Saturday, February 6, 2010

Electronic Medical Records

There have been many articles published about the benefits of electronic medical records. The US gov't has appropriated millions of dollars to aid in their development and adoption. The claim is that the adoption of such systems will save money by reducing the number of duplicated tests and also reduce patient mortality and morbidity because fewer errors will be made.

I will leave aside the question of the possible (unproven as yet) benefits of computer-generated hospital records and notes. I will note,however, that computerized hospital notes and orders use up more of a doctor's time, and that is the one dimension we already have too little of. It used to take me 5 minutes to admit a patient with pen and paper. Now it takes at least 25 minutes on a computer, in part because of "mission creep". By mission creep I allude for instance to the U.S. Census, which is mandated by the U.S. Constitution to do a head count every 10 years to re-apportion U.S. Congressional districts among the states, but now also requires you to tell the government how many bathrooms, bedrooms, and telephones you have, among other details. And when the computer crashes, all the residents are helpless.

I admit to two hospitals. In one of them, I cannot even admit a patient unless I answer the question: should the patient I am admitting have a flu and pneumonia vaccine at discharge (and there is no place to indicate that the patient refuses). In the other hospital (and only God knows why) they ask for the birth date of the admitting doctor. This information can't possibly help the patient, but someone wants the information. They even set up the answer matrix so that 00-00-0000 does not work.So now most of the doctors were born on 11-11-1911, because that is the easiest number to write that the computer will accept.

I am also concerned because not once in the past 5 years has a nurse called me from the ward to ask if my drug order was correct. The order goes by the computer to the pharmacy dept., who delivers it pre-labeled to the floor. The nurses seem to think that a computer-generated label must be correct, and besides, the pharmacy filled it. Of course the pharmacologist does not know the patient nearly as well as the nurse does, but it is efficient! I don't even want to discuss my reaction when I ask an intern for a patient's Hct. , and I am told that it isn't in the computer yet. It never occurs to them that the number is generated by the hematology laboratory, and you can call them up for the result.

Now let's look at electronic medical records, and their benefit for the patient. When I started practice, I used to give all my patients a photo-reduced copy of their EKG to keep with them , because when a patient hits the ER with chest pain, the doctor certainly wants to know what the previous EKG looked like. I found that fewer than 10% of patients carried the copy with them, so I stopped this practice.

Now a doctor's medical record system is supposed to be compliant with and interactive with the hospital system. This requires you to be HIPAA compliant which means the changing of your hospital password every 90 days, or else you cannot log on to the hospital computer. How in the world can two hospital computers exchange data? I can't imagine changing your password every 90 days for every hospital in the US, and the SmartCard still requires a password.

I also know that every doctor wants to review Xrays and MRI results personally before surgery, etc. So even if I have a transmitted electronic report about the result of an MRI, I would want to repeat it if the patient's treatment depended on it. This also holds true for cardiac echoes, stress tests, cardiac angiograms,etc. If you were a patient, wouldn't you want your doctor to personally review a study before operating? You also would not trust the result of a technically difficult lab test (e.g. N-terminal parathyroid hormone) unless you had total faith in the lab, and therefore you would probably repeat the test. And if different labs with different techniques have different normal ranges, this complicates matters still further. Similarly, where cancer is concerned, the pathologist and oncologist will want to review the biopsy slides personally, rather than rely on a written report.

All the above can be summarized by saying that no doctor would depend on a written report alone, but would want to see the actual data. If details are needed in the ER about a patient, then the ER doctor will probably communicate with the family doctor for information that is not in the computer system (e.g. the patient uses cocaine, is bisexual, etc.) because no computer system is totally secure. There is always critical data about a patient that is known very well by the family doctor but never makes it into the chart because of its sensitive nature. I would never write down in an office chart that a patient hates his wife or is having an affair if I know the wife has legal access to the chart, or vice versa. I am not being paranoid about the non-security of data in the hospital system, because in the last 2 years I have received communications from 3 Veteran's Hospitals that their information systems have been hacked.

I yield to no one in agreeing that computers are terrific for transferring information. What concerns me is the uncritical acceptance of information on a computer, and where the paper backup is when the system crashes. When patient lives are at stake, there must be accessible backup. As I recall, a few months ago LAX had to divert planes for 3 hours because of a malfunction in the computer program that linked their radar systems. There have also been recent articles on radiation overdoses received by patients because the technicians did not understand the intricacies of computer-operated radiation beams. And with a typical automobile having 30 computers with thousands of lines of code, it's a miracle that there aren't more recalls.

Let me close with the following anecdote (and I repeat that I am not a Luddite, and was an early fan of Wylbur): "Ladies and Gentlemen, welcome to the first fully automatic transcontinental airplane ride that is fully operated by onboard computers. We have 3 computers, and therefore double redundancy for safety. This system has been tested thousands of times. Relax as we take off, and be assured that absolutely nothing can go wrong, go wrong, go wrong, go wrong, go wrong,....."

13 comments:

  1. I already have had the same comment e-mailed to me by two experienced physicians: With electronic hospital notes, the house staff tend to cut-and-paste the same problem list from day to day. The result is that after two weeks in the hospital, the resident's notes are less accurate and less useful.

    ReplyDelete
  2. The code of the intern, if you did not see it with your own eyes it did not happen, has been complete obliterated by EHRs use in training. The ability to call B.S. on a lab result, radiology read, medication dose, nursing assesment, or historical fact in a patients record cannot be built into an electronic system. IMHO, The problem with current training that allows residents to lean on electronic sources of data is that their nose for B.S. is not as finely honed. I worry about the propagation of systematic errors as the next generation of physicians takes over for the current dinosaurs trained in the old ways.

    ReplyDelete
  3. I just heard first hand about another problem with EMR's. My aunt went to her new (younger) internist for the second time. She asked him a question about her cholesterol. He said it was on the computer, but when he logged on, it was unavailable at that time. If he were less lazy or computer-dependent, he would have done as I always do, and printed out the computer lab page and inserted it into the patient's chart. Instead he assumed he could always access the computer data whenever he liked. Surprise! We can read the Dead Sea Scrolls 2,000 years after they were written, but I'll bet my aunt's 2010 cholesterol results will be unaccessible in the year 2030.

    ReplyDelete
  4. In a similar vein, I inherited many charts when my older partner retired. I have his office notes going back to 1957, and the ink and printed lab results are legible, as are the Xray reports, consult notes, and operative reports. What are the odds of computer data that I log in today being available and readable in 20 years, let alone 50? If you think they can be, I have some great music on 8-track tapes I will give to you.

    ReplyDelete
  5. I book heart caths and we have a check off list, hand written may I add, to all the papers I have to fax to the cardiac DAR from the patient's chart. I then keep the chart on my desk until I get the hand drawn diagram of the cath with notes on the bottom the same day. I then get the typed report within a few days and it goes right into the chart. I also fax the diagram to the primary care doctor right away. We had a patient come from Florida with a CD of a procedure performed down there, we could not read it, we took it to the hospital could not be read...we decided to do a cath these CD'S were so bad and he needed open heart surgery. When he came for his HFU, I commented to him, that our handwritten diagram was better than his CD, he agreed and thanked us for taking such good care of him.

    In April all death certificates will be done online!!!! What will they do? Hold the body if the doctor doesn't sign online.....

    ReplyDelete
  6. I was just reminded by two of my radiologist readers of the lack of concordance between radiologists. They said that for chest Xrays the inter-observer agreement is about 90%, and for mammograms about 85%. You certainly wouldn't want to rely on a written report for these Xrays. And I also know that echocardiograms are VERY dependent on the skill and experience of the echo tech. I had one patient whose EF as judged from a cardiac echo increased (!?) from 25% to 35% in one year, despite the fact that he took no extra medicine to alleviate his low output state.

    ReplyDelete
  7. A recent study by the Rand Institute
    http://thehill.com/images/stories/blogs/randehr.pdf
    was unable to show consistent safety and health benefits from the use of electronic hospital records.

    ReplyDelete
  8. I am a nurse, and have experienced firsthand in separate facilities what happens when the computers crash. There is no back up plan. Documentation is impossible. This has occurred at nursing and rehab facilities, hospitals, etc. I much prefer the paper patient file we used at a privately owned small family practice, not affiliated with any larger healthcare organization.

    ReplyDelete
  9. Some study says, EMR (Electronic Medical Records) is the new standard in medical technology. Basically, hospitals are hoping to have the entire medical record of patients available on the computer system, including all notes written for that patient. You can see the potential benefits. Thanks for the post.


    -mel-

    ReplyDelete
  10. The problem is that (a) the doctors have to lay out a lot of money, which is very expensive for a single practitioner, and there is still no hard evidence that this interventions saves lives. Let me just note that a nurse would routinely call me at least once a month to clarify an medication or to ask a question about the wisdom of that medication in this patient. I never discourage such questions, because it isn't fair to the patient. In the five yeara that we have had EMR's for the in-hospital pharmacy orders, I have not been called once by a nurse. Did my orders somehow become more obvious and medically appropriate, or does the nurse assume that a pre-packaged drug dose,prepackaged by the pharmacy with the proper bar codes, is always the correct dose. This makes me a little nervous.

    ReplyDelete
  11. The NY Times of Tuesday, May 31, 2011, on p. B1, the first page of the business section, has a long article about the difficulty of securing the privacy of EMR's and how over 10 million records have been compromised to date. You have to assume that anything you put on the internet, no matter how "secure", can eventually be read by someone, and the punishment for doing so is rarely as bad as the release of your private medical knowledge can make you feel.

    ReplyDelete
  12. An article published online on June 29, 2011 by the Journal of the Informatics Association (a production of the British Medical Journal) showed that electronic prescribing had just as many errors as paper prescribing: http://jamia.bmj.com/content/early/2011/06/29/amiajnl-2011-00205.abstract

    please note jamIa, not jama

    ReplyDelete
  13. In today's (Thurs. Feb. 10) NY Times, on the first page of the business section, there is an article about how a certain % of those customers who paid their bills online via their Citibank accounts were sometimes billed and had deducted the same amount twice between July and December, 2011. Clumsy sentence, but you get the idea.

    ReplyDelete