I have lurked on this list for a year or so, but this disturbs me a little. AMNews usually has pretty good articles. It is not very insightful of the AMNews to accept such frightened text from someone as erudite as this physician who has NOT YET, according to his commentary, actually gone 'paperless'.
Mark Koch, MD Director of Informatics JPS Family Medicine Residency Ft. Worth, Texas, 76104 -----Original Message----- From: david derauf [mailto:[EMAIL PROTECTED]] Sent: Tuesday, February 11, 2003 12:46 PM To: [EMAIL PROTECTED] Subject: Paperless medical record not all it's cracked up to be Commentary. By Edmond Blum, MD, AMNews contributor. Feb. 17, 2003. Additional information New York City's Health and Hospitals Corp., the nation's largest municipal hospital system, is going paperless. In doing so it is going with the flow. The electronic medical record is touted as a cure for all that ails the paper record: illegibility, inaccuracy, inaccessibility and incompleteness. Beyond that it offers decision-support systems and facilitates population medicine via its databases. It has become an essential technology. Yet as a primary care physician in one of HHC's largest facilities, I have serious misgivings. So do many of my colleagues. The EMR has become an essential technology for health care and can be an immensely valuable tool, but going paperless is crossing the Rubicon, a critical boundary, a line which may be that of diminishing returns. The problems have to do with the limitations of how data can be entered, and how physician entry of those data may inadvertently inhibit patient care. As the Institute of Medicine has noted, this is the greatest single challenge that has consistently confronted system developers. In an EMR, data may be entered in free-text form, in coded form or in a form that combines both. Trade-offs between the use of codes and free text exist, but the more precisely the physician tries to represent a complex description in code, the slower and more costly the coding effort becomes. In free text, physicians can express complex information at the speed of thought. This is most evident in general internal medicine, in which the bulk of the patient record must remain as free text: history, examination findings, progress notes, discharge summaries, assessment and plan. This text, which conveys the interpretive and narrative parts of the medical record -- including the indispensable nuances and ambiguities -- is essential for diagnosis and management. The information it contains can be expressed only in natural language, not by structured databases and coded options. Without it, there is nothing to connect the dots of coded data into a coherent and meaningful picture. This text must be entered into the EMR via one of two general mechanisms: transcription of dictated or written notes, or direct data entry by the physician. As the Institute of Medicine has noted, "the latter approach [typed input by the physician] has generally failed because busy clinicians reject it." The expression "busy clinicians" evokes many things, especially for primary care physicians: a multitude of tasks to be performed, overcrowded schedules, unpredictable patient flow, severe time constraints and inadequate support staff; in other words, a taxing practice environment, one in which physicians can hardly assume the role of medical transcriptionists. Yet direct data entry by physicians may be mandatory, especially in the employment setting. When that happens, some major and hidden pitfalls in the paperless road will surface. "Busy clinicians" cannot defer their data entry to the end of the clinical encounter. They will, in fact, attempt to perform this task during the time spent with the patient. What will result is a scenario of two competing interfaces: physician-computer and physician-patient (in computer jargon, an interface is the place of interaction between computer and user). As cognitive psychologists have shown, when a person performs two tasks simultaneously, neither of which is automatic, the performance of each task degrades compared with the performance of each task alone because the tasks compete for attention -- a phenomenon that is called "interference." The paper record suffers little from interference because handwriting is such an automatic activity, and the paper itself blends almost seamlessly into the physician-patient interface -- there is only one interface. Thus, computer input requires undivided attention, and must remain separate from physician-patient interaction, which also calls for undivided attention. It therefore represents a bloc of time that must either be added to the time spent on patient care or taken away from it. If the former, the duration of the patient encounter will be substantially increased, an untenable option for "busy clinicians." If the latter, the time given to patient care will be substantially decreased. If we think of time as a resource, then computer tasks and patient care will be competing for that very limited resource to the detriment of both. Undependable hardware, cumbersome software, and deficient physician typing and computer skills will, of course, compound the problem. The use of computers in the intimate world of the physician-patient encounter can be extremely intrusive and needs critical evaluation. Computers, as pointed out previously, tend to represent a competing interface. They also tend to inhibit the asking of facilitating questions, or sensitive reflection, or the logic of the diagnostic process, or eye contact and the sense of personal attention that comes with it, or the formation of the physician-patient relationship -- the interpersonal bond that is so essential to the physician's understanding and effectiveness. All this can degrade the physician's clinical and interpersonal skills, thus impacting negatively on the raw data and meaningful information entered into the medical record -- the very opposite of what the EMR is meant to do. In computer jargon this is known as GIGO (garbage in, garbage out), a famous computer axiom meaning that if invalid data are entered into a system the resulting output will also be invalid. As for the physician-patient relationship, which is basically one of trust, how can it survive if the physician is so focused on the computer that he appears to the patient like the character so aptly described in Dickens' Martin Chuzzlewit: "Affection beaming in one eye; calculation shining out of the other." The goal of the paperless EMR is not simply what we want it to do, but how it can contribute to our overall goal: the quality of patient care. When the EMR impairs the quality of documentation and that of the physician-patient relationship, an invaluable tool turns into a dismal failure, and a panacea for what ails the paper record into a cure that is worse than the disease. Dr. Blum is general internist at a large public hospital in
