Fisheries, forests, & emergency care

In my last essay, I made the case that emergency care in the United States is better classified as a common rather than a public good. Due to the passage of the Emergency Medical Treatment & Labor Act (EMTALA), emergency care became non-excludable and because it is resource constrained, it is rival. Patients are streaming into the Emergency Department to service their emergent and non-emergent health needs. As a result, not only are tangible resources such as physical space, specialty coverage, and medication supplies strained, but intangible resources such as physician and nursing morale are at all time lows. Moreover, burnout rates amongst emergency providers and nurses are amongst the highest in medicine. However, because we effectively consider emergency care as a public good (non-excludable/non-rival) rather than a common (non-excludable/rival), it is susceptible to, and is in fact, falling prey to the tragedy of the commons.

However, the tragedy of the commons is a model, and like all models, it is a compressed representation of the world with foundational assumptions. Only if these foundational principles are met, can the predictions put forth by the model approximate real outcomes.  One of the key assumptions of the tragedy of the commons is the two hundred year old idealization of humans as the rational economic man, homo economicus. Homo economicus as first conjured by the utilitarian philosopher John Stuart Mill, defined man as one who “desires to possess wealth…a deep dislike of work and love of luxuries.” The English economist, William Stanley Jevons further mathematized this caricature by placing utility at the center of economic theory. Thereafter, in the 1920s, Frank Knight endowed the model with perfect knowledge and perfect foresight which enabled him to compare all goods and prices across all times. Homo economicus has become the smallest unit of analysis in economic theory and is characterized as a solitary agent, calculating in his utility, solely driven by competition, and insatiable in his desire to maximize his utility.

However as it is said, the map is not the territory and in recognizing the tragedy of the commons as a model rather than an inevitable law of nature, Nobel Laureate Elinor Ostrom through real-life case studies demonstrated that over the centuries people in many different places have experimented with many forms of institutional arrangements to manage physical commons such as fisheries and irrigation systems. Based on these case studies, she elucidated general design principles that communities have utilized to successfully manage the commons and avert the decimation of these commons. These communities overcame the inevitability of the tragedy of the commons. She broadly outlined design principles that eluded the tragedy. These included principles such locally appropriate rules, collective agreement, monitoring, graduated sanctions, conflict resolution mechanisms, rights to organize, nested enterprises, and clear boundaries.

From my perspective as an emergency physician, some of these principles can be applied to alleviate the burdens of the emergency care commons. Firstly, collective agreements and policies must be created by all stakeholders to more clearly define an emergent versus a non-emergent patient. Secondly, emergent patients must be risk-stratified on a triage score ranging from most emergent to the least. In both scenarios, clear boundaries are needed, so patients can be triaged to the appropriate venue of care. Emergent patients should be identified as such and seen in the emergency department and non-emergent patients should be diverted to non-emergent venues such as the urgent care or outpatient clinics. A classification technology such as machine learning is ideally suited for such a task. I can envision a model that can classify a patient as ‘safe’ for treatment in non-emergency care venues, and secondly, sub-stratify the emergent patients on triage score based on a defined criterion of risk for decompensation. However, effective machine learning predictions requires clearly labeled data and objective criteria to measure the accuracy of its predictions. In other words, clear boundaries are needed between emergent and non-emergent patients and between the various triage classes which currently do not exist.

The same law that transformed emergency care into a common – EMTALA – also stipulates that emergency departments must provide a medical screening examination (MSE) to any individual who comes to the emergency department and requests such an examination, and prohibits hospitals with emergency departments from refusing to examine or treat individuals with an emergency medical condition. However, a MSE is not clearly defined, left up for interpretation, and is consequently ambiguous. Nonetheless, metrics exist such as the risk of re-admission in 30 days that can measure the success of a MSE. If the various stakeholders can define and implement such a metric via shared agreements and policies, then machine learning technology can predict if the patient is safe for discharge to a non-emergent venue after receiving an emergency department screening. Similarly, if the risk for decompensation can also be defined, it can also risk stratify emergent patients so that the more risk-laden patients are evaluated more acutely. 

It is easy to fall into the trap of forgetting that world views at all levels (from the personal to the social and scientific) are woefully compressed representations and do not represent all of the data in the world. Even complex scientific theories are built on layers and layers of assumptions and presumptions. We easily draw illogical conclusions from accurate assumptions and logical conclusions from inaccurate assumptions. The tragedy of the commons is one such model and the foundational homo economicus that it is built on is an incomplete representation of the complexities of human behavior and societies. Consequently, its predictions are not laws of nature nor are they inevitable endpoints. Emergency care does not have to end in the tragedy of a collapsed health care system. I realize that emergency care is immensely complex because patients in the emergency department have layers and layers of social turmoil that perniciously affect their health.  Nevertheless, the current model of emergency does not seem sustainable and destined for ruin. Modeling emergency care as a common rather than a public good, employing design principles created for natural resources such as clear boundaries and collective agreements, and utilizing the results of those design principles to develop technology that can screen, triage, and classify patients can possibly avert the tragedy of the emergency care commons.