Never Settle: Issues, Impacts and Insights Column

Source: OJNI Volume 18, Number 2
by Judith Effken, PhD, RN, FACMI, FAAN

Occasionally, an article grabs me because it is not only paradigm changing, but also has profound implications for multiple disciplines.  My most recent attention-getter is an article by Michael Turvey (one of my PhD advisors, as it happens) and Sergio Fonseca (2014). In the article, the authors, who have been studying the haptic (purposeful touch) system for many years, synthesize the evidence collected over many years in various fields to conclude that current theories of how this and other sensory systems work are probably erroneous. One of the persistent puzzles about how haptic touch works has been that, unlike visual and auditory systems, in the haptic system there is no medium for transmission, such as air or water. Instead of the discrete sensors transmitting data to the brain for interpretation of what we are touching, they propose that the haptic system’s input is due to changes in the baseline level of tension in muscles, tendons, ligaments and, most surprising perhaps, fascia. Kinsella-Shaw and Turvey’s novel (1992) research on how spiders monitor the web strands for changes in the tension that might be due to prey before attacking their future dinners provides a clue of how this might. Turvey and Fonseca extend this idea using, as their own simplest exemplar, a tent. In a tent, the tension required to keep the tent erected is exerted in different directions by the pegs in the ground, the tent poles, and the stretched material itself.  These forces combine to maintain tent stabilization through isometric tension. Building on these (and other) insights, the authors propose that the haptic system can be understood as a tensegrity system, that is, an array (or map) of tensions that extends over multiple scales, from cell to whole body.  Further, based on a model of Reilly and Sirigu (2008), they argue that the brain contains two types of mappings that relate to motor control and the tensegrity system itself: a control map and a muscle map.  One of the persistent puzzles in medicine is the source of phantom pain post amputation. Turvey and Fonseca argue that, following an amputation, the control map of that limb is lost but the muscle map persists; hence the phantom pain.  This also explains the phantom sensations reported by individuals born without a limb.  Although the physical instantiation of the limb is missing, the instantiation in the brain’s motor map is intact, accounting for the perception of an intact limb.   

Why am I telling you this?  Certainly the paper is fascinating, albeit challenging unless you happen to have studied ecological psychology and/or the haptic system. Although I hope some of you will check the paper out for yourselves, my goal here is to highlight the parallels of the search for a usable model of the haptic system with our own discipline’s struggles to design an appropriate, usable model for an effective, efficient EHR. In my opinion, our current EHRs are using an inappropriate model borrowed from another discipline that simply doesn’t fit healthcare. 

As evidence for this contention, I cite the some of the problems with current systems that informatics experts who are also end users are reporting on various discussion boards. First, discussions about the current and future EHR requirements for meaningful use and the arguments for and against moving to ICD-10 or even skipping 10 to jump to ICD-11 are rampant.  Among the challenges to these two directives cited by clinicians are: ensuring accuracy of the patient record, finding the correct ICD codes for billing, wading through lengthy drop-down menus, and copying and pasting of previous entries no longer accurate. A second and different group of challenges is posed by the desire and need of clinicians and patients to communicate with a wide variety of disciplines (from nurses to dentists or pharmacists) anywhere along the continuum of care.  A third group of challenges emerges from the need for the EHR to accommodate many levels and types of patient data. To support patient care, as well as public health and clinical research, a variety of data elements are sought ranging from DNA to dental x-rays and family and environmental history in easily accessible and understandable textual, numeric and visual formats.  But the amount and variety of shared data and easy communication creates a tension with the simultaneous desire of each professional group to have its own “specialty” system—or, at least, specialty views of the data. Obviously, providing all these data in standardized, easily accessed formats can provide researchers with tons of data for research to improve outcomes, such as that currently being proposed by several federal and private initiatives, but legitimate privacy concerns must be dealt with.   

The second source of evidence I draw on is the persistent resistance of nurses to standardized languages.  My concern is that little seems to have changed in the overall approach. We still have multiple competing nursing ontologies in the United States, although the rest of the world seems to have coalesced around a single ontology with regional adaptations. The lack of agreement, as well as the sometimes obtuse formulations of the languages has turned nurses off to the clear value of a common language for communication of nursing problems and research. Of course there is always the tension of structure and free-form documentation.  Perhaps we need to get back to basics, looking at the purposes of nursing documentation and who needs to be able to understand and interpret what nurses say? 

Why have our health information systems changed so little over the years?  At their core, most of them continue to utilize the kind of relational database approach that other organizations use. Is this too constraining?  Lacking good theory, are we simply using models to build our information systems that don’t really fit the problem and the end users?  We think of the brain as a computer; and somewhat circularly, we conceive of the computer as a brain.  Recent research on the brain suggests that the physical brain may not work like current computers—or perhaps any computer.  Might a model more analogous to the tensegrity system described earlier (i.e., multidimensional, flexible, demonstrating similar functionality across time and multiple scales--e.g., cell to public health, and communicating information in a persistent, actionable format) better fit our needs?  

What might a tensegrity-like EHR model look like? Let’s assume that the EHR is first and foremost a record of an individual’s health from birth to death. Most of the time, the individual is conducting her life with relatively little interaction with the healthcare system. Perhaps the personal health record should be the core EHR. At fairly regular intervals, the individual interacts with the healthcare system (e.g., birthing, preventive and screening exams, vaccinations, eye care, dental care, etc.). Occasionally, the individual may encounter an acute or chronic care facility, or perhaps home or palliative care. The individual also interacts regularly with one or more pharmacies, as well as vitamin stores, school health nurses, etc. So our ideal EHR model must first be a multidimensional, longitudinal record of an individual’s health. Of course the individual is the product of inherited factors from her family, as well as influences of the environment in which she lives. Perhaps that too should be a part of the record.     

Carrington has proposed that the EHR should be considered primarily, not as a documentation tool, but as a communication tool (Carrington & Effken, 2011). What kind of model is needed if we are to design the EHR for communication, not just for recording and retrieval? Who communicates with whom? If the individual is the primary source of her own health data, as well as the recipient of the health care system’s interventions, might it not be appropriate that the primary language used for communication is something understandable by her?  Perhaps computers can actually help with this by translating each professional’s native terminology into more generic, understandable words. As in the tensegrity system, there may need to be many mappings of the individual’s health; mappings for various professions, for the individual, and for researchers if we are to provide maximize both flexibility and usability. Adequately mapping a person’s health will require doing so at multiple scales, from cellular to family to social and environmental.   

Can we tear down our penchant for data silos in favor of more inter-professional communication without making needed data more difficult to locate quickly?  Once I helped an organization design an interdisciplinary admission sheet as part of their effort to streamline and redesign their paper documentation system prior to going electronic. Over the years, after many suggestions from external reviewers, their current documentation system had grown increasingly unwieldy with data being collected for reasons that could no longer be recalled. It wasn’t easy, but the organization was able to integrate, within a single form, the information collected by nurses, physical therapists, dietitians, and others, when appropriate. This resulted in, not only a paper savings, but also a time savings because different professionals were no longer collecting the same information. From a patient’s point of view, I would love to have my personal record available (when I authorized it) to all the professionals who care for me. I detest providing the same information repeatedly. In some cases, such as mother and baby, the records (or parts of the records) of individuals actually need to be integrated if care of either is to be optimized. I am hopeful that the information exchanges that are being designed and implemented in Arizona and many other states will help provide the network level communication among providers will help us realize the safe information sharing that we envision.

Much of my research has a human factors focus. From that perspective, our current health information systems leave a lot to be desired. The commonly used spreadsheet format, derived originally from the banking world, is not well suited to the healthcare environment.  As a result, it is not surprising that end users report that health information technology is time consuming, interruptive, over-constrained by standardization, and error-ridden. A few months ago, I requested the summary of an office visit and received a 6-page “summary” in which much of the data apparently had been copied and pasted from prior visits. A 15 minute visit should not result in a 6-page summary in which potential problems have suddenly morphed into actual problems. During the next visit, I voiced my concerns; and the physician and I agreed we would start with a ‘clean sheet’ to record this visit. We even collaborated on the current note. Still, the previous record remains part of the record. Any model for a successful EHR must emphasize its usability and usefulness—and those criteria must be evaluated and met prior to implementation in a setting as similar to the multiple contexts in which it will be used as possible. 

 The latest motto for The University of Arizona is “Never Settle.” I hope fervently that we in the informatics community never settle for less than optimal health information systems. Never settling doesn’t mean throwing out our current systems without an alternative; but it does mean that we discard outdated or ineffectual models in favor of new models informed by contemporary, multidisciplinary theory and research in many domains. In my view, only radical new models will allow us to design health information systems that provide, not only accurate documentation of an individual’s health over time, but also facilitate the timely, efficient, effective communication and collaboration among the individual and his multiple providers necessary to help him achieve the best possible personal health outcomes while simultaneously contributing valid research data to improve long term public health outcomes.


Carrington, J. M., & Effken, J. A. (2011).  Strengths and limitations of the electronic health record for documenting clinical events.  CIN: Computers, Informatics, Nursing, 29(6), 360-367. doi: 10.1097/NCN.0b013e3181fc4139

Kinsella-Shaw, J. M., & Turvey, M. T. (1992). Haptic perception of object distance in a single strand vibratory web. Perception & Psychophysics, 52(6), 625-638.

Reilly, K. T., & Sirigu, A. (2008). The motor cortex and its role in phantom limb phenomena. Neuroscientist, 14, 195–202. doi:10.1177/1073858407309466

Related Content

Subscribe to HIMSS eNewsletters