "improvements." Learning new design takes computing power away from anything else the user is doing and increases the chance of error--the more stress the user is under, the longer it takes to learn a new layout. Keep important field where they have been visually; keep command buttons looking the same, and in the same place--don't just decide to move them for the programmer's notion of what looks better.
Have you ever seen the blog post "Your app makes me fat?" Brilliant, very clear explanation of why New! Improved! Shiny! With Bells! And Whistles! can make things worse for the user, not better, not even better months later. http://seriouspony.com/blog/2013/7/24/your-app-makes-me-fat
Especially for people who work on the computer, and whose work needs to be accurate and the output useful for others--when the computer is the tool, not a toy, a diversion, entertainment, or the *subject* of the work--unnecessary changes make them less productive, even when the intention was to increase productivity.
Spokespersons for Texas Health Presbyterian Hospital are STILL doing the CYA thing, claiming they're wonderful, their staff is wonderful, and they did EVERYTHING that could be done.
When clearly they didn't. With that level of denial, improvement becomes impossible. The people who cannot admit error even when it's pointed out do not belong in positions of authority.