by Judith Effken, Senior Editor
& Jane Carrington
Effken, J., Benham-Hutchins, M. & Carrington, J. (2013). You Can Lead a Horse to Water, but. . . . Online Journal of Nursing Informatics (OJNI), 17 (1), Available at http://ojni.org/issues/?p=2387
Several recent events precipitated this column: First, while attending the Annual Symposium of the American Medical Informatics Association (AMIA) in November, 2012, we attended a number of sessions in which participants were questioning our ability to provide significant meaningful use of health information technology. Sure, we have shown that we can use computers to document the frequency with which we provide patients with “stop smoking” information—but is this really the “meaningful use” we are looking for? Perhaps we can even charge for treatments accurately using the electronic health record (EHR)—although early evidence of substantial fraud in electronic documentation of charges suggests otherwise (e.g., Abelson, Creswell & Palmer, 2012). But it seems to us that our current meaningful use isn’t even close to what we need to be doing. If the electronic health record is the informational “water” to which we are being led (especially by the current national push for their implementation), why aren’t we “drinking” it?
The second precipitating event for this column was a Health Affairs article by Beth Ann Swan (2012), who is Dean and Professor at the Jefferson School of Nursing in Philadelphia. The article detailed the numerous communication breakdowns that she encountered when her husband had a severe stroke while on a business trip to Chicago and she had to negotiate the health care system across the full continuum of care in two cities. When posted on the AMIA list serve, the article struck a nerve. We are frequently told that, if we—or someone we care about—is hospitalized, there needs to be a health advocate to ensure the patient’s safety. When we go from nursing unit to nursing unit—or from physician to physician— even in the same system and building–we need to provide the same health care information repeatedly.
Perhaps even more troubling is the evidence that, despite having patient records available electronically—even at the point of care, nurses don’t use the available information to plan care for patients, instead relying on verbal report and/or the patient’s or family’s memory. Our own research suggests that nurses often regarded the information in the EHR as irrelevant for ensuring continuity of care (Carrington & Effken, 2012). In response, those nurses had grown increasingly reliant on verbal communication–even though verbal communication is known to be fraught with errors. Based on an ongoing multisite study, preliminary data and content presented by Carrington at the 2012 AMIA Symposium revealed that it was not just details that oncoming nurses assuming care for a patient did not recall. Some of the nurses could not recall or correctly identify the change in patient status about which they were told in change of shift report. For example, of seven clinical events or unexpected changes in patient condition, only one event was correctly reported by the nurse continuing care. Two patients experienced two events; and, in both cases, the nurse receiving the change of shift report correctly identified only one event. One nurse reported gaining knowledge of the patient’s status from the EHR by having read the physician history and physical, as well as receiving a verbal report; however, the nurse failed to report the clinical event that had occurred on the previous shift (Carrington, 2012). These communication issues have a circular pattern. Nurses view the information within the EHR as irrelevant, so they increase their reliance on verbal communication. Yet, cognitive limitations inhibit recollection of information shared through verbal communication in the complex healthcare environment in ways not fully understood.
Benham-Hutchins recalls repeating the same information about her husband’s condition to each oncoming nurse because they didn’t take time to read her husband’s record. As a result, she became her husband’s “living patient record.” Unfortunately, each of the authors has encountered similar situations—which become all the more puzzling in environments where the electronic record is accessible in every room, but remains untapped as an information resource.
Clearly the primary problem—nurses not getting adequate, accurate information BEFORE they care for a patient—is not new. In our considerable clinical experience, with the exception of student nurses who are required to review patients’ records the day before they care for them, most practicing nurses rely almost entirely on the oral change of shift report. How can health information technology help change this?
Most electronic records still replicate the discipline-based silos of paper-based patient records. Our own research has identified the complex information exchange patterns that resulted from hybrid records (not interoperable) (Benham-Hutchins & Effken, 2010). Preliminary results of a study to identify the information sought by nurses when assuming care of a patient has revealed that nurses prefer to rely on paper or replicate the paper-based Kardex systems (Benham-Hutchins, 2012). Is this reluctance to change from traditional handoff practices to using the content of an electronic record system based on habit, ritual, comfort with the familiar, electronic system design or interface issues—or all the above?
The vast amount of data present in the modern EHR system presents both challenges and opportunities. The volume of information in records can be a formidable navigation challenge—and the structured format of the information may result in even longer records. If a simple office follow-up visit can produce 5 pages of data, as it did on one of the authors’ visits recently, the volume of data in a hospital encounter is likely to be formidable. Recent studies of EHR data by the Center for Medicare & Medicaid Services (CMS) revealed that a considerable amount of patient data is cut and pasted from previous visits, which not only increases the length of the record (but with redundant data), but also challenges its potential accuracy and usefulness (Lowe, 2012).
As nurses become more comfortable with smart phones, windows based computers, and multiple applications (“apps”), are the antiquated designs of most EHRs, which continue to be based on spreadsheets designed initially for banking and similar businesses, interfering with full adoption? When nurses can search the Internet on smart phones or “Ask Siri” for directions, how can they be expected to navigate the silo-based, “old fashioned” interface of the typical EHR? Perhaps customized “home pages” with robust search functions and familiar windows- based interfaces might encourage nurses and other providers to use the system as an information source prior to assuming care of patients.
If an EHR eliminated discipline-based silos and instead was truly patient centered, would this help nurses and other providers build on the care of other professions? For example, must vital signs be taken by nursing staff at the start of each shift if physical therapy or respiratory therapy just recorded them as part of their activities? Perhaps the EHR system could prompt for nursing staff to assess vital signs if they had not been previously assessed by any professional within a specified period of time.
Verbal handoffs add subjective communication elements and provide a social context in which information is shared that may be missing if the information is gleaned from reading a health record. Is this good or bad? Does the oncoming nurse need to know that the patient’s family clashed with the previous nurse? Does this information serve as a distraction or support quality nursing care? Findings from separate ongoing studies by two of the authors (Benham-Hutchins and Carrington) indicate that nurses want specific information from the outgoing nurse, such as the last pain medication, but also verify this information in the EHR. Participants in these studies reported that a lot could change during the shift change – such as new orders or a change in patient status – so they double checked the information provided during the verbal report. Perhaps this redundancy is a necessary safety precaution, but it is also possible that the shift reports take too long and include non-relevant data.
One of the authors (Carrington) has characterized the electronic health record as a communication system; but both anecdotes and a growing body of research raise questions arise about its effectiveness:
- Is critical information as accessible in the EHR as it needs to be to change nurses’ patterns of relying so heavily on less accurate verbal information?
- Coiera (2000) has argued for clarifying the type of communication needed based on common ground or shared knowledge. For example, it could be argued that change of shift report is a time of limited shared knowledge. The nurse who will continue care knows less about the patients than the nurse who has been caring for the patients for the shift. Therefore, according to Coiera’s (2000) Communication Information Continuum, this is an ideal situation for face to face and technology communication methods. Consequently, we need to learn what patient information is best transmitted verbally and what patient information is best transmitted via the EHR.
- Patel and colleagues (2000) found that residents described patients very differently when getting their information from structured or free text in a patient record. We need to understand more about what effects the format of EHR data has on accurate understanding and recall of critical patient information.
- How much copying and pasting of previous nursing observations is occurring? What impact does this have on the use of the EHR and how much nurses trust the data?
- How obvious is the key information in an EHR to an oncoming nurse? How well discriminated are critical data from more routine data? What is the best way to highlight key information?
- Might there be an “executive summary” or clinician-specific webpage that would provide oncoming nurses with the core information they need about each patient with links to more detailed data?
- Finally, do we have the right documentation model, or is too much of what we are documenting based on past practice rather than science?
Nursing informatics researchers have spent considerable time on standardizing nursing language, albeit with somewhat limited success in broad implementation of these languages. The extent to which standardized language affects nurses’ usage of the record as a communication tool remains unclear. Recently, we’ve focused in the United States on increasing “meaningful use” of health data. However, one aspect of meaningful use has so far been largely ignored. The inclusion of health data in a record, no matter how accurate or timely, cannot be meaningful if it is unused. In education we call this “inert” knowledge because it has no impact on behavior. As the title of this paper suggests, you can lead a horse to water. . . . EHRs need to be radically redesigned, not merely to be a legal document of care received, but to facilitate accurate, timely communication of the critical information needed by care givers to provide the best possible care. Although solving this complex problem is well beyond the scope of this paper, we hope that raising awareness of the problem will be a first step toward the research that can lead to its solution.