Physician As Typist

Abstract

DOI: 10.1200/JCO.2010.29.4504 I stare at the primary care physician’s note in front of me. I have been concerned about our mutual patient’s hypertension. I believe it has been exacerbated by the use of bevacizumab, and I have referred her back for additional management. All I need is an acknowledgment of the problem and a treatment plan. The note that I have received is three pages long and is filled with unrelated laboratory values, scan results, and jumbled-up text. I can’t say I’m surprised. As the healthcare community in my area gradually takes to electronic medical records (EMRs), I am seeing more of such notes. The style is easy to recognize: a brief chief complaint and history of present illness written quickly and with the kind of abbreviations one might expect to see in a teenager’s text messages. Then, a prolonged description of the review of systems and physical examination detailing just as many systems as are necessary to achieve a certain level of billing. This is clearly the point-and-click part of the software program. Next are pages of detailed laboratory values, imaging results, and a problem list that includes specific billing codes—clearly the prepopulated portion of the note. Finally, the assessment and plan: here, we go back to the frenetic, teenage-texting style of a provider who is trying to be a good clinician, communicative, billing-compliant, efficient, and a quick typist, all at the same time. I finally find what I need, tucked away in a brief sentence, listed between hyperlipidemia and smokingcessation discussions. I get other notes, too, from providers that haven’t yet adopted an EMR system. I made a recent referral for a patient with hematuria to a urologist. In a day or two, I received a one-page summary of the problem, including a differential diagnosis, the findings on cystoscopy, and the plan for additional surveillance. It was, really, all I needed. Another oncologic surgeon with whom I share patients always mentions the patient’s profession in the first sentence of the letter. It tells me something about the care that a surgeon who cares to find out such details will provide. However, as our institution transforms from a hybrid to a completely EMR system, these unique styles are likely to disappear. A consensus has emerged in the United States that the health care system is too complex to be managed efficiently by using the current information infrastructure. EMRs are viewed as an essential next step in the digitizing of health care, and multiple governmental and regulatory authorities are nudging health systems in this direction. The American Recovery and Reinvestment Act, for instance, provides both financial incentives and punitive measures designed to accelerate the adoption of EMRs. Yet, in oncology, we know that there is no such thing as a free lunch, no efficacious drug without adverse effects; what, then, are we giving up when we allow EMRs to enter the physicianpatient relationship? A clarification: this is not a neoLuddite critique. Indeed, my own research has relied on the use of large health care data sets and interventions that use electronic order-entry alerts. I am an early and enthusiastic adopter of new devices, from smart phones to e-readers, and I blog regularly. I am part of my institution’s team to help implement our transition to a full EMR system. As an oncologist, I look forward to EMR implementation for the ease of writing chemotherapy orders and for not having to hunt down a paper chart whenever I need a patient’s height or weight. However, as a consumer of technology, I am also aware of its constraints. When I practice note writing on our new EMR, I am struck by how poorly it actually documents my encounter with a patient. This is not uniquely a medical software problem. In You Are Not a Gadget, a recently published critique of software design, computer scientist and Silicon Valley pioneer Jaron Lanier points out that “information underrepresents reality. Life is turned into a database...based on... the belief that computers can presently represent human thought or human relationships. These are things computers cannot currently do,” (italics added). Recall that there are two major narratives associated with the physician-patient encounter. The first is the narrative told by the patient to the physician. The starting point of this narrative is relatively uniform: the complaint that brought the patient in. From here onward, however, the narrative can be remarkably free-flowing and often tangential. To make sense of this free-flowing story, we as providers resort to a second narrative. The physician’s narrative repackages the patient’s tale, but in a format JOURNAL OF CLINICAL ONCOLOGY T H E A R T O F O N C O L O G Y VOLUME 28 NUMBER 24 AUGUST 2