The map is not the territory: the missing patient in the electronic medical record

Abstract

Purpose

- This paper aims to, by looking at the electronic medical record (EMR) from three points of view, bring light to the dynamics that are essential and are currently missing in the USA. The traditional paper medical record has worked for physicians, management and patients since the beginning of practice. Yet the development of the EMR did not begin with all the essential elements of the traditional record that were working, but instead shreds out important aspects of the patient.

Design/methodology/approach

- Triangulation between three studies - medical, information technology and management studies.

Findings

- An efficient EMR has to take into consideration more than just one area of study. The dynamics between departments and users of the EMR need an integrated process that includes the necessary pieces of all involved. This hole has not been addressed in academic literature.

Research limitations/implications

- The paper triangulates three areas - medicine, management and information management. Most research on the EMR focuses only on one or two of these areas' concerns. Looking at the three sides of the EMR is important to get a solid understanding of the dynamics that can occur relaying a patient's story through various departments and uses.

Practical implications

- There is a depth, space and volume crucial to the comprehensive nature of medicine. With a perspective or dimension, necessary dialogues can be addressed and more intuitive tacit knowledge from medical expertise can be made available. A prototype, filling the holes of the observed elements in this paper, is possible by using digital objects and including more information than the data of the day. Bringing accountability to the patient, more expertise to the fingertips of the physician and available data for management purposes area are the key ingredients for an effective EMR.

Social implications

- With a comprehensive EMR that works more effectively for those who input the data, the patient's story can be documented with more detailed efficiency. Filling the holes of the observed elements in this paper all support better healthcare and long-term results for the health of society.

Originality/value

- The paper triangulates three areas - medicine, management and information management. Most research on the EMR focuses only on one or two of these areas' concerns. Looking at the three sides of the EMR is important to get a solid understanding of the dynamics that can occur relaying a patient's story through various departments and uses.

1. Introduction

There have been many studies investigating and analyzing the phenomenon of the electronic medical record (EMR) at both individual and organizational levels. The EMR has been introduced and implemented in various ways, but literature reflects that the EMR has yet to fulfill its potential. Looking at the three sides of the EMR is important to get a solid understanding of the dynamics that can occur relaying a patient's story through various departments and uses. This is why the triangulation between the medicine, management and bioinformatics is crucial. Most research on the EMR focuses only on one or two of these areas' concerns. This paper takes into consideration the three essential areas, highlighting the unexpressed story and the dynamic inter-connectedness that is essential in making the EMR comprehensive and valid.

The use of the EMR is less popular in the USA than anywhere else in the world. Most theory identifies that it is the personal beliefs that act as a major determinant of intentions to use a new technology (Venkatesh et al. , 2003). It is important, however, to distinguish how beliefs are formed to understand the story behind the use and implementation. To ensure successful progress and development with this tool (Ilie et al. , 2009), the call for these details has been expressed by numerous authors asking to push the "contextual envelope" for bioinformatics research to deal with unique contextual issues (Chiasson and Davidson, 2004).

The traditional paper medical record has worked for physicians, management and patients since the beginning of practice. Yet the development of the EMR did not begin with all the essential elements of the traditional record that were working, but instead shreds out important aspects of the patient. Digital records were created with a focus on efficiency for the streamlining of information conveyance to the insurance companies, to enhance coding consistency and to potentially improve accuracy of reimbursement. Therefore, the input of the physician and patient, the two individuals essentially interacting and utilizing the EMR the most, has had the least impact into the EMR process and development. In essence, this undermines the physician/patient interaction.

2. The map and territory of medicine

Physicians use expert knowledge in an effort to optimize decision-making. The distillation of information (symptoms and signs) derived from the physician/patient interaction has been difficult to capture for sharing purposes. Although the EMR purports to be fully thorough in this regard, it, in fact, unnecessarily burdens the record with information that is not germane.

Traditionally, a healthcare organization maintains a few experts who can be consulted in situations when needed. Physicians in the same specialty may use different knowledge and approaches to treat patients. Physicians, typically Bayesian probabilists, tap into prior probability to evaluate their hypothesis with regard to their patient. Therefore, there is no unique way to determine the optimal method of medical treatment that applies to a specific patient. All medical decisions are based on hypotheses associated with a probabilistic estimate determined by the physician's experience and knowledge of the available options for treatment (Pizzi, 2009).

Sharing the story, not just the test results of the patient, is essential for adequate treatment of the patient. When Cleveland Clinic introduced their new training program for physicians, they included an educational effort to guide the transition. Many physicians left the didactic training sessions feeling frustrated and with the perspective that the EMR would slow down their patient flow and negatively impact their practice productivity. To help with the transition, they hired six licensed healthcare professionals with a minimum of four years of direct patient care; each expected to ease the physicians into the EMR by helping them enter a few patients a day, increasing the number as they became accustomed to the process.

Although the training for physicians was gradual and the physician was included in the on-going improvement of the EMR, nothing was done with the notes that physicians needed to increase their extraction of the episode. Through this transition format, a subset of each physician's EMR notes was reviewed for documentation completeness and coding appropriateness for compliance purposes. A careful review and audit of these notes produced new and more complete documentation for increased level of coding. This in turn also was able to allow physicians, who looked like they were less productive, to increase their scores, being accurately credited for the care that they delivered.

With the primary focus of the physician shifting to populating the EMR, the story of the patient is no longer coming from the physician, but instead being developed based on the ICD-9/ICD-10 codes of the billing department. The physician becomes a scribe instead of an interactive caregiver, and patient education and care is a secondary-tier activity to the process of data entry. The diagnostic decision-making, in a sense, becomes separate from the population of EMR fields with diagnostic codes. The humanness of the interaction is missing and is no longer unique, essentially defining the physician as a data collector and nothing more.

There are direct and indirect elicitation techniques physicians use to extract knowledge. Direct methods elicit knowledge from experts without analysis such as story-telling, case study and interview, and they depend on the expert's ability to articulate their knowledge (Hudlicka, 1996). Indirect methods such as observation require human intervention and require an internal mental structure or model of the expert's knowledge that, to this point, has not been duplicated by information systems technology (Hudlicka, 1996; Cooke, 1999). Those physician specialties such as endocrinology or psychology that rely heavily on indirect methods are limited by the missing human aspect that the EMR does not express.

The patient's medical story in a traditional paper file includes the epistemological data and the tacit expertise of the physician interpreted through the story of the patient. While the patient transfers much information through verbal exchanges, the effectiveness of this transfer remains a questionable issue (Herschel et al. , 2001).The usability of the information is highly dependent on the verbal communication skills of the patient and is also limited by the ability of the physician to gather, analyze, summarize and interpret the complete story for technological transfer (Klein et al. , 1989).

Therefore, the operating demands of the EMR may interfere with:

the physician/patient interaction;

documentation of the physician/patient interaction; and

tacit nature of medical knowledge.

Another question that concerns efficiency in the practice comes when it is necessary for prior documentation to be carried forward from a previous note. While technology has made it very easy to carry forward volumes of information, it is inappropriate to do this. If all documentation is forwarded, it makes it difficult for other physicians to discern what is truly relevant for the current visit. The EMR must be optimized to convey accurate and efficient documentation of a patient's condition rather than include a cloud of extraneous information that actually blurs the image of the patient.

The EMR is supposed to allow a physician to handle more patients. However, that means the physician needs to continue to be in a caring space satisfaction. The personal touch of the physician is a key ingredient for most patients to be satisfied with the physician/patient interaction. Many patients come to a physician because of the trust and comfort they feel when they are with them. If the physician will now be documenting on a laptop or tablet during the visit, positioning the clinical workstation so that the physician can face the patient and interact while performing the documentation is very important. Patients expect the physician's full attention, and having their back turned to them, or sharing the attention of the computer screen with them may be interpreted as not listening completely (Ossoff et al. , 2010).

It has also been noted that physicians will chose the most accessible information source when needing to face alternative decisions. Therefore, if access to the medical record is difficult, a physician will not use the EMR and rely on other methods of information (Ilie et al. , 2009). Having the paper chart conveniently situated just outside a patient's room becomes the more accessible method of information, while computer terminals may not be available in the physician's immediate examination room. Some researchers also found support for the claim that the more physically accessible an information system is, the less effort is needed to use it (Karahanna and Straub, 1999; Karahanna and Limayem, 2000). Therefore, more physicians may choose the paper chart as the primary clinical information source, despite the potential benefits of the EMR. If physical access to a computer terminal is time-consuming, these systems may be perceived at a relative disadvantage compared to the traditional paper chart (Ilie et al. , 2009).

One of the key connections between the physicians, IT department and hospital management is reflected in the bottom line. By diverting the attention of the physician away from the patient to populate the EMR, concerns about decreased productivity begin to emerge. The organizational structure includes the administrators in charge of overseeing the management of the hospital, who make the primary IT-adoption decisions, and the clinicians in charge of patient care, who make individual IT-usage decisions. In the long run, hospitals may benefit from their technological investments, but these benefits come at a cost to physicians when they need to increase the time spent learning how to use EMR systems and actually using the systems in their daily work. Physicians' productivity scales decrease during times of learning (Berner et al. , 2005; Ilie et al. , 2009).

2.1 Medical discussion

There is a need for a timeline of the patient's disease history for many areas of practice. There is a need for frequency analysis of various factors (high blood pressure, etc.) to have a better understanding of the patient's history. There are unspoken and unseen aspects that need to be recorded (e.g. the patient coming for comfort and information, not diagnosis). There are levels of expertise that must be incorporated into the documentation that stems from the ability of a physician to interpret the patient's story, not just base recommendations on the data and distinguish from ICD-9/ICD-10 codes. The ease at which the physician has access to the appropriate information for expedient diagnosis also remains a barrier.

3. The map and territory of healthcare management

When addressing the pros and cons about the EMR, most focus on the money that will be saved by processing the EMR, the independence that individuals will have in directing their care and the flexibility of those who are using them. The evolving process between management, patient, physician and IT continues to be criticized department by department without a view from a working triangular plan of process.

There are numerous clinical, academic, research and administrative uses for a patient-based EMR, which stores a person's complete medical history in a digitized form. The benefits that are argued include:

patients having more control over monitoring their health;

doctors and hospitals can simultaneously view and share information; and

research labs have new sources of medical information.

Hillestad et al. (2005) estimated that EMR implementation in the USA at a 90 per cent adoption rate could eventually save over $77 billion annually in medical costs. Costs of adoption would average $6.5 billion per year for inpatient systems and $1.1 billion per year for outpatient systems (Hillestad et al. , 2005), so the annual net savings to society could eventually reach over $68 billion a year. Despite these advantages, fewer than 5 per cent of US physicians and hospitals use full-scale EMRs (Cutler et al. , 2005; DesRoches et al. , 2008).

Several economic and financial reasons have been identified for this. Taylor et al. (2005) point out that EMR adoption would save an estimated 404,000 lives through improvements in disease management and prevention, but providers would bear "annual revenue decreases of $51.7 billion for hospitals, $11.6 billion for physician services and $13.5 billion for pharmacies". Second, as a network good with high fixed costs, EMRs' unit value rises exponentially as the number of units or users in the market increases. Thus, there is little benefit to be an early adopter (Taylor et al. , 2005). Third, by sharing electronic patient information with another healthcare provider, the provider's record transforms from the doctor's private property to a quasi-public good. Therefore, doctors and hospitals do better financially if their information sources are more inefficient and customized to the specific institution. Fourth, providers adopting EMR technology capture only 3.1 per cent of the total social benefits. Private insurance companies, Medicare, Medicaid and, eventually, consumers of medical care are the primary beneficiaries, yet these groups do not directly pay for the maintenance of the EMRs (Hillestad et al. , 2005).

The process of introducing the EMR has similar concerns as the credit bureaus did at their inception. It should have a positive effect on the patient, providing an incentive to pay attention to their own behavior if health insurance costs are tied to action (such as smoking or being overweight). By providing a more complete medical record, such as a person's credit report, the quality of medical care should theoretically be higher. There are also similarities in the competitiveness of the hospitals creating a hoarding effect of information, or making it difficult to collaborate with other EMRs. The credit bureaus had a similar issue to contend with, trying to keep individual credit history private. There were substantial hurdles that prevented access to credit information. Small retail shops kept individual customer histories on file and provided short-term credit. Like doctors today, they hoarded information, as there was no incentive to share a customer's credit rating with another store (Richardson et al. , 2010).

The deeper concern comes in the interpretation of the communication. The differentiation between the map and the territory is a result of the denudative communication between the physician and patient. There is no personality in the EMR. In a physical visit, a physician comes to know the person, and the relationship between context and content is revealed. Even though the language of medicine is universal, in communication systems, the sequences resemble stimulus and response rather than cause and effect when observing feedback. When kicking a dog, his behavior is energized by his metabolism, not by the kick. The metabolism may limit his response, but the systems which are in play limit or avail alternatives that are subject to probability (Bateson, 1972).

If the dialogues that occurred between the patient and the physician were to be included in the medical record, accountability of the patient could be increased. Using data of the day does not show the history of an individual's developing symptoms or repetition of avoidance to the physician's suggestion. The effective measure of verbal comments used to assess if the individual is complying with suggested treatment or avoiding the physician suggestions is eliminated and replaced with pure data. Yet, when a physician uses a paper medical record, their notes will include peripheral observations that are necessary to express the tendency of the patient to be compliant.

Conceptual support and knowledge creation are closely linked. To manage non-structured knowledge within practice, both intuition and reason are necessary. Tsoukas (2005) argues "the unarticulated background is an indispensable part of articulated knowledge". The possibility of a clear separation between tacit and explicit knowledge simply does not exist (Akehurst et al. , 2011). Thus, the map and the territory of the patient require expression for the physician to create a valid diagnosis.

3.1 Management discussion

There is a necessary interaction between the physician, patient, management and IT for the EMR to function efficiently and smoothly. If the physician is going to be reimbursed according to code levels, they will pay more attention to the codes to upgrade the episode of care or be diligent in coding according to the demands of the facility for which they work to maintain employable status. They will then pay less attention to the patient's basic medical concerns. Due to the lack of complete recording, the EMR used for personal use by the patient strains any accountability that could positively occur. The patient simply carried information without the history or documentation of their participation level in the physician's suggested protocol.

4. The map and territory of bioinformatics

The EMR is an emerging tool used in medical informatics to computerize medical records and to establish a knowledge-sharing platform among physicians (Hersh, 2009; Herschel et al. , 2001). It is argued that the EMR is a medical record that stores both the physician's tacit knowledge and data items that can be used to make valuable clinical decisions (Herschel et al. , 2001). However, discussions on approaches for eliciting knowledge from information stored in EMR are rare in the literature (Ting et al. , 2011).

Due to the advances and the tremendous growth in the volume of diseases and medical records, researchers have been encouraged to develop new challenging methods to work with the complexity and the growing size of these medical records. To extract their data features, data mining methods provide a variety of computational techniques that are becoming increasingly important viable approaches to efficiently and effectively extracting new information from these massive data sets (medical records). Digital objects can help users (physicians) to effectively navigate, summarize and organize the data information in an appropriate way and transform the processed data into useful information and knowledge to help them to find that for which they are looking.

Determining the most important features of the medical records data set improves the analyzing and extracting methods and provides a meaning to the data set. At the same time it greatly benefits the users (physicians) by increasing their understanding of their own work. It also can create a grounding and reminder for the patient to follow.

The Medical Knowledge Elicitation System (MediKES) has been proposed and developed in a leading healthcare organization in Hong Kong to bridge the gap between a physician's tacit knowledge and data compilation. It captures the physician's decision logic for the medical treatment into a uniform structure and then represents the individual physicians' decision logic visually by the concept mapping technique, which presents the information graphically, making it easy to understand. MediKES is the first attempt to elicit knowledge stored in EMR, for use by physicians to help them in making better clinical judgments. This is achieved by codifying the captured tacit knowledge, allowing them to exploit other physicians' tacit knowledge (Ting et al. , 2011). Through this diagram, the depth of the patient's story is represented in both informational and historical context. Figure 1 shows the architecture framework of the MediKES, which consists of several elements, including an EMR system, and modules for vector formation, case tokenization, rule construction and map construction.

The templates that are used are a key instrument in getting the information correct for the physician, and for the ease of collaboration. Clearly, one template may not fit the needs of patients, inpatients and various physicians. Therefore, the fine-tuning of the EMR will go through changes while more areas are exposed to the process. Not all physicians document the same way; therefore, an element of flexibility is necessary, not only in entering information, but also in receiving the stored information. Physicians and physician offices are generally created in specific routine format. If the template created does not list the organ systems in the manner matching the individual physician's style, then the physician may not want to use the template. The physician may examine more elements or fewer elements than listed on the template or ask different questions for the review of systems. Creating a unique template that is palatable to all is a tremendous challenge. Templates can be helpful when documenting compliance and insurance information, but not necessarily helpful for the physician (Ossoff et al. , 2010).

Using digital objects to reduce the number of data set scans could help to decrease the physician's focus on coding and more on patient care. The physician generally does not think in terms of the ICD-9/ICD-10 code, yet the EMR, to date, has been formulated using the ICD-9/ICD-10 codes as the universal language for the EMR. Insurance companies and hospital management teams may want to know how many individuals have a certain type of disease and how they are treated and the time limit of recovery for bottom-line purposes. However, this information is not relevant to the physician or the patient, who needs to use the EMR for an on-going purpose.

4.1 Bioinformatics discussion

The needs and repetitive behavior of the patient must be better incorporated into the EMR. With the ICD-9/ICD-10 codes dictating the EMR information, these needs are not taken into consideration. Giving priority to a dynamic description, with the ICD-9/ICD-10 code being secondary becomes crucial for the integrity of the physician/patient interaction. The significance of meaning behind the medical results must be a fluid and living piece of the EMR to express this crucial link between physician suggestion and patient results.

5. Conclusion and suggestion

The EMR needs to have topology of three dimensions to fulfill its potential as a universal record that works for physician, patient and healthcare facilities. There is a depth, space and volume crucial to the comprehensive nature of medicine. With a perspective or dimension, necessary dialogues can be addressed and more intuitive tacit knowledge from medical expertise can be made available. A prototype, filling the holes of the observed elements in this paper, is possible by using digital objects and including more information than the data of the day. Bringing accountability to the patient, more expertise to the fingertips of the physician and available data for management purposes area are the key ingredients for an effective EMR.

Peripheral observation and repetitive behavior serve as points of accountability. Unrecorded, the information strips the necessary elements that allow a physician to make a thorough and informed diagnosis and prognosis. Without the complete story, the patient themselves do not have access to information that allows a sensible trace of development and change. For the institution and its management, the lack of essential details can become a legal liability and an efficiency malfunction.

Figure 1.
System architecture of MediKES

[Image omitted: See PDF]

About the authors

Dr Faleh Alshameri is an Associate Professor of computer information systems. He received his Doctor of Philosophy degree in information technology from the Volgenau School of Engineering and Information Technology, George Mason University. His research interests include text mining, image mining, database and systems analysis and design. He has presented and published research papers in refereed journals and international conferences. Faleh Alshameri is the corresponding author and can be contacted at: [email protected]

Dr Debra Hockenberry has been a Visiting Professor at several universities in the United States and abroad. Her doctorate of management is in organizational development and change from Colorado Technical University. She has presented and published research papers in refereed journals and international conferences.

Robert B. Doll, Jr., MD, has been a practicing endocrinologist for 37 years, currently working with Lehigh Valley Health Network. He received his MD at Jefferson Medical College of Thomas Jefferson University. He is board-certified with the American Board of Internal Medicine in internal medicine and endocrinology.

References

References

Akehurst, G., Rueda-Armengot, C., Lopez, S. and Marques, D. ( 2011 ), " Ontological supports of knowledge: knowledge creation and analytical knowledge ", Management Decision, Vol. 49 No. 2, pp. 183 - 194 .

Bateson, G. ( 1972 ), Steps to an Ecology of Mind, University of Chicago Press, Chicago, IL .

Berner, E., Detmer, D. and Simborg, D. ( 2005 ), " Will the wave finally break? A brief view of the adoption of electronic medical records in the United States ", Journal of the American Medical Informatics Association, Vol. 12 No. 1, pp. 3 - 7 .

Chiasson, M. and Davidson, E. ( 2004 ), " Pushing the contextual envelope: developing and diffusing IS theory for health information systems research ", Information and Organization, Vol. 14 No. 3, pp. 155 - 188 .

Cooke, N. ( 1999 ), " Knowledge elicitation ", in Durso, and F.T. (Ed), Handbook of Applied Cognition, Wiley, New York, NY, pp. 479 - 509 .

Cutler, D., Feldman, N. and Horwitz, J. ( 2005 ), " US adoption of computerized physician order entry systems ", Health Affairs, Vol. 24 No. 6, pp. 1654 - 1663 .

DesRoches, C., Campbell, E., Rao, S., Donelan, K., Ferris, T., Jha, A., Kaushal, R., Levy, D., Rosenbaum, S., Shields, A. and Blumenthal, D. ( 2008 ), " Electronic health records in ambulatory care - a national survey of physicians ", New England Journal of Medicine, Vol. 359 No. 1, pp. 50 - 60 .

Herschel, R., Nemati, H. and Steiger, D. ( 2001 ), " Tacit to explicit knowledge conversion: knowledge exchange protocols ", Journal of Knowledge Management, Vol. 5 No. 1, pp. 107 - 116 .

Hersh, W. ( 2009 ), " A stimulus to define informatics and health information technology ", BMC Medical Informatics and Decision Making, Vol. 9 No. 1, p. - .

Hillestad, R., Bigelow, J., Bower, A., Girosi, F., Meili, R., Scoville, R. and Taylor, R. ( 2005 ), " Can electronic medical record systems transform health care? Potential health benefits, savings, and costs ", Health Affairs, Vol. 24 No. 5, pp. 1103 - 1118 .

Hudlicka, E. ( 1996 ), " Requirements elicitation with indirect knowledge elicitation techniques: comparison of three methods ", Proceedings of the 2nd International Conference on Requirements Engineering in Colorado Springs, CO, pp. 4 - 11 .

Ilie, V., Van Slyke, C., Parikh, M. and Courtney, J. ( 2009 ), " Paper versus electronic medical records: the effects of access on physicians' decisions to use complex information technologies ", Decision Sciences, Vol. 40 No. 2, pp. 213 - 241 .

Karahanna, E. and Limayem, M. ( 2000 ), " E-mail and V-mail usage: generalizing across technologies ", Journal of Organizational Computing and Electronic Commerce, Vol. 10 No. 1, pp. 49 - 66 .

Karahanna, E. and Straub, D. ( 1999 ), " The psychological origins of perceived usefulness and ease-of-use ", Information & Management, Vol. 35 No. 4, pp. 237 - 250 .

Klein, G., Calderwood, R. and MacGregor, D. ( 1989 ), " Critical decision method for eliciting knowledge ", IEEE Transactions on Systems, Man and Cybernetics, Vol. 19 No. 3, pp. 462 - 472 .

Ossoff, R., Thomason, C. and Appleton, J. ( 2010 ), " Challenges with the electronic medical record ", Journal of Health Care Compliance, Vol. 13 No. 2, pp. 51 - 54 .

Pizzi, N. ( 2009 ), " Information processing in biomedical applications ", in Bargiela,, A., Pedrycz, and W. (Eds), Human-Centric Information Processing Through Granular Modelling, Springer, Berlin, Heidelberg, Vol. 182, pp. 289 - 311 .

Richardson, C., Hall, M. and Madjd-Sadjadi, Z. ( 2010 ), " Lessons from credit bureaus for improving the market for electronic medical records ", The Journal of Consumer Affairs, Vol. 44 No. 3, pp. 546 - 556 .

Taylor, R., Bower, A., Girosi, F., Bigelow, J., Fonkych, K. and Hillestad, R. ( 2005 ), " Promoting health information technology: is there a case for more-aggressive government action? ", Health Affairs, Vol. 24 No. 5, pp. 1234 - 1245 .

Ting, S. and Wang, W. ( 2011 ), " Knowledge elicitation approach in enhancing tacit knowledge sharing ", Industrial Management & Data Systems, Vol. 111 No. 7, pp. 1039 - 1064 .

Tsoukas, H. ( 2005 ), Complex Knowledge: Studies in Organizational Epistemology, Oxford University Press, London .

Venkatesh, V., Morris, M., Davis, G.B. and Davis, F. ( 2003 ), " User acceptance of information technology: toward a unified view ", MIS Quarterly, Vol. 27 No. 3, pp. 425 - 478 .