in the Shadows, between the Lines

It has been estimated that an emergency physician makes 10,000 decisions in an 8 hour shift. I am not sure how that statistic was derived, but as an emergency physician I can experientially attest to the feeling of decision fatigue during many shifts. Many and the most salient of those decisions revolve around what tests to order, not to order, and is the patient safe to be discharged? Moreover, they occur in an environment shrouded with irreducible uncertainty, under the specter of unnoticed risks, in the cauldron of pain and fear, under the impositions of top-down metrics, and often during zombie hours. In such an environment, cognitive scaffolds that open up the space of affordances and ease decision-making burdens are necessary and needed. Due to their position in the delivery of care (future essay), bedside technologies that augment the physical exam offer a unique opportunity to help alleviate the decision burden. 

Two such technologies that have become staples of clinical medicine in general and emergency medicine in particular are the electrocardiogram (ECG) and the point of care ultrasound (POCUS). The pattern of ‘lines’ on a twelve-lead ECG and the ‘shadows’ on POCUS can not only refine the differential diagnosis, but can also serve as key nodes in the decision matrix. Both technologies also have the advantage of being easily operationalized, non-invasive, relatively portable, and “radiation-free.” Over the past twenty years, ‘reading’ ECGs has become an essential skill for emergency physicians and performing ultrasounds is increasingly pushed in training programs. Despite this push, POCUS continues to be underutilized in the community and ECG technology has yet to reach its full potential as a screening and de-risking tool for an emergency physician. 

Nearly eight million patients present to the emergency department (ED) with chest pain every year, accounting for the second most common cause of ED visits amongst adults. Most of these patients will receive a screening ECG within 10 minutes of presentation. The vast majority of (90-95%) patients will not have an emergent or dangerous reason for their presentation. So the task of the ED physician is to identify the minority of presentations that are dangerous and emergent. Unfortunately, ECG signals are insensitive for the majority of causes of heart diseases, and therefore, they only open a narrow affordance space. Furthermore, patterns are often ambiguous and open to interpretation. “Is this a 1mm ST elevation in two contiguous leads? Is that a “p” wave in lead 2? What was the Sgarbossa criteria? Is that a posterior wall myocardial infarction (MI)?” Sensitively reading an EKG requires an attentive eye and an attuned physician. Inferring diagnoses from EKGs impose an inordinate memory burden on the physician. Attunement, attention, memory, calculation are resources that are limited to a cognitively and affectively burdened ED physician.  

In order for ECG technology to reach its full potential as a screening and de-risking modality for the ED physician, it will not only have to decrease the cognitive burdens on an ED physician, but also be used as a reliable screen-out tool for the vast majority of patients who presents to the ED with chest pain. In the current ED workflow, ECG ordering happens reflexively during triage. For example, a presentation pertaining to ‘chest pain’ triggers an ECG order at triage. However, unless the test is unequivocally diagnostic, additional testing such as cardiac biomarker testing (next essay) and imaging studies are ordered without the benefit of a history and a physical exam. This eliminates the ECG as a bedside augment of the history and physical, and more importantly, leads to false positive screens and externalities stemming from delays in patient disposition (future essay). For ECG technology to be maximally effective, it requires software upgrades to improve classification of ambiguous patterns, the recognition of subtle signals, the identification of complex patterns, and the development of novel patterns that can reliably screen-out low risk patients (future essay).

The POCUS is another bedside diagnostic modality that has the potential to improve emergency care. It is an ideal bedside tool that can augment the history and physical exam. In doing so, it can decrease utilization of more invasive and time-consuming diagnostic modalities such as CT scans (next essay). It can also reduce the negative sequelae of incidentalomas. Despite these advantages, POCUS has yet to gain broad adoption outside of training programs and academic medical centers. This is because the current iteration of corded and wheeled ultrasound machines with clunky EHR interfaces are misfit to the mobility and time demands on an emergency physician. Additionally, POCUS is plagued by over-reliance on the operator – the emergency physician. It requires resources – time and operational proficiency – for relatively low-level tasks such as image acquisition, identification, and quantification.

Broad adoption of POCUS technology will require hardware and software improvements. Adoption will require ultrasounds to be increasingly portable and mobile to match the mobility of the ED physician, the space constraints of the ED, and location of patients in the ED. As currently implemented, the ED physician spends the majority of time moving the machine to the patient, acquiring images, and transferring those images to the electronic health record. Additionally, ultrasound will need to be embedded with computer vision and artificial intelligence, so lower-level tasks such as image-acquisition, labeling, and measurements are automatized. This will decrease the task-burden on the ED physicians and allow them to attend to higher-level cognitive tasks such as interpretation and synthesis. Without such hardware and software augments, POCUS technology will continue to be underutilized and ED physicians will continue to rely on more invasive, more time-consuming, and less risk-free technologies such as CT scans (future essay).

Bedside technologies that extend perception and increase the space of affordances have existed since the advent of medicine. Their privileged position – at the point of care – offers unique advantages to utilize these technologies to refine the differential diagnosis. For example, the stethoscope – invented in the early 19th century – extended the capacity for auscultation and increased the diagnostic capabilities of the physician. More recently, the development of the ECG in the 1920s and the application of ultrasound in medicine in the 1950s have gained prominence and have become staples of clinical medicine in general and emergency medicine in particular. However, maximizing their potential as tools for screening and de-risking will require hardware and software upgrades that match the demands, constraints, requirements, and goals of the emergency department and the emergency physician.