I often say that electronic health records (EHRs) is like Skynet in the Terminator. I expect to turn around from my screen someday, and Arnold will lift me by my throat saying, “You haff not been doing yuh meaningful use.”
We practice in a time when EHR confounds us by freezing, crashing and chaining us continuously to our work, as we spend evenings and weekends on documentation. For reimbursement purposes, we are instructed to include more and more useless details. As we pay more attention to the “iPatient” than to the real patient, we have confused the map for the territory. Is there a Dr. John Connor out there who will lead the resistance against Skynet EHR?
Perhaps instead it will be a hacker anti-hero that brings the whole system crashing down — an Elliott Alderson from Mr. Robot. Or perhaps a robot will herself take out the abusive, misogynistic corporate head, as Ava does in Ex Machina.
Alas, the machines are ascendant for now. Before we are liberated from robot rule, we will surely accommodate them further. Some advances in EHR do actually help us. For example, our hospital EHR won’t suggest a diagnosis of sepsis (though it could certainly calculate a sepsis-related organ failure assessment (SOFA) score), but once you choose the sepsis order set, it prompts you to repeat the lactic acid in three hours. Choose the order set for hospital-acquired pneumonia, and it prompts you to order antibiotics from the latest Infectious Diseases Society guidelines. However, as noted by National Nurses United:
Although these software programs may be called clinical decision “support,” and the embedded clinical practice requirements may be called “guidelines,” they often function as hard-and-fast rules that override physicians’ professional judgment as well as limit the full professional practice of nurses and other practitioners that care for patients.
No doubt there are techies who are working on making EHR more user- and patient-friendly. As long as we’re going to live inside The Matrix, can our user-friendly overlords at least make it pleasant for us, as Mr. Smith did for Cypher, and let us eat (synthetic glop that tastes like) steak?
The Terminator viewed the world through an interface that projected a variety of data real time in his field of vision — as if he had access to the internet of things. He had two hands free to do his thing, i.e., wreak mayhem. Can’t we have access to data on our patients in a way (a Google Glass-like display?) that allow us two hands free to do our thing? Or perhaps, like Tom Cruise in Minority Report, we could chart in the air in front of us.
In Homo Deus, Noah Yuval Harari contends that that artificial intelligence will soon augment human intelligence. How long before we will be able to instantly upload digital content from the National Library of Medicine? One can imagine a scene taking place in the near future in a remote hospital:
Nurse Neo to family doctor Trinity: Can you fix this subarachnoid hemorrhage?
Dr. Trinity: Not yet. (She speaks into her phone.) Tank, I need a program for the surgical approach to clipping an aneurysm. Hurry. (Her eyelids quiver briefly.) Let’s go.
However, when artificial intelligence merges with human intelligence, what role is left for the human physician? As technology advance, workers in nearly every occupation must fear being replaced by robots. The techies are certainly plotting to replace physicians altogether with robots. A robotic baby harp seal has been used as a stand-in for therapy animals in Japan since 2004. Will androids ever develop to the point where it will become difficult to tell them apart from humans, as in Blade Runner or Westworld? At that point, in what way can human physicians doctor better than androids?
Surely, the worth of the physician of the future will be their ability to carry out the human side of doctoring. A machine cannot comprehend a patient’s and family’s values. A robot without consciousness cannot understand the suffering caused by illness or the existential questions raised by life-threatening conditions. If artificial intelligence can figure out the right words to say, and the words are voiced by a life-like android — can such words console? Artificial intelligence cannot empathize. It cannot feel. On the other hand, might patients respond to a facsimile of empathy? Theodore Twombly falls in love with the operating system of his computer in Her.
The implications for medical education are that we must work to select individuals who are capable of doing these aspects of doctoring. Medical students should be evaluated less by their Step 1 scores and more by their ability to doctor. Future doctors should be interested in their patient’s feelings, their ideas about their illness, the effect that it has on their functioning, and their expectations of their medical encounter. We cannot simply depend on the natural empathetic characters of individuals. We must foster patient-centered skills in our future doctors.
Seiji Yamada is a family physician.
Image credit: Shutterstock.com