« What's the social value of Microsoft? | Main | The Spirit of Radio »
Why doesn't America have electronic medical records?
Ezra Klein poses the question:
I've never read a compelling explanation of why the nation's doctors and hospitals haven't broadly adopted electronic medical records. It's not as if they're allergic to technology. At this point, cardiovascular care employs every strategy but astral projection to keep our in rhythm. It's not as if it wouldn't be cheaper and easier for them. The man hours and costs from keeping track of files, printing out labels, finding lost manila folders, and getting sued because the nurse misread the doctor's handwriting are enormous. Theoretically, insurers should be pushing on this, but they seem behind the curve, too. And it's not as if there aren't tested programs in use -- not only does Europe do electronic records well, but the VA does them beautifully, and they've released their primary program, ViSTA, as open source, for free use by anybody.
I can think of four reasons.
1. Most of the benefits are reaped by the patient, and in the long run. Today's suppliers don't realize these benefits in the form of profits.
2. The United States has relatively weak data protection laws. Many people don't want outsiders to know their medical history, and information compilers fear lawsuits if the information leaks out or is hacked.
3. No single provider has an incentive to move first in this game. Why computerize if no one else has?
4. I haven't computerized my office (is Alex laughing?), I worry more about surviving until the next day.
The comments over at Ezra's are excellent. And if you think that electronic records are the source of vast productivity gains, just have Medicare mandate such a change. Readers?
Addendum: Here is Arnold Kling.
Posted by Tyler Cowen on June 27, 2007 at 07:10 AM in Medicine | Permalink
Comments
It's a public good. A lot of people would marginally gain from the public records, but no individual actor would benefit enough to fork over the cash for the IT needed for such a move. Clearly, either the hospital would need to see the cost-benefit advantage of such a move, or the health plans (including Medicare) would need to provide an incentive that would make it advantageous to the hospitals.
But as you alluded to a CMS mandate, this does seem like a situation where the government could intervene with a sharply directed mandate to provide a shortcut passed the market's slow-moving shift to EMRs.
Posted by: Christopher at Jun 27, 2007 7:30:29 AM
Some hospitals have switched over to electronic records, but like all changes in the medical field, a global switch is likely to be far slower than rationality might predict. Duke's medical system is almost entirely paperless; the military VA system (or at least some of it) has been paperless for longer. There are advantages and disadvantages to each system. My memory of the VA system is that it doesn't consistently save the medical end user time, but you win some and you lose some.
Posted by: at Jun 27, 2007 8:00:45 AM
Finally, something I actually know something about.
1. The United States has relatively weak data protection laws With the advent of HIPAA, that is no longer true.
2. Data capture is a significant problem. Input devices are slow, clumsy, and/or expensive. People still prefer pen and paper.
3. Interoperability standards are still evolving. This is a significant problem in compiling a unified medical record for a single patient across many providers.
4. As has been hinted, the biggest driver will be regulatory compliance - often at the state level, not necessarily the federal level.
Posted by: Jeff Burton at Jun 27, 2007 8:14:46 AM
Here is another reason. Do you want to hook your entire operation to Microsoft when Microsoft can change the operating system in six months and make all of the computerized medical system worthless.
Also, unlike other business, medical records have to be maintained. paper is relatively cheap to store and is technology independent. A computerized system forces you to pull all of your old records forward ever time the standards change. Do you want to bet your business that PDF will still be around in ten years?
Posted by: superdestroyer at Jun 27, 2007 9:10:13 AM
That Europse does electronic records well is a bit of an overstatement. Here in Denmark the struggle to make electronic records is way over budget and is still long in the making and I believe (via www.adamsmith.org) that a similar thing is going on in England.
Posted by: Peter at Jun 27, 2007 9:21:49 AM
Coincidentally, my wife recently did a study of this exact question. What she discovered was that there were many competing systems and each required large fixed startup costs both in terms of monetary outlay and in terms of human capital. Many specialist fields wanted the software to be modified to meet their special needs but could make it less compatible with other systems. Most small practices were reluctant to make such a large investment until they were reasonably sure that their records would be generally compatible with everyone else’s system. These things just take time.
Posted by: Michael H. at Jun 27, 2007 9:37:02 AM
I always thought that a lot of the problem was the difficulty of getting everybody to use compatible systems. The VA can make it work because they can dictate to every one of their facilities what software to use. My company has toyed with the idea of giving clinicians laptops and using electronic charting, but we would still be turning our digital record into a fax, so that the facilities where we see patients could have a paper copy. In a large university hospital I have seen at least three electronic records systems in use, none of which could communicate with the others.
Posted by: Gong Tao at Jun 27, 2007 9:42:02 AM
I'm diabetic and continue to be amazed that my HMO doctor is looking at a snapshot of my various tests, weight, BG , etc on each office visit. there is simply no longitudinal information available. Keeping a patient's vitals in a computer makes so much sense that I think it is malpractice NOT to keep them.
Regards
Posted by: John Johns at Jun 27, 2007 9:43:45 AM
It doesn't seem like the "doctors use the latest technology, so why not electronic records" is a good argument, since learning the latest procedures and such gives them a competitive advantage, whereas e-medical records mostly serve to give them no such benefit, at a large fixed cost plus lowered productivity in the middle of a learning curve which will probably never end. There's an article on NextGen about an anecdotal issue that plagues older doctors with regards to digital records.
Posted by: such.ire at Jun 27, 2007 10:00:33 AM
I recently had this conversation with my father, who is actually undergoing the transfer from paper to electronic soon (provided the conversion doesn't flame out like several previous attempts have done).
The problems are bigger than you might think:
1) The most crucial data for him (family practice physician) is the patient's description of his problem, in words as close to the patient's own description as possible. Filling out checklists of symptoms, etc is a very good way for a doctor's own view of what the illness might be to override the patient's own description, and often results in expensive and unnecessary tests as the doctor examines blind alleys. Current electronic systems make it inconvenient and awkward to put this information into the system. If it were in the system, there would be very little benefit from it being in electronic form, since the original words matter so much.
2) The costs of paper records are actually very low. He has 20 years of records in a single room, filed alphabetically. It's easy for a nurse to find, not susceptible to hardware crashes, software obsolescence, data theft, etc. The information can be mailed or faxed pretty easily. Putting new information in the file is as simple as inserting a new piece of paper, even if the information is in a very different format from 20 years ago.
3) He has very little idea what benefits might come from having the same information in an electronic format. There are no database queries he wants made, no data mining or statistical questions he finds necessary. He interacts with the information better when it's on paper.
Posted by: Zach at Jun 27, 2007 10:11:18 AM
I work in the IT and medical industry. I've worked on EMR systems in the past and have some knowledge of how it works, how it's sold, etc.
There are a lot of problems that arise when developing software for the medical field. When you develop software, you are taking business rules and processes and translating them into a pattern that can be interpreted digitally. The problem with the medical field is that (due to government regulation) these rules are constantly in a state of flux. That's federal rules: state laws and regulations add another layer of complexity.
Creating a simple patient charting system is easy. It's the details that cause all of the problems (billing, scheduling, etc, etc).
The medical industry is so complicated (again, due to regulations) that software engineers that are willing to invest time and money in development have difficulty connecting with individuals who are experts in the field and knowledgable enough to translate the rules into something that a normal person can understand. Most of these 'experts' are already highly paid consultants and speakers. One individual may be an expert in Medicare billing rules for wheelchairs, while another individual may be an expert in enteral nutrition. Each sub-industry has it's own quirks and oddities: bringing it all together is a monumental undertaking. What may be true today may no longer be true tomorrow if a beaurocrat changes his mind.
That helps to explain the private sector: but what about the government?
Expecting the government to mandate some sort of nation-wide software will only end up in further inefficiencies. I've seen government software projects (in particular was the software used to calculate loan qualifications in New Orleans after Hurricane Katrina hit). They aren't pretty, let me tell you. It is extremely rare that a well-suited, talented company receives a government contract to write software. There's so much cronyism in that industry that the mere thought of a government-mandated centeral EMR database makes my tear my hair in despair.
This just sums up some of the issues that I have personally seen in the industry. I could go on for hours on this subject.
Posted by: Brian Dailey at Jun 27, 2007 10:16:15 AM
Zach's comments are largely on point, which is why the individual doctor lacks the motivation and why providers in general lack the motivation. One of the great advantages of having patient info store electronically is that it makes him much more mobile. Sure, the doctor might lke to have a new patient's info right away, but either running a few basic tests, asking a couple questions, or requesting them faxed doesn't bother HIM that much. However, I do think it has a very negative impact on patient mobility.
(This is beyond the immense gains to be made by increased opportunities to track patient progress and outcomes to assess patient care)
Posted by: Christopher at Jun 27, 2007 10:16:26 AM
Patient mobility definitely goes on the other side of the equation, and ability to electronically transfer files in the case of an emergency arising on vacation, etc, would be a definite benefit. The cost/benefit equation probably also changes when you look at the truly massive recordkeeping that a large hospital or insurance company does.
Like most situations where person A derives a benefit from person B incurring a large hassle, I would expect the progress of electronic recordkeeping to be slow and gradual.
Posted by: Zach at Jun 27, 2007 11:03:23 AM
It would seem that a decent compromise in the short term would be some sort of hybrid scanning system where doctors would still benefit from the flexibility and ease-of-use of a paper-based system and there would be a low buy-in cost. Perhaps even no buy-in cost for doctors themselves if the govt. paid for these inexpensive machines.
After a checkup, the admin asst could just send those papers through some kind of fax-like scanner to create pdf versions of whatever was notated and save them under the patients SS# in an internal database. In the longer term, we can work on integrating these into a larger national database and slowly move toward a completely electronic system.
This is useful if for no other reason than fires, floods, or similar circumstances in which paper-based records are in jeopardy.
Just a thought.
Posted by: fustercluck at Jun 27, 2007 11:03:46 AM
Finally a topic I know about. Most doctors are small business operators and EMR represents a big capital expense and ongoing cost as well. Since payments are fixed by insurance and Medicare, EMR represent an actual monetary loss to the doctor with no way to pay for the cost, even in the long term. I don't think EMR increase productivity since most established doctors are already as busy as they can be.A lot of the patient care notes cannot be done by EMR and must be hand written or dictated/typed into the patient chart. There is no OS or software standard across the country that all have agreed upon for EMR, so communicating between sites becomes a problem. The benefit for EMR is eventually the patient and the insurance companies, for different reasons, and the doctor loses money with EMR.
Posted by: DrTom at Jun 27, 2007 11:14:21 AM
Acceptance is easy - if you want to accept Federal payments, you adopt the VA software. While paper is easy on the doctor, I bet the billing dept would love electronic records.
Other benefits would include warnings on drug interactions as they have done in private pharmacies for 20 years, procedural warnings & suggestions.
One poster touched on a data mining. Imagine studies that could be conducted in hours instead of years - using ten thousand subjects instead of hundreds. Or person specific data mining that would clue a doctor to another diagnosis.
Many reasons to proceed.
Posted by: Tom at Jun 27, 2007 11:36:43 AM
My health care provider/insurer is fully electronic. Every room has a computer that’s used to pull my file, write notes, access test results, etc. Everything always seems to go smoothly and longest I’ve ever spent in the waiting room is 5 minutes. I can access my own personal medical records online if I wish. The nurse/doctor types in my symptoms to the computer, and gets back a list of potential causes. Ultimately the Doctor makes the diagnosis, but the computer helps make sure he doesn't miss an important possibility. Further, my doctor (or a nurse) will reply to questions by email… usually within a few hours (which often saves a visit).
My wife gave birth at this health provider’s facility and it was a wonderful experience. We stayed in the same room the entire time (30 hour labor!). We had a private room complete with bathroom, birthing tub and a chair that converted into a bed for me to sleep in. The room was well stocked with supplies, which meant time was not wasted running back and forth to the supply closet. Since there was a computer in our room, no need for nurses to run back and forth looking up information. Our baby was examined in the same room he was born in (he never left our sight). We dealt with a single nurse during each 12 hour shift. When shifts were changed, the nurses handed off our case while in the room with us. That way, if the departing nurse forgot to explain a detail, I would know. Of course, all the details were right there on the computer next to the bed anyway. At departure, there was no need to talk to a receptionist, because the nurse was able to check us out right from our room.
My provider is the low-cost managed plan offered by my company’s health benefits program. I’m a mid-level manager at an investor-owned utility, so it’s not as though I’m on some fancy-pants executive plan.
Maybe I’ve just been lucky, but I’ve never had any problems with the many insurance providers and health care providers I’ve used during my lifetime. In my experience, Managed care is great. It’s a real shame that it was so horribly demonized in the 90’s.
Posted by: Whit Stevens at Jun 27, 2007 12:00:00 PM
Various kinds of EHR (not necessarily EMR's as they are technically understood within the field) are NOT prohibitively expensive for midsized practices (they probably are too expensive for practices of fewer than, say, five doctors). You just have to look past the dozen or so largest companies that have name recognition because they provide enterprise-level solutions.
And the idea that keeping paper charts is cheaper than electronic ones is laughable -- the sheer rent costs for enough storage for that much paper are enormous. And that doesn't even begin to get into the typical workflow of a paper chart during a routine (nothing wrong!) checkup: Retrieve the chart from the closet, move it to the exam room with the doctor, move it back to the receptionist's desk, refile it in the closet, retrieve it from the closet again when the progress notes are transcribed, add the progress notes, refile it in the closet, retrieve it again from the closet when some lab-work comes back, add the lab results, refile it in the closet. In a well-designed medical record system, every step above is obviated. In a poorly designed medical record system, all but "retrieve and put in the exam room" are obviated.
(Oh, and god forfend you're talking about a large multi-specialty clinic where the paper charts aren't all stored in a closet that's just a few feet from the exam rooms -- then we're talking about faxing or couriering stuff around. Also, there's the cost of getting that transcription and those lab results to the hands of the staff in the first place -- which is ALSO obviated by a decently designed EHR system).
That the people making decisions about buying EHR systems don't understand the above (and they don't, as a rule) just speaks to the degree to which medical businesses are mismanaged.
Posted by: Michael B Sullivan at Jun 27, 2007 1:21:32 PM
What do plastic surgeons do? Since they're in the private market, it'd be interesting to see how they've handled it.
Posted by: Matt at Jun 27, 2007 1:42:52 PM
The obvious answer is immigration limits and labor supply. There is a shortage of IT labor in the US to make the transition, or in other words, good programmers would rather work for a high-paying, option-granting internet startup than for tedious drudgework in a highly regulated industry, where as Brian Dailey says, medical consultants are likely to earn more than the programmers. The "high capital costs" mentioned above are just a consequence of the difficulty of hiring enough programmers for it.
Note that people like Steve Case and Jim Clark both went into this area after the dot-com boom bust. they would have done medical records startups before internet startups if they thought it was more profitable to do so.
Posted by: DK at Jun 27, 2007 2:18:25 PM
Kaiser Permanente is going electronic at all their locations. They have already done so here in Central California. But they have an incentive to do it, since they are an insurer that provides almost all their care internal to the firm, and can do lots of helpful database queries internally to optimize care based on that data. Most other health care insurers probably do not have the same incentives to switch as Kaiser does.
Posted by: Sisyphus at Jun 27, 2007 2:20:17 PM
So I just started working in this exact field. My job description entitles me to handle most electronic medical records. As for why we don't, we do. More hospitals are getting used to storing records electronically. However, when it comes to the legality issue, hospitals still need paper documents to show consent, original source of diagnosis, etc. These of course are being scanned into the record databases of the larger hospitals but people seem to need hard copies for actual proof. Laws are also beginning to allow for more validity behind electronic documents. The real question is: why isn't every hospital linked together electronically? We are still resigned to use fax and telephone as a means of confirmation between institutions. It would immensely help doctors and people in my position to gather information from other hospitals more quickly so that proper diagnosis and treatment are possible.
Still, to answer most of it, it would come down to just one reason. Money.
Posted by: MikeLee at Jun 27, 2007 2:34:15 PM
So, this is all well and nice, but YMMV, depending on vendor, project, execution, hell, even how the contracts are structured (who's responsible for maintenance, will the state or the contractor be running things, . It's not for no damn good reason that Texas canceled the deal with Accenture to handle the CHIP program. Accenture dropped the ball big time -- instead of building and running the service even at the level of quality the state had been providing, they succumbed to the capitalist impulse to cut costs to increase profits too drastically, and the resulting degradation of service cost them the contract and the state lots of time and money to rehabilitate the program. Partially.
These matters are not simply economics, or improvement of service, or politics, or any of many other issues. They are a convolution of all of the issues.
Posted by: agm at Jun 27, 2007 4:07:41 PM
In such a system, the initial, and most important, issue is setting data interoperability standards. This is a place where it might be reasonable for the government to step in and organize it. Once standards for the data exchange are set, it does not really matter what kind of software the particular doctor uses, as long as it sticks to the standards. So IT developers can compete and introduce products designed for particular purposes.
The standards have to be open, license-fee-free, and expandable in the future. It works very well with ISO (and other) standards, there is no reason why it should not work here.
The reason why internet flourishes is exactly this. You have basic communication standards everybody has to stick to, and user-end devices can be extremely varied.
The problem why this is not yet introduced is that no particular big player in the medical business is interested in it, since it would only undermine their position (while customers would benefit). A small practitioner would gladly buy (or maybe even get it for free as an open-source solution) his software before others do so if he knew that this software will be able to communicate with others in the future.
Posted by: pinus at Jun 27, 2007 6:21:02 PM
pinus: HL7 is an annoying standard, but it's a standard, it exists, it works, and most medical systems can read and write it.
A problem with it is that there is a really wide variety of medical information that can be written, and so the standard is really, really, really large in an attempt to satisfy the many kinds of data that could be interchanged, and that leads to some confusion.
But, ultimately, it's not the obstacle. If there was a lot of money in the business of connecting healthcare systems, people iron out the kinks in HL7. The field languishes because the healthcare industry isn't really interested in finding good EHR products and buying them, not because the technology isn't there.
Posted by: Michael B Sullivan at Jun 27, 2007 7:13:04 PM
