Is Electronic Health Record Safety a Paradox?

Abstract

Electronic health records (EHRs) are now the most common method for documenting and communicating patient information, providing the most comprehensive view of patients’ intermittent experiences with the health care delivery system. Although use of EHRs has improved patient safety, the largest area of undesired consequences from EHR use has also involved patient safety: EHR-related errors have resulted in significant patient harm and even death.1,2A recent integrative review of the literature on the impact of EHRs on nurses’ cognitive work provides insights and highlights the impact of EHRs on nursing practice.3 The researchers found that most studies indicated EHRs did not support nurses forming and maintaining an overview of patients.3 Nurses found navigating EHRs to be challenging because of the scattered and fragmented presentation of information; they also experienced difficulty seeing the chronology of events and understanding clinical implications of various data.3 The template-driven design of documentation, as well as the restrictions on free text, made it difficult for nurses to express their clinical reasoning, and perhaps worse, understand the reasoning of others.3 Nurses found summary reports and handoff tools insufficient to support their work.3Although some suggest EHR risks to patient safety are primarily caused by nurse or user error, the reality is quite different. Understanding the underlying cause of EHR safety risk allows us to solve the problem and improve EHR safety. A statement in a recent article on clinical decision support (CDS) malfunction shed light on the nature of the broader EHR patient safety problem: it is “difficult to troubleshoot issues that appear in production [use] but not in the test environment.”4(p503) That statement captures the essence of use error, which is distinctly different from user error.5Use error is a “user action or lack of action that was different from that expected by the manufacturer and caused a result that (1) was different from the result expected by the user and (2) was not caused solely by device failure and (3) did or could result in harm.”6(p3) Use errors occur during the use of health information technologies, in this case, EHRs. Ideally, manufacturers should identify potential use errors while designing and testing EHRs, before they are used in the patient care environment.Because of the widespread use of EHRs and the potential impact of associated risks to patient safety, it is important to examine the current state of EHR safety. Is the delivery of health care safer or less safe with EHRs? What lessons have we learned for the use of EHRs? What is being done to improve HER safety, and is it enough? This issue of Technology Today considers these important questions.Although the use of computers in health care spans more than 5 decades, widespread use of EHRs did not begin until the past decade, following the passing of the Health Information Technology for Economic and Clinical Health (HITECH) Act in 2010.7 One of the key requirements of the legislation was that EHRs improve patient safety through features such as electronic medication prescribing and CDS. Some improvements have occurred, but there have been undesired consequences as well.Information on EHR safety as a major problem has come from multiple directions over the years. In a survey of members of the American Health Lawyers Association and the American Society for Healthcare Risk Management, 53% (of 369 responses) reported at least 1 serious EHR-related safety event within the preceding 5 years.8 The Joint Commission released a sentinel event alert on the importance of evaluating health information technology–related adverse events, including prevention and mitigation strategies to avoid patient harm.9 In 2019, the ECRI Institute highlighted inaccurate and missing information in EHRs as one of their top 10 patient safety concerns.10A study published in 2020 sought to answer the question: How did the safety performance of EHRs change in the United States from 2009 to 2018?11 Researchers found that EHR systems overall have only demonstrated modestly improved safety performance over 10 years. Of the 8657 hospital years observed, the most basic safety standards were met less than 70% of the time.11 The researchers did find a correlation between EHR vendors and medication safety performance, although specific vendors were not identified.11 The researchers concluded that hospitals that use EHR systems should routinely evaluate them for safety performance.11A 2020 report from Boston-based Coversys covering 10 years of nearly 12 000 events in 20 000 closed medical malpractice claims identified serious safety issues with EHRs.12 Specific findings related to EHRs included: The authors concluded that past malpractice data provide important signals for vulnerabilities in today’s EHRs.12The relationship between EHRs and patient safety is complex. But at the center of this safety paradox, you will often find the issue of usability. Although many nurses may not know the word usability or even the definition, they do spend a lot of time in the EHR recognizing what impedes their work and what helps it.Usability refers to the ease of use, usefulness, and satisfaction in achieving the desired goal when using EHRs.13 As noted by a Pew Charitable Trusts report14 quoting Norman,15 “Good usability should make it obvious to end users what they need to do.” This is true for EHRs; usability issues can be frustrating to clinicians and make their work unnecessarily more difficult.14Electronic health records play a central role in nursing practice today, which means even small usability issues can have a major impact on nurses’ workloads.7 For example, prior to EHRs, the documentation of patient pain required writing a few lines on paper. Now it requires locating a computer, logging in, finding the patient list, scrolling through the patient list to identify the correct patient, selecting the correct patient, responding to any pop-ups or alerts that may appear, navigating multiple screens to find the needed data elements, finding the right data element lists, scrolling through multiple lists of different data elements, clicking on multiple data elements, deciding if they have located all the desired data elements needed to document appropriately, closing the patient record, and logging off. The scenario is repeated when a nurse later documents the patient’s response to the pain intervention; this whole process is in contrast to simply jotting down a few words or lines on a piece of paper. And it is repeated with every assessment, every intervention, and every evaluation on every patient multiple times a shift.One might assume that nurses would become quicker at documenting over time as they learn the steps, shortcuts, and workarounds required to efficiently use EHRs, but that has not proven to be the case.7 Usability issues persist unless identified and fixed.14 Meanwhile, new features and content are routinely added, by vendors as well as health care organizations, that include usability issues.7 As a result, nurses’ use of EHRs includes activities supported by good EHR usability that improve their workflow and patient safety as well as poor usability that does the opposite.A seminal study by Lopez et al7 examined EHR-associated workload and usability effects over time, including whether EHR usability predicted workload. The researchers found a significant increase in perceived workload following EHR implementation that lasted the length of the longitudinal study, 2.5 years (P < .001).7 The study is considered to cover the longest time period known to examine these variables. Researchers concluded that improving EHR usability should improve patient safety.7Poor EHR usability increases cognitive burden, which in turn creates threats to patient safety. For example, the structure of EHRs with poor usability may not match how nurses think and work. This requires additional cognitive effort for nurses to integrate the EHR into their complex and dynamic workflows, creating opportunities for missed data in decision-making.A recent report by the American Medical Informatics Association took a retrospective look at previous predictions made at their 2009 policy meeting.16 More than a decade ago they predicted EHR usability issues would increase clinicians’ cognitive load.16 Today, they find the magnitude of the burden was significantly underestimated.16The amount of data in EHRs continues to grow, negatively impacting patient safety as it increases the cognitive burden of clinicians. Drivers of information overload in the EHR extend beyond clinician-required data and include billing, quality improvement, avoiding malpractice, and compliance.17 The increasing amount of irrelevant data that must be waded through, complicated by screen sequences, poor data or information displays, excessive alerts, the mismatch between EHR workflow and clinician workflow, and more can all lead to information overload and cognitive burden.18,19 All of this information is compounded by the need to recall, organize, and synthesize data and information from diverse screens, increasing cognitive burden and creating an environment for errors.20 A study by Ahmed et al21 confirmed the effect of cognitive burden on error rates in clinical decision-making by comparing the use of a conventional EHR and a streamlined EHR displaying relevant information.Nurses know when they have to navigate multiple screens or wade through a lot of data to find something they need. But sometimes safety risks related to usability can be hidden where nurses cannot see. One example is an alert malfunction that occurs when a CDS feature does not work as intended or expected.22 A study on the malfunctioning of CDS alerts in 14 sites across the United States found many malfunctions caused alerts to fire for the wrong patients (false positives) and fail to fire for the right patients (false negatives).22 Three primary causes of error were build errors, conceptualization errors, and the introduction of new concepts or terms.3 For example, a malfunction occurred because an internally used identification (ID) number for a medication was updated but the associated alert logic was not updated to include the new ID number.22 Across the 14 sites, similar malfunction patterns recurred and included challenges during system upgrades, during code sets and values maintenance, during environment migration, and during design and implementation of CDS.22Another example of a hidden safety risk is autopopulation. Newly introduced EHR safety risks can be the result of attempts to improve EHR functionality for end users. Listed on ECRI’s Top 10 Health Technology Hazards for 2021 is the autopopulation of medications resulting in fatal errors.23 The functionality was updated to make searching and selecting medications easier. Once a few letters are entered into the drug entry field, several suggestions are generated for the end user, increasing the risk for an incorrect selection.Unfortunately, there are currently no standards for routine testing of EHR safety as there are in other high-risk industries such as the airline industry.14 Yet, we know EHR safety risks exist and that some cause harm. What is or can be done to improve these risks? The answer is complex.First, current testing of EHRs focuses on the work of EHR vendors and includes initial design and development.14 This testing often excludes how well the newly designed and developed functionality works in the practice setting.14Second, once EHRs are received by health care entities there is further design, configuration, customizations, and updates. These processes make EHRs sold by the same vendor different across different organizations.Third, testing for use errors that may lead to risks to patient safety requires an extensive evaluation of actual human-machine interaction. Nurses have a right to use safety in the technology they use. Use safety is defined by the US Food and Drug Administration (FDA) as, “freedom from unacceptable use-related risk.”24(p3) Unfortunately, the FDA stance on use safety does not apply to EHRs.Fourth, health care entities are increasingly working with external vendors to provide different content, such as CDS content. Who is responsible for ensuring the CDS content in the EHR is accurate and current at the point of use? Who is responsible for ensuring changes in content are implemented in a timely fashion and implemented in all relevant EHR areas?This complex landscape makes it difficult to identify the source of a particular EHR patient safety risk. It also makes it difficult to identify, prevent, and mitigate the risks themselves. Compounding the situation is a dearth of comprehensive EHR safety testing capabilities.14Blijleven and colleagues25 looked for origins and solutions to potential patient safety risks by examining the workarounds performed in EHR systems by nurses, physicians, and clerks working in a large university hospital. They identified 15 clinician rationales for workarounds including EHR usability, which the researchers defined as a high cost in time and effort for users in completing a task.25 One example of a usability-related workaround included copy-pasting patient data, which Blijleven et al25 found to decrease patient safety. While copy-paste functionality was implemented when clinicians encountered user-interface challenges, the practice was found to decrease patient safety and clinician effectiveness with negligible influence on efficiency.25Several leading organizations have developed approaches to preventing EHR-related risks to patient safety.14 These include Pew Charitable Trusts,14 American Medical Association,14 American Medical Informatics Association,26 Office of the National Coordinator for Health Information Technology,27 National Institute of Standards and Technology,28 and the American Nursing Informatics Association.29 Still, EHR safety risks persist. The Table outlines available resources and is well worth exploring for nurses involved in health care safety, those wanting to learn more about EHR safety, or those seeking to improve the safety of a current EHR.Given the complexity of the relationship between EHR safety and usability and all the ways problems can present, the solution to EHR-related safety risks may lie in taking a different approach entirely to EHRs versus spending time, effort, and money on researching and fixing current-state systems. Resources should focus on eliminating nonvalue data as well as simplifying, integrating, and automating data collection.30 Safety of current EHRs has become a paradox: although EHRs were created to improve patient safety, they have created new safety risks that can injure patients. Are we repeating the same big miss when dealing with EHR safety by not recognizing the complexity of health care, communication, and collaboration, as well as the differences between simple and complex software, and the impact of failures on the safe use of EHRs?31 The amount of empirical evidence on EHR safety and usability alone is staggering, making the paradoxical nature of EHR safety unacceptable.