Why an EMR doesn’t necessarily deliver better patient care

Regular readers of this blog know that the mere introduction of an electronic medical record doesn’t necessarily guarantee better patient care.

There are multiple reasons for that, including the fact that many systems are archaic in nature, counter-intuitive, and doctors are forced to learn multiple systems.

Yesterday, the WSJ’s Health Blog posted a study showing that hospitals with an EMR don’t necessarily have better quality measures.


According Rand Corp.,

trying to introduce an EMR system to an already complex health-care workplace can cause “a myriad of unintended consequences” in terms of workflow and communication. That’s especially true with the full-bells-and-whistles systems, which include things such as computerized physician order entry system … “The complex systems are more difficult to implement and use,” [the lead researcher] says.

Measures such as pneumonia and heart attack quality scores weren’t improved, and in fact, “quality improvements at hospitals that started using an EMR system for the first time during the study period or upgraded to a more advanced system also mostly lagged those at hospitals that made no change to their EMR capability.”


What does this mean?  Hospitals and physicians cannot simply upgrade to an EMR and expect better patient care immediately.  It’s a painstaking process that must include physician input at every turn.  It’s the doctors that provide the care, and they must be given the appropriate health IT tools to make their lives easier, and not be burdened with the obstruction common to many EMR implementations.

There’s no doubt that, eventually, doctors in the United States need to step into the digital age and document, order, and write prescriptions electronically.  But the current generation of health IT isn’t ready for prime time yet, and as such, little immediate quality improvement should be expected of the current electronic medical record push.

 is an internal medicine physician and on the Board of Contributors at USA Today.  He is founder and editor of KevinMD.com, also on FacebookTwitterGoogle+, and LinkedIn.

Comments are moderated before they are published. Please read the comment policy.

  • Bladedeoc

    But, but, but . . . computers! Information age! Synergy!

    It is my experience that EMR has been from the start a plan developed by insurers and regulators, incentivized by bureaucrats, and managed by non-clinical IT drones. But perhaps I’m bitter because we are in the midst of a incredibly bungled order entry system roll out.

    • Vox Rusticus

      The inside pitch was one between the IT companies and insurers and the government who pitched that the data these records would generate as a “mine-able” resource from which “cost savings” could be extracted and that the data would be free to them. This encouraged government to try to force these systems on providers, greatly assisting the IT companies in bringing product to buyers, who would be forced to purchase or suffer penalties.

      The outside pitch was a lot of hand waving about efficiency and cost savings and better coding and other promises that practices and hospitals would reap benefits, ignoring the fact that they would bear the cost of creating the databases (that sit in isolation.)

      The IT and hardware industry have coopted the government into creating by force and punishment a sales market for their goods with absolutely no requirement that these products show any utility to anyone at all.

      Great work if you can get it.

    • horseshrink

      Gosh … you stole my thoughts!

      In my experience, the point-of-care end-user has the least interest in using EHR technologies.

      When EHRs give us the same “gee whiz … I really like this” reaction that smartphones provide, then spontaneous desire will fuel EHR adoption.

      We just aren’t there yet.

      If we aren’t there by the time incentive payments dry up, office based EHRs databases will be dumped to paper and the equipment donated to people who want to use it.

  • http://www.occampm.com/blog Michelle W

    As always, the important thing to remember is that technology is a tool to be utilized by the user. Whether we’re discussing the wheel or the iPad, neither of these inventions are worth much on their own. It is only their skilled use that makes them valuable, just as their misuse can have tremendous consequences. The car is an excellent case in point: more people die from automobile accidents each year than plane crashes (compare the Census Bureau stats here and here), and they pollute the enviornment; yet we still find them a valuable, useful part of our daily lives. With EHRs, it’s the not a one-size fits all scenario. Just as there are three different major gaming consoles, there are many different types of phones and tablets, and the number of content management system options seems to multiply exponentially online, so too there are different types of electronic records to fit different needs.

    Just this morning I posted a video I found online that lists 17 benefits a provider might derive from EHRs. None of these are guaranateed, and no, medical providers should not expect this technology to “transform” their practice: they’re the ones who are the transformative power in medicine. Technology is simply another tool to help them effectively do their job. As Dr. Halamka MD wrote at his blog today “Automating a dysfunctional manual process will not yield a successful performance improvement outcome. Before any technology project is launched, the business owners need to understand their own process flows and goals for improving them.”

  • http://acoready.com ACOdave

    Technology is a tool that we utilize for patient care, it does not replace the fundamentals of patient care. For some apparent reason people love to throw technology to solve problems, while technology will solve a lot of the problems. It should be used as a tool. The practitioner are still required to think

  • pcp

    “Why an EMR doesn’t necessarily deliver better patient care”

    An EMR doesn’t deliver patient care period.

  • Stalwart Hospitalist

    Keep in mind that the elements of the ORYX core quality measures that hospitals are being judged on are often not affected by the presence of an EMR per se.

    For instance: it is not readily apparent how installing an EMR in the hospital can make a hospital system get a patient with an acute STEMI into the cath lab within 90 minutes a greater proportion of the time.

    Also: an EMR will probably not make it easier to make sure the patient with pneumonia gets blood cultures taken prior to the administration of antibiotics.

    ACOdave is correct: the EMR is just a tool. A powerful, complex tool, to be sure, and one that comes with its own risks, but still just a tool.

  • horseshrink

    I smiled to read this article:
    “Six Proven Technologies For Boosting Nurse Workflow, Care”

    EHRs “did not make the list.”

    The other stuff sounds useful.
    1. Workflow management systems: = collecting information from multiple sources and integrating it into a single display
    2. Real-time location systems (RTLS): used to locate equipment, patients and staff. A nurse with whom I work vouches for the usefulness of this from prior work at another hospital.
    3. Wireless mobile VoIP communication: For example, when a patient’s heart monitor goes off it can send a message to the nurse’s wireless device, which can also be used to respond to the message.
    4. Wireless patient monitor technologies: “Going, going gone” is what one hospital called the solution it used to alert the nurses when patients were trying to get out bed. This technology even comes with the ability to record a family member saying “Mom don’t get out of bed,”
    5. Delivery robots: At a Washington Hospital, two robots deliver routine medication carts to the units.
    6. Interactive patient system: two-way communication and delivery of multimedia content at the bedside. For example the system can prompt patients to view educational videos on such topics such as hand-washing and fall prevention which frees up nurses’ time for patient care.

    Maybe, one day, EHRs will become a “proven technology for boosting nurse workflow, care.”

    • pcp

      Multiple studies have shown that hospitals have to significantly increase nursing staff when they go to EMRs, and staffing does not ever return to pre-EMR levels. But microscopic improvements in “patient care” are attributed entirely to the EMRs, not the greater nursing presence.

      • horseshrink

        An elephant in living room is data entry. Takes too long.
        So, docs are trying “scribes” now to free them from the template driven EHR documentation time suck. (I don’t know how the overhead incurred by a scribe compares to regular transcription.)

        EHRs also cannot capture our normal handwritten conceptual shorthand, including symbols, impromptu arrows, lines, tables and illustrative graphics – like a drawing of where Joe got cut with the hunting knife and how large the injury was.

        I favor very basic templating … e.g. HPI, Past Med Hx, Social Hx, etc., in the manner to which we are already accustomed. Beyond that, the EHR geeks and nerds should do what Google does … mine the data for us. We don’t serve the computer. It is supposed to serve us.

        My experience with IT gnomes reveals a disturbing tendency by some to believe that THEY are the customers!

        • ljpmt

          In addition to comparing the costs of scribes versus transcriptionists, what about the costs of hiring “health information analysts” or “data integrity specialists” that is occuring at hospitals with EHRs in place? What exactly do these folks do?

          • horseshrink

            Several years ago In our state hospital system, a new position was created and manned statewide for each treatment team – the “treatment team assistant.”


            To try to mitigate the productivity, data input and data access problems created by the electronic health record.

            No joke.

  • jsmith

    “We don’t have the measures we need to [assess] the return on investment” in health IT, he says.
    That quote from the WSJ says it all. This entire EHR craze is a gigantic exercise in putting the cart before the horse.
    Kevin says that we docs need to step into the digital age with EHRs eventually. I could not disagree more. We docs only need to adopt technology that has been shown in numerous well-designed studies to improve pt care, lower costs, or both. If you believe we must go electronic even if there is no evidence for the benefit of doing so, you are a victim of the propanganda machine you are criticizing in this post.

  • http://www.innovativepdf.com ted rudolph

    Any new technology is going to have growing pains. The horse was better than the first cars; You could do accounting faster on paper than you could on the first computers. You can do faster patient care with pen and paper now over most EHR systems. This is the beginning stage of getting EHR adoption. Thinking that these systems don’t help now is only looking at the situation in the present. You have to look to the future with these systems. They are only going to improve. And with that improvement will come an improvement in patient care.

    I’m afraid that if not for the government promoting EHR use they would never get adopted. And the medical community would still be analog in a increasingly digital world.

    • horseshrink

      If the industry creates technology we want … why gosh golly, we’ll use it!

      Didn’t need the feds to carrot and stick smart phones into our hands!

      Anyway, here’s a comment I submitted to ONC today – re: an arena in which the government might play a useful role – standards:

      I favor broad EHR data constructs standardization and thus agree with the proposed data metatagging project.

      I think EHR technologies will improve in their usefulness to the extent that market forces are in effect.

      EHR customers are currently paralyzed once they purchase a product and build a patient database. Vendor specific database idiosyncrasies obstruct customers from changing products at will. Ask disgruntled customers why they don’t buy a new product. They can’t — without great expense.

      Migration expense + new product expense = prohibitive expense.

      (This is painfully evident in the system in which I work.)

      Broad standardization of data constructs can allow customers to change EHR products at will.


      Unleashing of previously stymied market competition forces. Product quality increases and cost diminishes.

      Product quality improvements would more responsively follow the needs and desires of the point-of-care end-user and thus fuel a spontaneous grassroots desire to adopt EHR technologies (e.g., as seen with the smart phone phenomenon.) To achieve a national objective, an endogenous desire to adopt EHR technologies is much more important than temporary exogenous, financial incentives.

      I am hopeful that the development of a “standard language” using metatagged data elements is a move in that direction.

      • http://www.innovativepdf.com Ted Rudolph

        New technology is one area that sometimes is not adopted quickly. The telephone was initially turned down by Western Union as being a “useless toy.” Google was turned down by Alta Vista and Yahoo. You can practice great medicine without an EHR right now. That problem is what about in the future? Right now I can access everything about my life except my medical records.

        I agree that the present EHR systems turn doctors and nurses into typist. My doctor sits in front of a keyboard typing away as he ask me questions about my health. He only looks up after he has entered all of his notes into the EHR. It’s awful, but this is just the beginning. The technology will only get better.

        Totally agreement with the government and standards. I don’t think that we can count on the EHR companies to product any standards. But, we should use the ISO group for standards. Adobe is one company that gave their technology to the ISO and now PDF is an international standard(ISO 32000).

  • http://diagnosticinformationsystem.com Bob Coli, MD

    One crucial physician workflow process where the current generation of health IT isn’t yet ready for prime time is the dangerously archaic and counter-intuitive design of the interface physicians must use to view, analyze and share cumulative diagnostic test results. The costly problem, ignored since the 1960s, is that all ambulatory and hospital EHR and HIE systems use infinitely variable formats to report diagnostic test results as incomplete, fragmented data that is hard to read and navigate.

    There is a simple solution to this costly and dangerous data management problem, just no easy solution. EHR and HIE platform vendors need to upgrade to a vendor and platform-agnostic standard format to report all test results as complete, easy to read, clinically integrated information. Doing this would add substantial workflow value and competitively differentiate their brand from other ONC-certified EHR products.

    Historically, physicians have had absolutely no way to incentivize HIT vendors to compete for customers by improving the usability and reducing the total cost of their products. However, three recent developments offer some hope that this long overdue physician interface redesign could actually happen.

    One is that ONC is planning to sponsor an open community of practice for developers of health IT with a goal of “creating solutions to challenges that prevent rapid and widespread meaningful use of electronic health records” (1)

    Another is that usability testing has become part of the CCHIT Certification approach, at least for ambulatory EHR products. CCHIT usability testing includes “a set of twenty assessments that assure that screens are readable, well organized, consistent, and easy to navigate and that information is highlighted and flagged within the context of a clinician’s regular workflow.” (2)

    The third is the growing recognition by trial attorneys and potential defendants of medical malpractice liability being a “hidden danger in EHRs.” (3)

    Physicians, patients, hospitals, payers and lawyers all have an interest in whether this specific and correctable “system-wide EHR failure that creates problems in patient care processes” remains hidden or is finally corrected, either by market forces or government fiat.


    (2) http://ehrdecisions.com/2010/12/20/first-do-no-harm

    (3) http://boards.medscape.com/forums?14@@.2a0594eb and

  • http://www.meyersmedmal.com Jivanmeyers

    As was pointed out earlier, making a health record electronic is not useful if the original health record and its use was dysfunctional. An established tradition of non-communication between members of the health care team is in no way improved by making the entries legible. further, the existence of default entries (prior entry remains in the same field when next entry is to be made and verified or is copied with a key stroke to the same field) encourages physicians to assume the absence of change rather than conducting such examinations and asking such questions as would confirm the absence of change.

    I represent a child who as a newborn in a neonatal intensive care unit developed a herpes eye infection and later herpes meningitis. A physical examination is recorded each day of the child’s admission. The child’s first physical examination included all normal findings. The same findings, word for word, appear each day to the end of the child’s hospital stay. Nevertheless, at the time of discharge the child exhibited severe brain damage and was blind in one eye.

    At 2 weeks of age a lumbar puncture was performed. Progress notes for the next 5 days include a default entry “culture results pending.” In fact, the laboratory reported to the electronic record that cultures were negative and final on day 3. A further lumbar puncture was obtained 9 days after the first lumbar puncture. This time the CSF specimen was submitted for PCR DNA analysis for herpes. The result reported 2 hours later confirmed the presence of herpes. The diagnosis of herpes meningoencephalitis was made and appropriate treatment with acyclovir was finally initiated.
    Though this disaster was caused by the failure of those involved to have adhered to appropriate standards of practice, the health system’s electronic health record certainly facilitated the unfortunate result.

    • horseshrink

      In busy private practice and correctional settings, I eventually settled on semi-templated handwritten documentation. I could check a box for some variant of “normal,” but most of the room on the page was reserved for handwritten observations/thoughts. There were no “default” answers. What I saw and thought was amply evident by the end of each appointment. Templating here organized the information into basic sections, but did not create identical notes from encounter to encounter. (Also, handwriting allowed me to interact with my patients in a more normal, human manner during sessions, by the end of which my notes were usually done.)

      A serious risk with EHR default answers is an actualization of the old joke that WNL (“within normal limits”) actually means “we never looked.”

  • doctor

    Agree with the points you raise. Would also raise the point that a main goal of using an EMR is to comply and conform with government regulations and not to provide better care. The regulations are aimed more at minutia (enough “bullet points” in your HPI or ROS?) than at any meaningful efforts for good care.

  • http://www.emrandhipaa.com EMR and HIPAA

    I’m not sure about the ROI of an EMR in a hospital, but in an ambulatory practice there are some good ways to achieve a great ROI from an EMR. Here’s just a short sample list of benefits: http://www.emrandhipaa.com/benefits-of-emr-or-ehr-over-paper-charts/

    Granted, some of the benefits are hard to quantify with a number like legibility of charts. You can’t easily put a $$ on that, but it’s extremely valuable.

    Of course, like any project, you can lose a lot of money implementing an EMR if you choose the wrong EMR for your practice and/or don’t have or get the necessary skills to implement the EMR properly.

Most Popular