Given what is now known about how the case of Thomas Eric Duncan at Texas Health Presbyterian was handled, the attempt to blame the hospital’s electronic health record for the missed diagnosis sounds pretty lame.
But people are still doing it:
Critics of electronic medical records have found a case they will be talking about for years.
Consider this argument from Ross Koppel and Suzanne Gordon:
While it is too early to determine what precisely happened in this case, it is not too early to consider the critical issues it highlights. One is our health care system’s reliance on computerized technology that is too often unfriendly to clinicians, especially those who work in stressful situations like a crowded emergency room. Then there are physicians’ long-standing failure to pay attention to nurses’ notes. Finally, there is the fact that hospitals often discourage nurses from assertively challenging physicians.
Long promised as the panacea for patient safety errors, electronic health records, in fact, have fragmented information, too often making critical data difficult to find. Often, doctors or nurses must log out of the system they are on and log into another system just to access data needed to treat their patients (with, of course, additional passwords required). Worse, data is frequently labeled in odd ways. For example, the results of a potassium test might be found under “potassium,” “serum potassium level,” “blood tests” or “lab reports.” Frequently, nurses and doctors will see different screen presentations of similar data, making it difficult to collaborate.
Another technological issue is the flatness of electronic records: Much of the information looks the same — a series of boxes to check and pre-formatted text that makes highlighting an urgent or important issue difficult. Electronic records, with their cut-and-paste functions, create what doctors call “chart bloat.” The announcement that Duncan’s electronic records totaled 1,400 pages illustrates this phenomenon. Poor record presentations may well have contributed to the hospital spokeswoman’s initial statement that Duncan’s temperature was only 100.1, when in fact the hospital’s records show it increased from that to 103 by the time Duncan was discharged four hours later.
A Textbook Case
This is the kind of the thing that overworked doctors and critics of electronic health records have been saying forever: This stuff is great and everything, but there are problems and you’re ignoring them. And we need to do something about it.
On the other hand, if you’re one of the people who thinks that EHRs are the problem and not the solution, consider this:
Clinical decision support could have stopped the current outbreak dead in its tracks.
All it would have taken was a single electronic warning and Thomas Eric Duncan would most likely not have been sent back home. The CDC would have been notified. Duncan would not have potentially infected scores of people in the Dallas area. Amber Joy Vinson would never have boarded a Frontier Airlines flight to Cleveland.
And we’d all be talking about something else other than Ebola, most likely football and the ISIS crisis.
Risk and Opportunity
There’s no doubt that we’ll be hearing a lot about innovation in the weeks to come. People will be building apps and inventing alogorithms and coming up with helpful new subscription based services designed to stop the global pandemic.
History teaches that most of these ideas will be stupid, a few will be smart and a few may even be game changers.
The game changers may not be ambitious services that use Big Data to categorize all people with 99.9 degree fevers and stomach aches and other Ebola-like symptoms into cohorts and index them using brilliant algorithms, although they could be.
They may be simple ideas that are easy to implement. One early example:
Data Points :
Privacy: Much of the weird slowness of the CDC response to the outbreak has to do with the invisible electric boundary lines that divide the health care system into a thousand different units and sub-units.
If the agency seems tentative, it is.
That’s because twenty years of political battles in Washington have made it so. We need to have a grown up conversation around privacy and transparency. Where do we stand on the rights of individuals versus the rights of society? And where are we on the rights of a business to put its own interests ahead of society, as appears to have happened in the early days at Presbyterian with potentially cataclysmic results?
Early Warning : Critics have pointed out that the healthcare system has been carefully designed to allow high risk patients carrying infectious diseases to penetrate as deeply as humanly possible with in it before they are detected. Unlike the healthcare systems in places like Liberia and Sierra Leone where there are only a few healthcare workers, the U.S. healthcare system deploys a massive show of force to get things done. Front office staff. Nurses. Doctors. Nurses assistants. Lab techs. Physicians. Specialists. Sub specialists. Hospitalists. The mind boggles at the number of people involved.
In a case where things go South, a potential Ebola paitent will have contact with six to ten people before anybody figures out that somebody needs to do something. They’ll sit around a waiting room for ninety minutes and then enter the system, exposing God knows how many other people.
It shouldn’t take a a specialist in infectious diseases from Harvard to tell us that’s a big problem.
Why not screen patients before they reach the front lines? Let’s figure out a way to do it remotely. Do intake electronically. Have people check in using their mobile devices. This seems like common sense.
Protocols and Guidelines: Officials have been shocked by the confusion among healthcare workers about which procedures and guidelines they should follow. It’s hard to understand the scope of this problem until you study how many competing sources of information there are out there. As the CDC releases new guidelines, there is going to be complete confusion about which ones to follow. The old ones? The newer ones? Version two? Version four? The EHR is the most logical place to get updates to people.
EHRs traditionally have not been about pushing this kind of content. After all, people have other sources of information: why should vendors worry about it?
They should worry about it because it’s their problem.
The focus group results are in. Faced with criticism about the quality of the information their systems generate, it’s been clear for some time that EHR companies need to evolve into content companies. They need to learn to help organize information and build tools that offer users context and allow them to sort and filter information.
In the end, building a smart EMR may be as much about the quality of the content you deliver as the interface you come up with.
/ john irvine
Categories: Uncategorized
Wow! At last I got a webpage from where I know how to truly obtain valuable data regarding my study and knowledge.
FDA has made it pretty clear they don’t want that tar baby.
Excellent idea to throw the EHR under the bus, and in front of the FDA.
Yeah, and it’s equally hopeless to joust with untraceable troll Curly_Moe_Larry MD, comedic open mic night genius whose money shot is limited to making fun of my name.
“Under the direction of the Food and Drug Administration Safety Innovation Act (FDASIA) of 2012, the FDA is working with FCC and ONC to propose a strategy and make recommendations on an appropriate, risk-based regulatory framework for health IT that promotes innovation, protects patient safety, and avoids unnecessary and duplicative regulation.”
http://www.fda.gov/MedicalDevices/ProductsandMedicalProcedures/ConnectedHealth/ucm338920.htm
@Adrian_Groper has missed the point. Vendors would not be liable if the FDA had approved their devices.
And it is hopeless to explain to @Booby_Gevalt that EHRs and their ordering and decision support components are indeed, medical devices.
“The two most important things the public wants from EHRs are: 1) patient safety–ie keep me alive, don’t miss critical symptoms, save a life.”
___
No, that’s what they want from their doctors and nurses and clinical institutions. You seriously believe a viable MedMal defense on the stand under Oath would be “Epic caused the patient’s death, not my failures of adequate process and clinical judgment?”
It would certainly comprise an interesting case, and, again, under the prevailing Joint-and-Several Sharknado tort regime, Deep Pockets Judy might be found to have contributory liability, but I wouldn’t want to be the clinicians in the crosshairs facing that plaintiff’s Counsel.
Of course the Epic EHRs did not flag Ebola or African nation—but it’s surprising there was no flag for 103 temp or severe pain or both.
Should anyone leave the EHR with a temp that high? No.
The two most important things the public wants from EHRs are: 1) patient safety–ie keep me alive, don’t miss critical symptoms, save a life and 2) don’t sell or disclose my data w/consent.
We get neither: EHRs aren’t safe and all our data is sold today by 100,000 health data suppliers covering 780,000 live daily data feeds. See: http://tiny.cc/z042nx
The only time PHI should be disclosed w/o consent is to save a life or prevent an epidemic.
@Curly MD – Be careful what you wish for. As vendors become liable for clinical decision support, your responsibility will dwindle and so will your malpractice insurance. And your role. And your pay.
For example, airline pilots are licensed professionals that do not control their own technology. Is that where medicine is headed?
“The vendor”?
Good luck with that. Though, in a “joint-and-several-liability” legal world, I guess we’ll see.
Epic is simply a huge database beneath a GUI set comprised of a large array of modifiable templates. It’s not a “device.” The templates and the workflows are user-modifiable — they call it “physician personalization” — so, I repeat:
“the failure here is proximately one of process, not technology. To the extent that clinical management permits technology to do their thinking for them is an abdication of professional responsibility”
And, yes, I’ve searched a lot of EHRs for a lot of keywords (at my REC we, being “vendor neutral,” served clinical clients using 40 different EHR platforms. In fairness, such searching is typically part of report generation modules, not ad hoc on-the-fly queries a clinician might quickly need in exigent treatment circumstances.
Lots of room for improvement, to be sure. But, delightful as it may be, Judy bashing is not gonna get us very far.
@@@Now, I’ll take it as a given that physician and nursing workflows within Epic were negligently silo’d at the time TD presented. Who is culpable for that?@@@
Answer to your question Bobby: The vendor.
Did you ever search an EHR for “travel” ? For anemia? For “sleep apnea”? or anything? Not a function of any EHR. Just silos and silo searching. Pathetic that lives depend on such crap devices.
“The debate should be about a complete redesign of the EHR around a patient centric workflow.”
Absolutely agree. We should be discussing new ways of designing electronic systems that support clinical work.
http://ehrscience.com/2014/09/29/from-clinical-work-to-clinical-productivity-its-the-little-things/
http://ehrscience.com/2014/05/12/developing-clinical-care-systems-from-the-outside-in/
Agree that Joe Flower has done a pretty good assessment.
I would add that physicians in general practice under crappy working conditions that promote fast thinking, cognitive overload, and oversights galore.
We do 20% of what should/could be done and are hoping we score 80% of the gains…which we maybe do but we miss a lot, and when we miss something like TD’s case it can be catastrophic for PR and public health reasons.
(For the many people who are harmed or killed by our oversights, it is catastrophic for them personally but as a society we are somewhat used to this, even though we tsk-tsk over the many deaths attributable to our suboptimal work and keep emphasizing better quality and safety work.)
EHRs — as currently constructed — are a big part of the problem. But we certainly need well-designed machine help…not sure how soon it’s coming, but we need it…so that we can focus on the human interactions that cannot be replaced by computers.
If there was not an EHR running the case of Duncan, and the prior system of care with its highly evolved safety nets was in place, Duncan would have been identified as being 5+ sick and admitted.
The social, human and financial costs of the failed EHR directed management of Duncan are huge.
You shoulda been at Health 2.0 2014. It’s coming.
Don’t get Vik started on Frieden. 😉
@Talos – “I am going to say that this solution still needs work.”
I have never argued otherwise.
Nope, not the “cause” but sold to us as part of a solution to those 100K. Since we have not seen a significant lowering, I am going to say that this solution still needs work.
We’re going to be bored to death if all we are doing is helping billers and making executives and ONC smile.
We need what IBM is calling “cognitive” computing: intelligently process online information. Watson; DARPA’s automated reading of medical literature; neural networks for non-text processing; high speed classifiers, Bayesian networks, support vector machines. See Science 10 Oct.
I want AI…help in diagnosing porphyrias, Mediterranean fevers, hemolytic anemias, antibiotic selection, et al.
We need ongoing summaries of viral and bacteriological loads in our neighborhoods from local and state health department laboratories.
The fun has got to come: We want to live longer, be smarter, keep our hair, have muscles, be sexier, have more energy, lose fat and have joie de vivre. Ditto for our patients.
We are not here to be productive input factors for firms.
Again,
“To the extent that clinical management permits technology to do their thinking for them is an abdication of professional responsibility…”
I’ve just gone back and again re-read THCB’s post “Statement from the Dallas nurses.”
Not one word about Epic.
Now, I’ll take it as a given that physician and nursing workflows within Epic were negligently silo’d at the time TD presented. Who is culpable for that?
If the permanent storage was:
Under individual control
Within an individual’s ‘cloud’
Unconnected to other accounts
And:
Access rights were granted by the individual
Recorded in the cloud by the individual accessing the information
Prosecution for unlawful entry or sharing of such data
No contract of adhesions permitted (assuming this is the correct term)
Would permanent storage then satisfy the basic needs of the patient?
And, of course, the insinuation is that HIT is a major “cause” of those 100k?
Yes, bullets kill people at somewhere around 30,000+ per year…and hospital-related mistakes, the ones EHRs are supposed to help prevent, kill over 100,000+ per year.
Really nicely summarized, Joe. I agree with your pithy assessment as well as with nearly all of what’s in this comment thread.
Nice! Have you been fencing with gun enthusiasts recently?!
Ok. The Vinod-ization thing was probably unfair.
The end of doctors, great conversation starter. Going by his recent comments it sounds like he’s modified his tone. I doubt he believes it either.
I think I just wanted to say “Vinod-izaton” …
Bullets kill people 😉
“Although they’re on the front lines of care delivery, and the most frequent users of EHRs, an overwhelming 98 percent of the 13,650 licensed RNs polled by Black Book for its latest EHR Loyalty Poll say they’ve never been included in their hospitals’ IT decisions or design.”
http://www.healthcareitnews.com/news/nurses-not-happy-hospital-ehrs
@John – I’m strongly in favor of both decision support and alarm triggers but clinical technology and mass surveillance need to be transparent and under socially acceptable controls.
I would be surprised if Vinod Khosla believes that open, peer reviewed medicine and personal control of technology are a thing of the past.
I see your broader point on decision support, Adrian. That’s fair.
Saying that you’re against the “Vinod Khosla-ization” of medicine through widespread adoption of clinical decision support is one thing, saying you’re against alarm triggers for pandemics is another ..
I’m sorry. I want an alarm.
This case is pretty compelling evidence that one is needed.
“Bobby, I’ve never spoken to Hannity or Beck—I spoke to CBS.”
__
You take everything so literally.
#1. I don’t think what I am talking about is “magic”, although in our healthcare system it evidently qualifies .. .
Saying that we should screen people with a deadly infectious disease using electronic outreach is not Stephen Spielberg, it’s basic communication.
And makes sense given the variables and risks we’re facing …
You say that’s a privacy risk? Delete the information after you’ve screened them. Create SnapChat for medical information.
I don’t care what you do.
#2. Saying that EHRs need to evolve into content companies instead of the click based billing and coding data silos they have become is not a great leap into the unknown, it is an obvious statement.
A simple web-based customer relationship mangement tool for doctors and caregivers built on the Highrise / Salesforce model would do more to answer the complaints raised by doctors and nurses than 95% of the solutions that are coming along ..
Will the current players in the market embrace this reality?
Perhaps. Perhaps not.
But it needs to be pointed out.
It’s the permanent storage that’s the problem isn’t it? The cure is the poison.
When they start designing EHRs for use by physicians and not for administrators they will be much better and will start to do the wonderful things for which they have potential. However, blaming an EHR for missing Duncan is nonsense. That’s like saying the EHR was at fault because the fact that the patient had no pulse or BP was written down in the wrong place. Some things require direct communication and immediate action. That was one and they just blew it.
Steve
See the works of Jerome Carter, MD, at EHR Science. http://ehrscience.com/
Best analogy of the thread, hands-down. The airplane one. I wonder what proportion of airline pilots bristle at the notion of being “machine minders?”
Actually, I think I am describing a Wufoo form.
These are steps that are essential for voluntary wide acceptance of EHR’s. Thank you for clearly presenting them.
The only reason I can see that others are pushing top down EHR’s without these steps is that they are only able to lead by using force.
You are describing surveillance on an epic scale 🙂
People (physicians and patients both) would need to have tremendous trust in the governance of our technology to allow this kind of surveillance. Gaining this kind of trust requires full transparency and a separation of public health from commerce. That would start with open source EHRs and legal protection from mining of public health data for commercial or law enforcement uses.
It is interesting to me that almost all of the physicians I know who have abandoned traditional payment systems (i.e. cash only or direct pay practices) continue to use EHRs. But the EHR apologists like to portray their criticisms as evidence of an anti-technology bias or anti-progress. It seems to be in their best interest to continue to frame the debate as EHR vs Paper when no one cares about that comparison anymore. The debate should be about a complete redesign of the EHR around a patient centric workflow.
John, I entirely agree with you.
My beef is with those who think this would have prevented the index case – i.e. prevented TD. Not those who say it will stop another TD.
Once the index case has occurred, it makes sense to use a screening heuristic, and not only use it but assign the burden of proof against its positive result.
A simple heuristic would be: symptoms + travel and contact. If positive on 2/3, then it’s Ebola unless proven otherwise.
EHRs cause 40-50 million people to be harmed every year?
That’s quite a statement.
Can you break that down for us?
And you can define “harm”?
I’m quite prepared to believe that 40-50 million people’s lives are impacted by them in some meaningful way that may be bad
Throw in the number of doctors and nurses traumatized by them and I’m sure the number is a high one …
But maybe not 30-40 million, those are World War II numbers
We have an early warning system. His name is Tom Frieden. I suggest we listen to him.
The screening idea is an idea, not Monday morning quarterbacking, although it is Monday morning, isn’t it?
It seems like a good idea to start thinking about what we can do to catch things earlier. If you can do this remotely, great. Substitute have you been to Cleveland? or were you on Flight 44 from Minneapolis? for have you been been to Liberia and you see my point …
If you were the person doing intake at a crowded ER this weekend, I’m sure you’d think this was such a bad idea
Screwdrivers and chainsaws are the choice of the craftsman. Today’s EHRs are more like the plumbing, or maybe an airplane – a piece of infrastructure that’s not a professional or a craftsman’s choice.
Plumbing, and especially airplanes, are highly regulated and the professionals that manage them are highly constrained. It would be a fair debate to have if we want our clinical technology highly regulated by the FDA and our physicians relegated to the role of watching over the machine. Building clinical decision support into institutional EHRs is taking us down that path, minus the debate. Unfortunately, taking technology choice out of the hands, and the responsibility, of the licensed physician is now being done for commerce rather than health.
John, I agree completely. And I agree with (almost) everyone in this thread. What I am hearing that I agree with is something like:
“EHRs are not the problem. EHRs as most are currently constructed and are being rolled out across the industry are massively the problem in their lack of decision support and their lack of usability. This becomes actual decision hindrance, by slowing the clinician, by hiding the most relevant information in acres of dross, and by laying its own information-gathering pathways on top of and in front of whatever decision-making pathways the clinician follows, instead of paralleling and assisting those pathways.
Under the bus is the proper storage option for most current EHRs.
Blaming the EHR for not catching the disease is like blaming a screwdriver for not being a chainsaw. The EHR is a tool, programmed by humans and fed information by humans, so its GIGO. EPIC is built along the lines of being patient-centric, so getting it to perform population-centric tricks is like mating elephants — it can only be done at a high level with a lot of noise and effort, then waiting 2 years for the result.
Many have suggested that having an early warning system could have diagnosed TD more promptly.
When would the early warning system have been put in place? During the first case of Ebola in Liberia? Before the outbreak?
Imagine the news headlines.
“Texas implements early warning system for Ebola. The rest of the world braces for the index case.”
So when exactly would this have been instituted?
When TD boarded his flight?
“Liberia to mission control: Houston we will have a problem. Our first Ebola patient is on board.”
When?
Anyone can be a genius in hindsight.
a) Guns don’t kill, people kill.
b) EHRs don’t miss diagnosis doctors miss diagnosis.
I find it useful to line similar statements from opposing political stances together. It creates cognitive dissonance (if there is a shred of intellectual honesty).
If you support (a) you should support (b). If you are against (a) (presumably from the anti-gun brigade, and as has been recently described ,a granola cruncher) you should be against (b).
You can’t just decide when you want to that inanimate objects are culpable and humans lack agency, and then the opposite when it doesn’t suit one’s world view.
I think physicians are responsible for obtaining relevant history and speaking to the patients. However, EHRs haven’t helped matters, which is a damning indictment for so expensive an investment.
So yes, guns don’t kill, people kill but guns help killing.
So yes, doctors miss diagnosis, not EHRs, but EHRs help missing diagnosis.
Dr. Peel,
You stated:
“Today millions of patients in the US act every year in ways that put their health and lives in danger because they don’t trust EHRs that their physicians use. 37.5 million hide information annually and 5-10 million delay or avoid treatment for cancer, mental illness, or STDs every year because they know the data are not private.”
I agree completely. I advise friends and family to carefully consider what they tell their physician unless the physician gives some assurance that they will keep info our of the electronic record. But this issue seems to be widely ignored by the media.
In theory it is possible to design electronic record systems that give patients excellent control over what gets shared and when…..many who post on this site pay lip service to this goal….but I suspect this will not become a design priority until the issue gets highlighted in the media and by patients/consumers….(and their docs)….demand it.
In the meantime, it is still wise to be selective in what you tell your doctor unless you have a serious medical issue that mandates your privacy concerns are secondary…..and if you don’t have a serious condition, rely on your spouse or other confidant to bring the medical folks up to speed in the event you develop a serious condition.
Finally, you mention some statistics of how many withhold information from their docs….it would be appreciated if you could provide sources for that information. Thanks.
Proof of efficacy is one of the first thing on the lips of those that support EHR’s yet when it comes to EHR’s we see no proof required.
Was the EHR a proximate cause in Duncan’s missed diagnosis? We will never know, but then we can’t be sure a person driving an automobile while texting is a proximate cause of the accident. Empirically though, we can believe both to be proximate causes.
Moreover, who is the EHR assisting? Auditors or the physicians at the bedside? We know that answer but supporters of the EHR choose to place that answer under the rug.
John I. asks: “Why not screen patients before they reach the front lines?” Duncan, I understand, was ‘screened’. Didn’t he call the CDC and wasn’t he told to go to a hospital? It appears Duncan recognized that he could have Ebola. Did the CDC call the hospital and notify the front lines of a potential threat? Did the CDC make sure the hospital was prepared? That is known as a top down failure.
This was a failure from the top down and now many are telling us that we should rely upon systems that are top down where a priority is the audit process interfering with the safety and treatment process. I think a lot of people have things backwards.
The EHR desrerves to be vetted for sfaety, efficacy and usability. If that means throwing it under the bus, so be it.
The EHR facilitated the double whammy: failure to suspect the diagnosis, and the failure to admit a critically ill adult with acute renal failure, thrombocytopenia, and a fever of 103 F after 1 gram of Tylenol was administered.
EHRs have been known to cause errors, even by the most competent of physicians. I do not believe for 1 second that the ER doctors at Presby were that stupid so as to not make the right decision based on the abnormalities, regardless of travel history.
The fact that a note said “no fever or chills” suggests that the doctors did not see all of the data which was silently submitted into the EHR without warning or notification, ie the silent silo syndrome. This is a known defect that has repeatedly caused delays in care and errors in treatment.
Thus, it becomes a no brainer that these devices are dangerous and must be subject ot he F D and C Act.
Dr. Peel is spot on.