I was recently speaking with a friend of mine, a pulmonologist at a large academic medical center in the Midwest, about his COVID-19 experience. I was especially interested, in the context of iterative experimentation, to learn how his hospital was working on improving the care of COVID-19 patients, especially those in the ICU, which he oversees.
It’s real problem, he said. On the one hand, there are specific initiatives he’s trying to evaluate, in a classic, controlled fashion, so he can figure out if the intervention is effective and should become part of the standard of care.
That’s the goal.
In reality, however, here’s what he says is actually happening: most of the front-line doctors are hearing about the very latest approaches, generally from social media (such as Twitter or medical podcasts), and are trying to immediately apply these methods to their patients. As a result, the care patients receive depends (to some degree) on the specific physician involved, as well as the extent to which that physician has been influenced by other opinionated doctors.
At a recent Boston innovation conference (discussed here), Dr. Paul Biddinger, an emergency medicine physician who leads emergency preparedness at the Massachusetts General Hospital, made a remarkably similar observation. He praised the “unprecedented information sharing” associated with the COVID crisis. But he also expressed concern about the “practicing by anecdote,” and more generally the “temptation to fall off what have been the time-proven methodologies of science.”
The tension here captures an especially distinctive aspect of American medicine, a characteristic beloved by some and lamented by others.
Intrinsically, doctors value their independence, and the opportunity and obligation to be masters of their own domain. This is how U.S. medicine generally works: once doctors are trained and licensed, they more or less can treat patients as they see fit, within a fairly wide band of professional expectations. Care generally isn’t cookie-cutter, and doctors are typically free to offer a range of potential approaches, constrained mostly by what the patient’s insurance is likely to accommodate or reject (and by the hassle of navigating that process).
To be sure, physicians are supposed to stay up to date, and not commit blatant malpractice like treat a strep infection with a blood thinner, say, but there’s a lot of room for individual interpretation and improvisation, which many doctors have embraced, believing it acknowledges the individualized nature of each patient/doctor encounter. Call this the “libertarian” view of medicine.
Contrast this approach with what might be called the “systemic” or “operations” view of healthcare, described eloquently nearly a decade ago by Dr. Atul Gawande in his classic “Cheesecake Factory” New Yorker essay. In this view, while doctors are the ones actually laying on hands, the expectation as well as the mindset is that these providers should be more focused on consistency than individuality; care should adhere clearly to “best practice” guidelines, and while modest customization is permitted, significant deviations should be evaluated deliberately and programmatically, not ad-hoc.
In this view, much of what’s wrong with American medicine is precisely the individual physician’s insistence on acting, well, individually, leading to dramatic variances in care, and often, it is said, the suboptimal treatment of patients. Medicine practiced in this way tends to view physicians as providing the “customer service,” but adhering to master pathways and algorithms. Not surprisingly, many (but not all) doctors in systems like this tend to view themselves as interchangeable cogs and data entry clerks, rather than empowered inquisitive physicians in the Judah Folkman tradition.
The uncomfortable question raised by advocates of a systems approach is whether it promises better care of patients but lacks traction because it challenges the self-esteem of doctors.
The COVID-19 crisis highlights the dilemma. In the unlucky event you were hospitalized with COVID-19, would you want your care driven by a standardized algorithm, rigorously followed, or would you want your doctor to improvise, and potentially apply the latest and greatest – although it might not be so great, and perhaps may reflect little data and instead just the strongly-held opinion of a persuasive, social-media-savvy clinician? Or, it might incorporate a valuable emerging insight, the sort of adjustment that could take a relatively long time to incorporate into a standardized protocol.
On the one hand, most physicians wouldn’t want to practice, and don’t like practicing, in a climate of rigid pathways and defined approaches – that’s not why most doctors went into medicine. Moreover, the physicians who self-select for environments with great autonomy may be more likely to have the imagination or creative insight that the clinicians in more rigid systems lack.
But it’s also possible – especially as more care becomes templated and algorithmic – that the consistency and transparency provided by these systems offers not only a higher average level of care, but also enables a degree of continuous, iterative improvement in care that is far more difficult to achieve in less structured environments.
One worry is that in such a command-and-control framework, you’re potentially relying on centralized planning – on detached, “enlightened” supervisors to suggest modifications, rather than benefiting from the experience and insights of front-line providers, doctors whose imagination and creativity could well be crushed in a system that valorizes provider conformity, and expects providers to perform their job consistently and empathetically, but looks elsewhere for insight and originality.
Historically, the physicians so many of us have looked up to are the brilliant, iconoclastic individuals who figured out on their own a new way to do something. It will be interesting to see if there are future heroes who distinguish themselves by their ability to apply deliberate, iterative experimentation to raise the standards of a system.