One late evening on pediatrics call, a frantic young couple brought in their few weeks old baby. She had spiked a fever which refused to go down and was fussier than normal. The cause of her symptoms could have been anything — at best, a mild respiratory infection, in which case we would simply watch her and manage her symptoms, but at worst, it could be meningitis, an infection attacking the membrane covering her spinal cord and brain. It’s a grave condition that would be fatal if left undiagnosed, but diagnosing it meant doing a lumbar puncture, an extremely invasive procedure for anyone, let alone a baby, involving inserting a needle into the spinal column.
The situation was a classic dilemma: to tap or not to tap. After an hour of deliberation with the residents, I found myself holding down a writhing and crying infant by her legs and torso and while resident stuck a needle into her lumbar spine. After a few attempts, we finally drew enough spinal fluid for analysis.
The next morning, the lab results came back — she didn’t have meningitis after all. Did we make the wrong decision by putting the baby and her parents through needless suffering and risk? When science fails to give us right answers, we’re forced to resort to our instincts, to dig deep into our clinical memory banks for guidance and have faith that our training was sufficient. And that ambiguity and nuance, as our attending at the time told us, capture the art of medicine and is what drew me into this field.
And yet, the art of medicine is slowly being pushed out. Much of clinical decision making, at least the simple kind, already uses personal health data. Go into any family doctor practice these days, and you can find a provider plugging your information into an online calculator to determine what tests to do and what drugs to prescribe. Are you an older male with a history of high blood pressure, high cholesterol, and smoking? According to the algorithm, other patients with similar health profiles have benefited from taking a daily baby aspirin. Evidence-based medicine has eclipsed medical dogma — now we do what is scientifically sound instead of what we feel is right.
But what about more complicated decisions? At a conference I attended last month, a computational researcher and his colleagues presented the results of their latest machine-learning project. Over the past year, they designed a remarkable and elegant algorithm that could distinguish thousands of normal tissue from cancerous tissue of any type with up to 96 percent accuracy. The group plans on applying the same strategy to blood samples, opening up the possibility that a simple blood draw could lead to earlier and less invasive cancer diagnoses. Even the most talented pathologist in world would be hard pressed to recapitulate such efficiency.
Medicine is not immune to disruption. And with the rise of computers, we as physicians may see our autonomy and creativity, some of the very reasons why we decided to become doctors, marginalized. You’d think that the dim prospect of computers taking over medicine would frighten me — that after 26 years of schooling and endless loans, I’d find myself unemployed and my skills obsolete. But computers will never be infallible and there will always been a need for human oversight — after all, we humans built these machines.
What we should be more concerned with is the depersonalization of medicine. Stripped of human interpretation and judgment, medicine becomes cold and distant: Face time supplanted by screen time, probes and wires instead of a human hand and ear. Some resident physicians already spend 40 percent of their time in front of a computer screen and only 12 percent of their time with direct patient care. Will those numbers eventually turn to 100 and 0?
My generation of physicians may one day be practicing more as overseers rather than decision makers. But along with this change comes the standardization of care and improvement in outcomes by closing the gaps of inefficiency and erasing human bias. For our patients, that’s a good thing. And ultimately, isn’t that what matters most?
Details in this article have been omitted or changed to protect the identity of those involved.
Steven Zhang is a medical student who blogs at Scope, where this article originally appeared.
Image credit: Shutterstock.com
