How should doctors make decisions?
In the last post, we reviewed the conceptual difficulty of applying risk estimates derived from population studies to individuals, and I proposed that medical decisions should NOT be modeled on the following scheme:
- Step 1: weigh the probability of future pluses and minuses.
- Step 2: make a decision on the basis of “probability of pluses > probability of minuses.”
Yet that scheme implicitly serves as a rationale for pay-for-performance measures and justifies legal impositions on the content of informed consent procedures (e.g., precise disclosure of risks, probability of benefit, risks of alternative treatments, etc.).
Today, I will add a some perspectives that I think buttress my case.
Disclaimer: This discussion is not meant to entice you to violate the standard of care, to forego the established procedures for informed consent, to incite you into civil disobedience, or to put your medical license at risk!
Let’s start with two anecdotes:
A few years ago, the New York Times detailed the extraordinary story of Michael DeBakey’s aortic repair operation. At age 97, the legendary heart surgeon suffered a dissection of the ascending aorta which he diagnosed himself. He wavered about the decision to have it treated and, a few weeks later, ended up with multiple organ failure, became unconscious, and was essentially at death’s door.
Now, it is clear from the details of the story that if the usual procedures to gauge risk versus benefit and to obtain informed consent had been followed, Michael DeBakey would have died unoperated. In the midst of a very hostile dispute among the medical staff and under pressure from the attending surgeon, the hospital’s ethics committee allowed the operation to proceed at the 11th hour. DeBakey survived against all odds. He returned to a productive and enjoyable life, clearly thankful to his caregivers.
One may argue about the ethics of granting exceptions to the standard of care for famous physicians. One may argue about the obvious violation of DeBakey’s own advanced directives. One may also argue about the economic wisdom of embarking on a costly surgery for a very old patient. But it remains that the decision made by the surgeons and the ethics committee flew in the face of everything we know about risk-benefit calculations and yet, as far as the individual patient DeBakey was concerned, it was certainly the right decision.
What is also striking in this story is how the deep personal connection between the medical team and the patient clearly influenced the ultimate outcome. That is by no means unique to this case. Intangible human factors undoubtedly play important roles in all medical encounters, but how can these be factored into an explicit risk-benefit analysis?
The other anecdote is given by the late Geoffrey Hartzler in a 3-minute video interview. Hartzler, who was a pioneer interventional cardiologist in the early days of the procedure relates how, in 1980, he decided to use coronary angioplasty to treat an acute myocardial infarction. This had never been done before and Hartzler makes it clear that his decision, which he believed to be logical, flew in the face of the logic of established wisdom.
Hartzler confronted uncertainty with a bold move for which risk-benefit calculation would have been a farce. And his bold move was not part of any investigational protocol. He put his own career at risk, I imagine. But there is no doubt that thousands of lives have been saved as a result of his disregard for the diligent addition of pluses and minuses.
Countless medical innovations have emerged in a similar fashion, and countless undeniably excellent medical outcomes proceed from “uncalculable” medical decisions, wholly distinct from the linear risk analysis that the health care system is trying to impose on doctors.
From these two anecdotes, the idea seems to emerge that the physician’s role is to assume the risk, not to to engage in a dispassionate risk analysis and invite the patient to decide.¹
Importantly, patients don’t seem to care about risk analysis either.
Dr. Saurabh Jha, an academic radiologist and a witty and insightful essayist made the following observation in regards to why the public is so taken with Dr. Oz, even though the products he promotes are so obviously quackery:
[P]eople don’t listen to Dr. Oz to be “informed.” People aren’t interested in numbers, uncertainty, that more research is needed, that science is a provisional assumption. People want certitude and solutions. This is why chicanery is fertile. God may be dead, but prophets are still alive and kicking.
Science can handle Oz. Scientists can’t seem to handle that the public may prefer Oz to science. Be that as it may. But then the fault, dear Brutus, lies with us, not Oz.
I am not suggesting that doctors turn into quacks, but simply that Dr. Jha’s analysis applies equally well to people’s attitude toward informed consent and shared decision-making.
Someone once told me of of a study about shared decision-making in cancer treatment. In healthy people, the split was 80/20 between those who wanted more say in medical decisions and those who preferred to defer largely or entirely to the doctors. In patients who had personally experienced cancer, however, it was the opposite. Autonomy was no longer so prized²
The informed consent “procedure” makes me cringe. Telling patients in the midst of an acute illness that a cardiac catheterization carries a 1% risk of death, a 5% risk of bleeding, or whatever the case may be, makes a travesty of concern for the patient. And the song-and-dance isn’t any more convincing with elective procedures.
In the face of uncertainty, patients look to doctors for answers, not scientific odds. Imposing this quantified notion of informed consent on patients may be more paternalistic than the paternalism it claims to supplant.
Herb Fred, a master clinician, educator, and writer, once summed up optimal medical decision-making as Mutton’s law: “Know what to do and when to do it.” On teaching rounds, he still occasionally quips at indecisive residents to remind them that “M.D. stands for make decisions.”
Risk analysis and risk calculation are bureaucratic impositions on the art of medicine. They may claim to promote “best practices” or empower patient autonomy, but the claim does not stand up to scrutiny and risk analysis contributes to distancing the doctor from the patient.
We would do well to remember that if we want to help heal patients, our job is to make decisions, not calculations.³
1. One could certainly retort that to assume a risk, one has to know what it is. That may be true in some sense, but I doubt that to assume a risk one would have to know it with the kind of precision that justifies an “arithmetic” of care. There is obviously more to be fleshed out here…
2. If anyone knows of that reference or one like it, please send it along.
3. This statement obviously invites further reflection on more general theories of decision-making, of which I’m but a very beginning student.