Health IT’s role in reducing medical errors

Two landmark reports by the Institute of Medicine (IOM) changed Americans’ perception of their health care system and launched today’s drive to improve the quality and safety of medical care in America. The reports were To Err Is Human, published in 1999, and Crossing the Quality Chasm, released in 2001. 

Both these reports highlighted the important potential role that health information technology (HIT) could play in improving health care quality and reducing medical errors. In fact, Recommendation #9 in Crossing the Quality Chasm called for “renewed national commitment to building an information infrastructure” and said: “This commitment should lead to the elimination of most handwritten clinical data by the end of the decade.”

The end of that decade is now just three months away, and not to mince words, we’re behind the ambitious schedule that the IOM report envisioned. Nonetheless, we have at last made the substantial commitment that was called for in the report.

Last year in the HITECH Act, Congress and the President authorized $27 billion in Medicare and Medicaid incentive payments for providers who adopt and make meaningful use of certified electronic health records (EHRs). At the same time, the Act created $2 billion in new programs to support the transition to HIT-assisted care. And this summer, the regulatory framework was completed for Stage 1 of the Meaningful Use path toward an EHR-based future in health care.

With the engines of change now in place, it is time to bring closer focus to other key issues for achieving the full potential benefits of HIT. One of these is the issue of improving patient safety.

We know, both in theory and practice, that HIT-assisted care can reduce errors and improve patient safety. In particular:

  • Reliable access to complete personal health information is the foundation of safe and effective care. EHRs are inherently superior to paper in delivering such access.
  • Even more uniquely, EHRs can use their computing power to automatically cross-check personal information and other sources. With such backup, clinicians can be automatically alerted when drugs or other treatments may be contraindicated because of allergies, potential drug interactions, or other factors.

At the same time, however, it would be naïve to suppose that HIT-assisted care can deliver its full patient safety benefits in a single stroke – or that HIT will not present its own safety issues. Clinicians need to become familiar with new EHR systems, which will take time. EHR systems themselves need to evolve and improve. We need to ensure that the “decision support” information they provide is accurate and personalized. Their interfaces need to grow in user-friendliness. Even safety alerts need to find the right medium and avoid producing “alert fatigue.”

These challenges can be met – and indeed, the very “fix-ability” of HIT-based care can be one of its primary safety benefits. HIT systems tend to record and expose patient safety problems when they occur, while paper-based care too often hides them. And EHRs are amendable to rapid, systemic correction of problems – while corrections in a non-systemic, paper-based clinic can take years to accomplish, even when they are identified.

How can we maximize patient safety through HIT-based care? What roles and actions by government, the private sector, and health care providers themselves can help achieve the full potential benefits that were sought in those seminal IOM reports?

As we address these questions, there is no better source of guidance than the IOM itself, building on the same expertise and convening power that produced its initial reports 10 years ago. For that reason, the Office of the National Coordinator for Health Information Technology has contracted with IOM for a follow-up one-year study. In this study, IOM will:

  • Identify approaches to promote the safety-enhancing features of HIT while protecting patients from any safety problems associated with HIT and preventing HIT-related patient safety problems before they occur;
  • Identify approaches for surveillance and reporting activities to bring about rapid detection and correction of patient safety problems;
  • Address the potential roles of private sector entities such as accrediting and certification bodies as well as patient safety organizations and professional and trade associations; and
  • Examine existing authorities and potential roles for key federal agencies, including the Food and Drug Administration, the Agency for Healthcare Research and Quality, and the Centers for Medicare & Medicaid Services.

As this study is carried out, we will move where appropriate to improve surveillance, reporting, product safety, and clinician performance. But at the same time, we will anticipate a “deep dive” in knowledge synthesizing and a new round of productive recommendations from the IOM.

There is every reason to believe that HIT-assisted care will be transformative for American medicine, but no reason to think the change will be easy or instantaneous. We are returning to the IOM as a key partner in helping to refine the course that it first helped to chart a decade ago.


  1. Back Brace Dan says:

    One of my favorite subjects in college was Information Technology and the endless benefits it offers so many facets of our daily life. I didn’t even think about the medical industry application till I switched healthcare providers and the new provider had all of my records/charts/x-rays stored digitally. My Dr. had 2 offices and he’d often have me meet him at either, so his adoption of HIT made it easy for him to turn to his computer screen, pull up my x-ray and explain what was going on – in either office, without need to remember to bring the x-rays with him everywhere. Thanks for furthering a great cause.

  2. Sherry Reynolds says:

    Clearly patient safety is a key goal and outcome of a high quality healthcare system and the inclusion of IOM is an exciting development.

    Many people who are familiar with the IOM research tend to focus on 1) safe, 2) effective, 3) efficient, 4) timely and 6) equitable care but for some reason many people seem to forget that the fifth dimension of a high quality healthcare system is “patient centered – encompasses qualities of compassion, empathy, and responsiveness to the needs, values, and expressed preferences of the individual patient. It is such a fundamental core value and goal, (versus a project) that you need to include patient centered design principals in every program.

    Central to patient safety is trust in the almost sacred conversations and relationships that occur between health care providers and patients and it is critically important to protect that. Although health IT can be an amazing tool to extend that communication and education beyond the walls of the exam room it also provides new powerful avenues for other interests to insert themselves into that conversation and impact patient preferences, the standard of care and even safety.

    For example, some states, many medical schools and clinics have blocked pharmaceutical representatives from providers offices or at a minimum clarified their relationship to pharmaceutical sales reps and their gifts. This is a result of ample social science evidence that demonstrates that when you give pharmaceutical companies “access to prescriber identified prescription data, providers are subjected to undue influence over prescribing practices, it raises costs, promotes irrational drug selection, compromises patient safety and threatens the professional integrity of the medical profession.” Now without disclosing any PHI to drug firms we are seeing EHR’s which essentially allow a virtual drug rep into the exam room and it raises genuine concerns for patients around trusting the entire healthcare system.

    One EHR vendor recently announced that they are now the fastest growing and largest “free” electronic health record systems with over 43,000 using their system covering over 1 million patients. According to their web site they are “free” because their business model delivers targeted ads to providers at the point of service. I wonder how many patients are aware that their providers are being subjected to drug ads in order to see their patients medical records or if they would lose trust in health IT if they did? Certainly no one would ever expect to allow a drug sales rep to sit in the exam room with patients or capture their data during the check in process like another EHR vendor does nor allow them to peek at their problem list and then deliver targeted direct to consumer ads on the check in screen. Patient’s shouldn’t have to wonder if the clinical alerts their providers are receiving were developed by pharma or if the educational materials they are being directed to where sponsored by a medical equipment firm.

    Clearly E-prescribing is essential to improving patient safety and reducing the medication errors that according to the IOM in 2006, affect 1.5 million people and costs the nation at least $3.5 billion annually, not including expenses for lost wages and productivity., but the “cost” to implement health IT shouldn’t be a virtual pharmaceutical sales representative in the exam room with the patient and provider. Patients themselves are often the best at catching and preventing errors in healthcare and many hospitals have implemented patient safety councils that are highly effective.

    Perhaps we will also see virtual patient safety councils (or design principals) embedded in our health IT systems and family and patient work-flows will be as important at mitigating errors as provider needs are and pharma can work with providers, patients payers and providers to develop safer, more effective, efficient, equitable, timely, patient centered health care tools and systems.

    The real source we need to return to is the patient.

  3. Bob Coli, MD says:

    One important EHR feature the new IOM study should examine is the flawed user interface all EHRs and HIEs use to report patients’ cumulative diagnostic test results.

    Since the mainframe computing era, this key clinician interface has used infinitely variable formats to report results as fragmented, incomplete data that is hard to read.

    The logical solution is to use a standard content exchange format to display all results as clinically integrated, complete information that is easy to read, analyze and act upon.

    Significantly improving test results reporting will facilitate the viewing and sharing of this vital information by physicians and patients and could complement other IOM recommendations for using HIT-based care to help maximize patient safety, improve health for individuals and populations and reduce per capita costs.

  4. Alex says:

    I agree with Sherry. Healthcare, pharmacy, medicine and hospitals have become more of a business instead of a means of turning sick people healthy. The patient should become the focus of the health industry again. I hate to call it an industry, but that’s what it is at the moment.

  5. Dr. G. Holt says:

    Doctor Defined Defaults – Flexibility, Efficiency and Safety Benefits

    HIT software / CPOE etc should allow doctors to define their
    OWN sets of default order and treatment sets for conditions
    that they commonly encounter. (This is akin to the definition
    of macros in spreadsheets, although considerably more powerful).

    When a doctor encounters a patient with a particular condition,
    (s)he can activate their default order and treatment set for that
    condition. This may include standard blood tests and standard first line
    medications, scans and other investigations.

    The doctor then may change ANY PART of this default treatment plan
    for their specific patient. However, since we are all human and human
    memory and attention are subject to error, it provides the doctor’s
    OWN default approach as a starting point. As such, it functions
    as a checklist to reduce error and increase efficiency.

    It is essential that each doctor be allowed to define and modify
    THEIR OWN defaults, and choose WHEN to use them – otherwise
    it will meet with considerable – and appropriate resistance.

    This being said, Doctor defined default templates could be SHARED
    among practitioners within a group, or hospital or even more
    broadly. A website – akin to “open source” software websites –
    could allow doctors to post their favorite DDDs and to allow other
    practitioners to view and use them. This sharing would expedite
    and improve best practice transfer.

    Additionally, high quality DDDs – incorporating published
    clinical guidelines and/or the results of recent meta-analyses
    – could be made freely available and adopted as the INITIAL
    default in that area by many doctors. Again, it is essential
    that individual doctors be allowed to choose whether or not
    to use any particular default, be allowed to freely MODIFY and
    customize the default for their own use, and be allowed to
    change any part of the default for any particular patient where
    they deem this to be clinically appropriate.

    CPOE and electronic prescription software standards should
    define a common and easy to use language/interface for creating
    Doctor Defined Defaults and modifying them.
    All such systems should be required to facilitate the use, exchange,
    modification and definition of Doctor Defined Defaults.
    DDDs may lead to increased efficiency and safety while maintaining
    clinician flexibility.

  6. Dr. G. Holt says:

    A Potential Bias in Comparing Error Rates

    Estimates of error rates may be biased against HIT if
    errors are more readily detected using HIT systems than
    with control (paper-based) systems.

    Any effectiveness studies will need to reflect this
    potential methodological concern.

  7. Gavin says:

    The first time health information technology impacted my world as a physical therapist was when I had a brief (and I mean brief) meeting with an ortho surgeon about one of his patients I was working with. Orhto surgeons are always super busy so if your interactions with them don’t happen quickly they don’t happen at all. He was able to turn to a computer terminal, pull up the patients file and x-rays to get himself oriented about the patient and answer my questions within 3 minutes.

    Instant access to HIT is invaluable for creating efficiency in the clinic and potentially averting any missteps on the part of health care workers who are working in consort with patients.

  8. Dr. David Blumenthal says:

    Thank you for taking the time to share your thoughts on this blog post. I appreciate your comments about personal experiences with health IT as well as your input on areas that the IOM should examine to help achieve patient safety goals.

    Some of your comments specifically addressed:
    •How EHR software should allow doctors to define, save, modify, and share their own default order and treatment sets for conditions they commonly encounter
    •The importance of examining the interface all EHRs and HIEs use to report patient test results
    •The need for patient-centered design principles in health IT systems

    What are some other areas that are important to consider when achieving patient safety goals through health IT?

    I encourage you to continue this valuable discussion!

    Thank you,

    David Blumenthal, MD, MPP
    National Coordinator for Health Information Technology

  9. Mark Kelly says:

    Patient privacy and security rights are a passionate issue of mine. I wonder if HIPAA needs a revision to keep up with the times so that it continues to meet its intended goals. From what I have seen at some medical institutions it seems the security stuff is often left to admin staff when it needs to be part of everyone’s job.

  10. Dr. Joseph Kousa, MD says:

    I used and implemented many EMRs for since 2001.

    Most EMRs are still at their infancy. A lot of them not designed by actual physicians who work on them day to day. Most EMR interfaces are crash prone and awkward and not efficient as it takes many clicks to navigate to target screens. Many interfaces actually cause doctors to rather make mistakes. For example, the EMR I use has some fields formatted to accommodate 4 digit numbers but the field actually is only one digit wide, so you could type quantity of tablets of 1200 at a time when you meant 120. Additionally, the lines so close instead of refilling Synthroid you could easily check the next line for something else. On and on. I was attending a conference at CCF foundation and the lecturer mentioned she made similar mistakes and the patient was angry and left her practice.

    Medical care delivery is intricate to document and report and must be secure to communicate over wide range of systems. An EMR must have the platform to accommodate the intricacies of healthcare delivery and reporting. There must be specific standards and protocols that take into consideration operability, efficiency and security.

  11. Autoapprove List says:


    I am now not sure where you are getting your info, however good topic.

Leave a ReplyComment Policy