Dreaming of the EHR

The Dream:

I am at the APA convention.

It is in a crummy hotel attached to a mall. I am going down to the street level in an elevator and it stops at a level where there is a big cinema complex. An 8 year old boy runs excitedly into the elevator and jams in to my left. The elevator is crowded. He looks at my name badge and says: "What would George Dawson say if the Watson computer said he wasn't doing a good job?"

I glance down at him and say: "George Dawson would not care."

The kid says: "That's not good" and laughs intensely.......


It is all a dream. I had that dream early Saturday morning. It doesn't require detailed analysis. That last thing I did on Friday was try to review a 22-page paper record that was generated by a modern electronic health record (EHR) system. With the exception of a few paragraphs it was largely unintelligible. It contained bits and pieces of information. I was looking for imaging (CT, MRI, ultrasound reports) and ECG data, but instead could find only a few lines that summarized fewer results. There were no dates - no hospital admission or discharge date.  Although the hospitalization was longer than a week - there was no medication administration record or MAR - showing the specific dates when medications were changed.  There were no comprehensive reports that I am used to seeing for the past 30 years from Radiologists. There was no discharge summary. The documentation was basically unacceptable as a source of clinical information and yet it was created by a very high end, enterprise wide EHR system. It brought back a memory of a mandatory meeting I had with a "coding specialist" about 10 years ago. That person let me know that I had "passed" the documentation review in that I had ticked off the necessary "bullet points" so that documentation specialists would approve my EHR note for the day and the associated billing document that had to be submitted with each note.  She showed me how I could do things faster by ticking off a series of check boxes and electronically signing the note.  She was shocked when I told her that I really could not sign my name to that document because there was no sign that an intelligent human being had seen the patient.  For all of these reasons, very poor documentation in the EHR is always on my mind.

My attempt to read the last report may have been enough of a reason for the dream, but I also spent time on Thursday with a colleague who really dislikes the EHR for additional reasons. We regaled our spouses with tales of incomprehensible reports.  In addition, his reports require a synthesis of many imaging, lab, and clinical reports.  He previously used a system where all of the reports showed up in a queue and he could go down that list in chronological order to dictate the report.  In the new system, he has to go to tabs to find all of the reports he is looking for.  Some of these tabs are hyperlinked and the reports don't load very well.  In the end, he and his colleagues end up printing out all of the reports on paper so they can dictate then more efficiently without having to search for what they need in real time in the EHR.  That reminded me of an experiment I did about 8 years ago with the same EHR.  I went in and read all of the clinical notes looking for chronic diagnoses that were not addressed.  I came up with an additional 8 diagnoses from 340 clinical notes buried in the EHR. There is generally no good way for physicians to mine data on their own patients to make sure that they have done the most thorough assessment of their problems.  On the other hand administrators can get detailed numbers of mouse clicks by nurses in primary care clinics and rate their productivity in terms of mouse clicks, screen views, or tasks completed.  My colleague's theory was that the current EHR is selected for the administrative capabilities like monitoring doctors or nurses rather than any inherent advantage for medical staff.  The major evidence for that is that many EHR vendors have permanent staff in the hospital and they are making constant modifications to the EHR.  In many cases there are meetings of all the physicians in a particular department about these modifications.  The hospitals and clinics purchasing these systems are purchasing incomplete products that require what seems like constant revisions.

It has been about 15 years since the blight of the EHR hit physicians.  It was originally called the electronic medical record (EMR) but I suppose some business type decided that they could really solidify the corporate stranglehold on medicine by eliminating the word "medical" from another phrase.  Corporate psychology also dictates that they give the impression that they are maintaining health rather than treating medical problems.  That is another good reason for eliminating the word medical from the corporate lexicon.

The marketing of the EHR has been masterful.  The political hype promised untold savings.  National candidates seemed to suggest that we could actually "save" enough with the EHR that it would cover a substantial part of American health care inflation.  Any physician involved in the implementation phase of enterprise EHRs knew that was a bold faced lie.  There is no way that annual multimillion dollar a year licensing fees as far as the eye can see are going to save anybody any money.  In fact, I am certain that many clinics and hospitals have had to cut staff and services just to bankroll the EHR. Nobody has ever followed the money on the EHR debacle as far as I know.  Congress is well known to invent businesses and turn people into billionaires overnight.  All it takes is a few Congressional mandates about the need to use an EHR and electronic prescribing.  There is no mandate to keep things cost effective of make sure that independent practitioners can afford it.  There are mandates to implement EHRs and electronic prescribing and the White House brags about it.  The following graphic and text are from a White House document on the EHR entitled: More than Half of Doctors Now Use Electronic Health Records Thanks to Administration Policies.

The problem with the White House statement is that despite spending about $3 billion dollars a year and in some cases $44,000 per physician,  the value of the EHR for the reasons already stated remains in question.  It is very handy to be able to pull up lab results and x-rays on a computer screen.  It is also very handy to be able to send electronic prescriptions to any pharmacy in the country.  On the other hand, it is reasonable to expect that a multimillion dollar piece of software will write a report that any hundred dollar database software from the 1990s could write.  That same software should be capable of allowing physicians to search their own patient results for quality and report writing purposes.  In the end we are left with very expensive, high maintenance systems, and massive amounts of information that is either buried in storage because it is not easily accessible or because it is worthless and generated primarily for justifying a billing document.

That is one of the many real costs of having a health care system run by bureaucrats and politicians.  


George Dawson, MD, DFAPA


Attribution:

Photo at the top of this blog is by Paco Burrola on Flickr and is used courtesy of this Creative Commons license.

No Information From The EHR - An Ongoing Problem

Like most physicians - I like the concept of an electronic health record (EHR).  It is just that the real EHR as it exists is a far cry from the concept.  The proponents of the current EHR,  especially those who want it mandated by legislative activity continue to brag about the savings and all of the benefits.  Any physician looking for information or an ability to enter and move information without ending up in a click fest of mouse clicks knows the reality.  Any physician looking for a note that reflects an intelligent conversation between a physician and a patient is also left wanting.  Reading the electronic or printed out version of the EHR usually results in very choppy documentation.  Lists that are the result of not very intelligent coding by EHR IT engineers, notes produced strictly to meet billing and coding bullet points, and notes produced because they could be rapidly compiled with features like smart text.

All of this can be a nightmare for a compulsive physician like myself who wants to use all of the relevant information in patient care.  My career has been treating patients with complex medical conditions who are also on complicated combinations of medications.  Many have known heart disease and take combination of medication that can adversely affect their cardiovascular status and interact with psychiatric medications that I prescribe.  All of that needs to be considered.  Since ziprasidone (Geodon) hit the market in 2001, psychiatrists have been preoccupied with the QTc interval.  The QTc interval is the electrical interval that corresponds to the contraction and relaxation of the left ventricle.  In cases where this interval is too long it predisposes the patient to ventricular arrhythmias some of which are potentially fatal.   The FDA had a warning on ziprasidone about the potential for QTc prolongation and subsequently came out with warnings about citalopram.  In the course of clinical practice, many psychiatrists had already encountered this issue with older antipsychotic medications and tricyclic antidepressants.  The FDA makes these pronouncements but gives physicians no guidance on what to do about the clinical situations.  I have a practice of looking at ECGs and any Cardiology evaluations that have been done.  That is the only way the QTc interval can be determined and even then there are various factors that can affect it.

Rather than order an ECG, I will ask whether they have already been done and get the patients consent to have them faxed to me.  That result is frequently disappointing, especially in the case of the EHR.  I will often get a series of cryptic sheets, that look like a sparsely populated medical record.  There are often no coherent notes from physicians or if they are there, they do not contain standard information that I am looking for.  I have never seen an ECG tracing contained in these stack of records.  The best I can hope for is a brief note that lists an impression like "NSR - no acute changes."  An added bonus would be an actual description of the critical intervals.  For the tracing at the top of this page it would say:  "PR interval - 164 ms; QRS duration - 100 ms; QT/QTc - 434/415 ms."  That is really all of the information I need to know.  But the most important issue with the EHR is that all of this visual information is usually lost, unless I submit a second or third request and it usually has to say "send me the ECG tracing."  The medium that purports to provide a lot of information to physicians and put it at their fingertips is a bottleneck.  By the time I see the information I need to see, it is not necessary.  I have moved on and not recommended a treatment that I could have recommended if the ECG was normal.  That practice has been reinforced by getting an ECG after the fact and realizing that not only was there a prolonged QTc interval, and it was read that way by a Cardiologist but reported as "normal" in the EHR.

I will be the first to admit that there is minimal evidence that my tight QTc surveillance has saved any lives.  But my threshold is really to prevent any complications.  I am not treating acute heart conditions.  I am trying to make sure that I don't cause any by the medications that I prescribe, by ignoring a critical drug interaction, or by not recognizing the significance of a patients physical illness and how it needs to direct the therapy that I prescribe.

That doesn't end at ECGs.  I would throw in imaging studies (CT and MRI), EEGs, and even routine labs.  If the EHR is supposed to convey the maximum information why wouldn't all of the visual information of an episode of care be included?  Why can't all of the brain imaging studies be sent along as a disk or e-mailed to me?  Why do I have to read a 200 page fax and try to reconstruct all of the lab results  in a coherent manner that are spread randomly across those pages so that I know what happened in the hospital?

The EHR as it currently exists is a tremendous burden to physicians.  It takes far too long to enter data and quality notes about care are rare.  If you happen to lack online access to the program where the record is constructed, good like trying to piece together the information that you need for clinical decision-making.  Politicians are good with ideas, but none of them seems to be aware of the real problems that exist in these systems.  Despite that lack of knowledge they continue to insist on the wide implementation of these systems and that is really a tax on physicians that is being used to subsidize the development of EHRs and fund this industry.

Hopefully that will pay off someday, but the current problems have been there for at least a decade and there are no signs that they will be going away soon..



George Dawson, MD, DFAPA  

Enduring Problems Of The Electronic Health Record

I think the national debate is coming back to the more reasonable position that the heavily hyped electronic health records (EHR) will not save up hundreds of billions of dollars due to "efficiency."  But then again again any physicians not working as an administrator hyping the EHR could have told you this based on their experience over the past 10 years.   If I had to think of a reason, I would imagine it is the companies trying to build a moat around their businesses.  Software engineering can't possibly be this bad.  Wall Street jargon considers moats or barriers to direct competition with a company to be a good thing.  Let me illustrate with a real world example.

Let's suppose you are working in a clinic that is not online with the largest managed care (MCO) company in your area.  The only way you can get electronic access is to pay a huge licensing fee, but in many cases the software company will not even accept that licensing fee.  It will just conclude that that you are not big enough to do business with them.  At any rate, you need electrocardiogram information on a patient from that MCO because you are looking at a new abnormal ECG on that patient.  You need to know if the pattern on that ECG is new or it has always been there.  You request the records from the MCO.  They fax you 50 pages containing the lowest possible amount of information per page.  There are two one line references in that 50 pages to an electrocardiogram.  One says: "Prolonged QTc" and the other says "Normal".  There is no graphic information (the tracing) and no numerical information (the intervals with the associated times in milliseconds, the machine read out).  So after the work put in by you and your staff to request this data, you have just read through 50 pages and found absolutely nothing useful.  A review of all of the pages shows scant information on each page.  As an example, one entire page contains a chest x-ray report, when it could easily be printed on an area 1/20th that size.  Some entire sheets contain 1 or 2 lab values of 3 to 5 digit numbers.

I am convinced that the multimillion dollar licensed legacy wide EHRs are designed this way.  There is really no other explanation for providing such an abundance of low to no information records.    Their intention is obvious.  Make sure everyone is using their system and at some point make sure that the government is forcing people to use somebody's system.  All physicians should be using electronic prescribing right?  It is only a matter of time before politicians mandate access and an extremely expensive portal will be required.

There was a time when the medical record was coherent.  Maybe I was spoiled by reading what sounded like fine literature by comparison.  There was one Cardiologist in particular who wrote incredible notes for consults.  Reading those notes gave you all of the medical information you needed and it also left the impression that you had just read something written by a highly intelligent person.  Somebody you probably wanted to have a conversation with.  Somebody you could learn from.

What has happened to the medical record leaves a bad taste in my mouth.  It reminds me of when an EHR consultant was showing me their latest time saving way to create a choppy, incoherent progress note, and sign off on a billing document at the same time.  She assured me that the "compliance people" would find it completely acceptable for billing purposes.  When she asked me what of thought of their system she seemed taken aback by my response.

"I would be ashamed to sign my name on that note."

That was about ten years ago and the electronic health record has not changed much since.  It will still kick out a phone book sized print out containing minimal to no useful information.

George Dawson, MD, DFAPA