With the adoption of electronic medical records (EMRs), healthcare systems in the United States are moving rapidly toward digitization. This movement is being driven by a variety of influences, including regulatory and governmental authorities, with the aim to increase accessibility to information, reduce errors, and improve compliance with quality and performance measures. The overarching objective is to allow physicians to gain control over the ever-increasing complexity of the current healthcare system. Indeed, the American Recovery and Reinvestment Act provides for both financial incentives and punitive measures that are designed to accelerate the adoption of EMRs.1
I personally welcome the introduction of EMRs into healthcare. My own research program has benefited from the advent of information technology. Our group has utilized large healthcare data sets to conduct analyses that would not be possible with clinical trials alone and has used electronic order entry alerts to improve patient outcomes.2,3
From a clinical oncology perspective, writing chemotherapy orders is perceived to be safer and easier in EMRs, and most of us appreciate not having to hunt down paper charts for writing orders or for finding patient heights and weights.Note-Writing in the EMR
Having transitioned to an EMR at my institution earlier this year, however, I find note-writing to be the EMR’s major shortcoming, as it tends to disrupt many of the intended uses of the traditional patient-progress note. The progress note has a long and storied tradition in healthcare, and it serves many purposes. The first is documentation, as it captures essentials of the visit and reminds physicians of their discussions with the patient. A second purpose is communication, with versions of progress notes (“letters”) sent to other providers for either information or to request input. A third—and more recent—purpose is compliance, with notes documenting mandatory preventive measures (for example, thromboprophylaxis in the inpatient setting or influenza vaccination in the outpatient setting) that can later be queried for quality-of-care audits.
The EMR first causes disruption outside of the progress note itself. Patients do not feel as engaged with their physicians, who are looking at the computer screen rather than the live patient in front of them. While many of my patients are forgiving of my focus on the screen, I find it difficult to practice mindfulness when multiple items on the screen (eg, laboratory results, flagged messages, imaging results) are all vying for my attention and competing with the actual patient. One can accept disruptions if this leads to improvement over current capabilities. However, the primary uses of the progress note—documentation and communication—are disrupted as well, because EMR notes rely heavily on templates and pointand- click software applications. Although convenient for note-writing, such choices made by software designers (and not physicians) force us to reduce our patients to the “boxes” that have been provided by the software designer. For instance, if I wanted to document that the patient at a recent visit was anxious, but the choices offered by the program are “sleeping,” “active,” “cooperative,” or “distressed,” it might be easier to pick “distressed” even though it doesn’t accurately convey the situation. This is not the physician failing to write a good note; it is a software program deficiency that narrows available choices for physicians and leads to an inaccurate capture of reality.
My concerns about this issue are affirmed by a recent critique of current software design by one of Silicon Valley’s own pioneers. In You Are Not a Gadget , computer scientist Jaron Lanier states that information underrepresents reality and that current software does not have the capability to capture the nuances of human interactions (let alone the complexities of physician-patient interactions).4
Even worse for future EMR development, Lanier describes the process known as software lockin, where software choices made early in program development become entrenched for all future iterations of the program.4
In other words, the choices made in currently existing programs, if accepted by end-users (physicians) as a sacrifice necessary for transitioning to EMR, will be kept for the next several generations of medical software. As Lanier points out, software lock-in “removes design options based on what is easiest to program, what is politically feasible, what is fashionable, or what is created by chance…[and] narrows the ideas that it immortalizes, by cutting away the unfathomable penumbra of meaning that distinguishes a word in natural language from a command in a computer program.”4Take Ownership of the Note