Slashdot Mirror
Why Digital Medical Records Are No Panacea
theodp writes "As GE, Google, Intel, IBM, Microsoft and others pile into the business of computerized medical files in a stimulus-fueled frenzy, BusinessWeek reminds us that electronic health records have a dubious history. Under the federal stimulus program, hospitals can get several million dollars apiece for tech purchases over the next five years, and individual doctors can receive up to $44,000. There's also a stick: The feds will cut Medicare reimbursement for hospitals and practices that don't go electronic by 2015. But does the high cost and questionable quality of products currently on the market explain why barely 1 in 50 hospitals have a comprehensive electronic records system, and why only 17% of physicians use any type of electronic records? Joe Bugajski's chilling The Data Model That Nearly Killed Me suggests that may be the case."
Digital Medical Records Are No Panacea... but they are pancreas!
Be relentless!
Major credit card companies either can't or won;t take the necessary precautions to protect credit card information. So what if there is a breach, identify theft, headaches, etc?
Now what makes you think hospitals, private doctors, etc. are going to be able to protect their data any better? They have less money then the credit card companies.
Can you imagine a million patient digital medical record breach? The black mail or power that could be leveraged over people?
When my wife was in the hospital with a broken ankle I tried to get a copy of the X-ray, because it was on a big monitor out of view of the patient. The user interface of the DICOM viewer did not provide a way to print or save the image... presumably to protect patient confidentiality.
The next day I went in to the hospital to pick up the "films" for her doctor, and they gave me a copy of the same files on a CD, completely uncontrolled, and I used OsiriX to convert them from DICOM to JPEG so my wife could see them.
Having the files in digital format is great, but let's have some appropriate level of controls. If the patient wants the images on a flash stick, it's THEIR records, let them have it!
This article reads like a lifetime made for TV movie. Heavy on emotion devoid of logic.
The author was repeated asked for his medical information, his doctor's written instructions were ignored and different departments within the hospital did not communicate. Therefore the problem is Obama's computerized data record system that doesn't exist yet.
The whole time I was reading it I was waiting for the author to tie his experience to how computerized medical records are bad. He never did, his experiences were caused by humans that did not care enough about patients to read computerized records OR paper records.
The author fails to explain how his experience proves anything other than that particular hospital is terrible and that the health professionals employed there are less than friendly.
... and here we have just a single anecdote about how the system did not work in one instance. If we are playing the anecdote game, I'm sure I can find a similar example where non-computerized health records lead to bad care. Of course, while the anecdote game is very effective at playing at human emotional response (we tend to assign more weight to a story that we can associate with a single person versus aggregate statistics), it's useless as an actual policy question.
Since every complicated system has failures, even the critical ones like hospitals and air traffic control, the important policy question is not whether it works in all instances, it's whether it produces overall better care than the system it's replacing and whether that improvement is worth the difference in price. If the new system actually reduces costs, then it's a good idea so long as it doesn't degrade care (since, ultimately, reduced cost means either more health care or more dollars to satisfy other wants).
I'm not going to comment on the data myself, since you should read the studies for yourself and draw your own conclusions.
http://journals.cambridge.org/action/displayAbstract;jsessionid=7C274D08947B0625B3B540BEF2E70367.tomcat1?fromPage=online&aid=416400
http://content.nejm.org/cgi/content/abstract/348/22/2218
(PDF)
http://www.pubmedcentral.nih.gov/articlerender.fcgi?tool=pmcentrez&artid=1421388
PS. Of course there's no panacea for our medical problem. The question is whether EHR are better than the system we've got, not whether they represent the best possible system. The perfect is not the enemy of the good.
PPS. I have a sneaking suspicion, reading my post (yeah, some /.ers actually read their own posts before hitting submit :-P) that I will be accused of not having the proper sympathy for the guy in TFA. That's not true. I have sympathy for him as an individual, but I'm not going to let that sympathy for him cloud my judgment on the merits of a system.
For example, suppose there was a highway by you that had no center divider, just a grassy median. Suppose also, for the sake of argument, that installing a jersey barrier (http://en.wikipedia.org/wiki/Jersey_barrier will lower the injury/fatality rate in accidents by a statistically significant amount by preventing out-of-control cars from going into oncoming traffic. Now, hypothetically, someone could be in an accident where the jersey barrier caused him serious injury or death (say, by flipping his car even though they are designed to minimize that chance) where the old system would have been just fine (say, because there was no oncoming traffic at the time of the accident). Does someone that still says we have jersey barriers not have sympathy for that guy? No. His death is regrettable but because we can't make a perfect road, we have to settle for the best road we can make.
The problem is that you can point to someone that's injured (and provoke an emotional response related to his regrettable accident) but the only thing the jersey barrier proponent can do is point to the statistics that say there are fewer serious injuries since they've been installed. There's no emotional resonance to the thousands of people that travel without incident each day because they don't make a good story. "Man drives to work safely" isn't news, but because it happens much more often that "Man killed in car wreck", it's actually much more important in the grand scheme of things.
We aren't privy to all the stories where EHR made things smoother, cheaper or helped prevent calamity. Largely, these will be small victories, unsung
Like all software, digital medical records can be done badly. But they can also be done right. Joe Bugajski's story is gripping, but I want to compare it with the story of my mother.
My mom was in her mid-50s when she became ill, apparently healthy but in fact hiding a serious alcoholism problem. I'll skip the details, but suffice to say that a lifetime of drinking can destroy your body's natural blood-clotting system, leading to internal bleeding. So don't drink, kiddies.
Anyway, once she was medevaced to Queen's Hospital in Honolulu, we never saw a single obvious piece of paper. Everything was recorded digitally. But the key difference between my Mom's story and Joe Bugajski's is that the data was *available* once entered. I got a chance to look over the doctor's shoulder as he reviewed her chart. He was able to look at blood tests, x-rays, up-to-the-minute vitals, every piece of data the hospital recorded, at his fingertips in seconds. And he drove the software like a pro.
In the end, my mother died, but it definitely wasn't because of bad recordkeeping software.
I RTFA, and there is a very telling reader comment at the end...
The moral of the story, then, is that no amount of even well-organized information can compensate for a break in the continuity of care. The allergist tossed this guy to the wolves with a post-it note stuck to his forehead. The current system couldn't cope with that, and it's hard to imagine any system that could, because the hospital et. al. can't morally or legally just follow the instructions on the post-it note; they have to start from scratch.
The allergist had to know this, but dropped the ball anyway. Find a new allergist.
FATMOUSE + YOU = FATMOUSE
My wife works at a hospital with digital records, and it seems to work fairly well - no worse than paper charts anyway.
The major issue that I have is that they use it only to a fraction of its potential. They use it just like they did charts, with no real capability increase other than stretchability and speed.
For instance, they could use it to prevent some medical mistakes by requiring an override if a physician changes an order. Right now one doctor (or even a nurse) can simply walk over and change the order given by another doctor. At the very least, another doctor who is on call should okay the change so there are at least two eyes on it.
Another example is medications. A groggy doctor woken up at 4AM can and will make mistakes, sometimes as severe as mixing mg and micrograms. You can bet that a dosage 1000 times higher than indicated will not be good for a patient, and currently they rely on the pharmacist to catch these errors. The computer could be programmed to require an override by a second doctor before allowing such orders.
Also, due to lawsuits, everything at the hospital is a CYA system, and patient care suffers. Computers could be used to help this situation, too - but I'm getting carried away now :)
W..w..W - Willy Waterloo washes Warren Wiggins who is washing Waldo Woo.
I'm a resident physician, and so I've used various EMRs in different hospital and clinic settings, and they pretty much all suck in different ways. EPIC, which is based in Internet Explorer of all things, is the worst, but seems to the the one that's being adopted at the most hospitals.
The UI design is just horrible, but beyond that I had a hard time putting my concerns into words until I read an article somewhere that talked about something called "cognitive support to the physician." That is what most EMRs lack.
As a physician, I want an EMR that lets me rapidly get at important clinical information and give me targeted alerts that I need to make a decision. Instead, the systems are centered around billing and cover-your-ass medicolegal documentation. In the paper chart word, these issues had already diluted the meaningfulness of the chart. (Ever see a hospital chart - maybe 10-20% of it has meaningful clinical data in it, the rest is full of useless legal/billing/redundant crap.) Many EMRs just translate the same troubled paper chart system into electronic format, but then the ease of electronic data entry means that even more useless information is included/required, making it that much harder to find the info you really need to make a clinical decision.
I have to say that the best EMR I have used is still good ol' CRPS at the VA. It's not as slick looking as the newer ones, but the data is easily accessible and I have never had to waste my time looking up a billing code. It's been chugging along for over a decade, sharing data between hundreds of sites across the country. (And the issue in the first article about the EMR causing more deaths because you can't put in orders while the patient is en route - not an issue in CPRS, we do this all the time at our VA.)
My understanding is that the code for CPRS is open and free to anyone who wants it. I would gladly choose CRPS over the ability to type my notes with colored fonts in EPIC. They were considering adapting it for the large county hospital system where I work now, but in the end went with EPIC because... wait for it... it was easier for billing.
Good points.
Any system can only be as good as the people that use it. I can't help but feel while reading 'The Data Model That Nearly Killed Me' that the problems encountered actually had very little to do with the electronic record system at all. It seemed more like an incompetent system was in place as a whole. The data model didn't seem to do anything wrong, it was the people using it, or not using it. Not saying whether it is actually a good electronic system or not, impossible to tell...but enough people had enough direct access to critical information, without even thinking about the electronic system, that this guy should not have had the problems he had.
Is it really the data model's fault that not only did no one use information provided on entry to the er, they didn't even READ it? Sounds to me like the real problem is that new systems were put in place without new processes or training being put in place...and then on top of that the users of the system failed to even fall back on the logical concept of direct communication!
I do not for one second believe that this situation wouldn't (Or for that matter hasn't) have happened even with the use of standard physical medical charts instead of the electronic record system in place. There is really nothing at all in the story that makes the problem specific to the system or the model being used in that system. Can't believe that had a physical medical chart been used that the same mistakes the medical staff made in this case would have somehow miraculously NOT been made on paper as well.
Basically, what I take as most important from this guy's story, is that that is NOT a medical facility I ever want to step foot into under any circumstances, electronic records or not!
No Comment.
I would go even a step further than that and posit that a good portion of his problem was stemming not from the system as much it came from the active resistance of the people attending him in using the system.
I don't directly work in healthcare, but I do work in a corporate environment for a large healthcare company that recently (in the past decade) made the switch from paper to a 'global' electronic system. At the start, stories like this were common, as people fought the system rather than use it.
Yes, not all systems are equal and it's entirely possible to design and implement an completely unusable one. But there is no avenue for improvement when the default behavior to burrs in the system is to revert to a far more inefficient (and porous) paper method, which, due to the introduction of the electronic system, is not even being monitored as well as it was when it was the only method.
In the end, the improvements that were introduced and enabled by converting to an electronic system far out weighed any of the temporary and transient issues such as this.
The administration either has an undisclosed agenda or no idea what is really wrong with the health care industry. I work for a large medical institution in their IS department and I spend most of my time moving medical data around. In the short time I've been here, I have run across several roadblocks to providing efficient, safe and effective medical treatment.
The most detrimental entity in all of health care has to be the private health insurance industry. Insurance companies have spent a great deal of time and money developing strategies to MAKE MONEY. They are not in the business of making people well, they are constructed to make profits and protect those profits at all costs. They have nearly perfected the art of delaying or denying treatment for sick people all in the name of the almighty dollar.
The lack of standards is truly astonishing as well. There are dozens of large companies vying for stimulus money to develop electronic medical records. Do you really think they'll be working together to provide a single solution that can be transported all over the country? These companies are also out to make a buck and it better serves their interests to develop the one standard format and be the holders of the golden goose than to work collaboratively on a solution that fits all (or most) needs. See: Blue Ray vs. HD-DVD or VHS vs. Beta-max. I would estimate that 9/10s of the stimulus money directed to these companies will be an utter waste, and the remaining 10th will got to produce fortune for a single organization.
Whenever a format *is* declared the winner, it will likely be so inadequate that it will be routinely altered and hacked to fit the specific needs of each institution. It will be rendered nearly useless. HL7 is great example of this. It's designed as the de facto format for transmitting health care information from one site to another, however, I have yet to see two institutions or vendors do it alike.
Pricing and billing are two other concerns. Both are seemingly completely arbitrary and vary widely from one facility and/or patient to the next. A simple lab procedure, let's say a white blood cell count (literally counting white blood cells), could be done in one location for X while in another location for 6X. The worst part, you have no way of knowing what that charge will be until you are billed. Then, if you have insurance, they get to choose whether to pay all, part or none of the bill based on what loopholes are available to them.
My personal opinion, I represent no one other than myself, is that the single most effective action that any government can do to help solve the health care problems is to do away with privatized health insurance as we Americans know it today and replace it with a system that is much more socially responsible. A standardized digital medical record will be a good thing, but it will likely show very little impact on patient care.
[T]he VA is run entirely by the government. What the rest of the US is going to wind up with is a huge train wreck of competing standards and products by proprietary vendors who don't want to interoperate.
Once again it's probably worthwhile to note that this was a major part of the motivation behind the original ARPAnet project which grew into the Internet. The US Dept of Defense was trying to deal with a growing problem. They were collecting all sorts of fancy electronic gadgets that generated and consumed data, but most of them would only talk to other gadgets from the same vendor. It was clear that this wasn't an accident. Every vendor wanted a to be the sole supplier, and one way they all saw to do this was via proprietary data formats.
The ARPA gang's solution was to build what they called Interface Message Processors (IMPs), whose job was to talk to a proprietary gadget in its native language, translate the gadget's messages into a standard format, and transmit that to another IMP, which would translate it into the native language of another recipient gadget. They knew from long experience that their vendors wouldn't cooperate with this, and would do everything in their power to sabotage the ability of other vendors' gadgets with their own. So the ARPA people farmed out the task of building the IMPs to people who had a history of successful communication with their competitors, the people in academia.
That was about 40 years ago. Now, with four more decades of experience, we can clearly see that the problem hasn't gone away. There is no prospect that gadgets or data systems built by different corporations will ever interoperate sanely. Private companies have a strong motive to sabotage such communication whenever they can get away with it. So, as in the past, the only way we can get useful medical data systems is the same was we've done it with the Internet. We need government-run projects to develop and enforce the standards. Building the low-level gadgets can be a job for the corporate world. But if we ever want to be able to use the data for any meaningful purpose, we must make sure that the corporate world can't control it.
Actually, of course, we have no guarantee that government agencies will do the job right, either. There's no shortage of incompatible data formats in government databases. Unless the job is handled by people as competent as ARPA was back in the 1960s and 70s, it'll still be a huge, expensive failure. Sorta like the medical data systems we have now, which were mostly developed in-house at hospitals, and even the nonprofit hospitals have a poor record of interoperability. (Yes, I've worked on some of their systems, and it's not a pretty sight.) So we should be watching how the governments deal with the problem, and be quick to criticise the crappy standards that we know they'll design.
Otherwise we'll end up with medical records based on a standard similar to the Avian Carrier Protocol, but it won't have been published on April 1. You should also read the wikipedia article to read of a real implementation. But most managers in both corporate and government circles don't have a sense of humor good enough to prevent such things from becoming actual standards.
Those who do study history are doomed to stand helplessly by while everyone else repeats it.
Better to get rid of the lawyers first.
If the medical journals say that there is a 0.0001% chance of deadly condition Y being present given the patients symptoms, and a $5k MRI test has a 0.001% chance of detecting Y, then the doctor is going to have to order it. Otherwise when the 1-in-1-million patient dies from undiagnosed Y the jury will be handing money to the plaintiff hand over fist.
The expectation of modern juries is that every patient gets tested with every modern technology available, has access to experimental technologies that are just emerging, and has a board of doctors meeting in a conference room with House to discuss every aspect of the patient's care.
I was a medic in the USAF during Viet Nam. I had a strong technical background, so I worked on a medical records database project from 1975-77 at the Air Force Rocket Propulsion Lab in the Mojave desert.
We hand coded, on punch cards, for a Control Data host, about 650 records. Took 6 months.
I thought at the time, "there's got to be a better way!"
In the late '80's, I was CEO of a medical software company that created a networked medical transcription application integrated to "ChartChecker", an expert system for ER physicians, that would analyze a patient record and tell the doctor if he had passed or failed the encounter and was therefore at risk of malpractice litigation. We got the chart through the network from the transcriptionist to the analysis engine and had a result in 30-40 SECONDS. With voice-to-text, we actually did near realtime analysis.
Massachusetts approved a statewide 25% malpractice premium reduction for any ER doctor that leased our system. At the time, the minimum annual premium was around $30,000 and our system leased for $5,000. The average ER doctor stood to net $2,500 a year and that doesn't factor in the reduced chance of litigation.
This was 20 years ago. We spent a LOT of time with the VA, BIA, DoD, CHAMPUS, the Navy and Air Force. I saw a WORKING digital dogtag in 1991.
And where have we gone in 2 decades?
Not far. Not far enough by ANY yardstick.
We have sufficient technology; what we need is a national standard medical record that is mandatory for all who deliver medical services in the U.S.
This is a problem that should have been solved 20-30 years ago.
I am my own gestalt.