Radiolab recently aired a show called “The Bitter End” that discusses the end-of-life care preferences of physicians and non-physicians. Physicians are much more likely to decline “heroic” measures, such as CPR, mechanical ventilation, feeding tubes, etc. This comes as a surprise to the hosts and, presumably, to other non-physicians. It’s a good show. I recommend it. (Full disclosure: I like Radiolab.)

In the show, Ken Murray argues that physicians decline these “heroic” measures for intellectual reasons. He argues that we know the data, which includes a study that reported that, of people who receive CPR, only 8% are successfully resuscitated. (Of those 8%, only a portion of them return to their full previous function.)

I don’t think physicians decline CPR and other “heroic” measures because of evidence-based, numerical data alone.

The experience of performing CPR and attending to patients who are critically ill contribute to physician preferences against CPR. It’s emotionally taxing. All physicians have seen the trauma we cause with these “heroic” measures. Yes, performing CPR can lead to cracked ribs and punctured lungs. Mechanical ventilation can lead to severe cases of hospital-acquired pneumonia. Intravenous hydration can cause massive tissue swelling. The consequences of heroic measures are often devastating.

Physicians are taught “first, do no harm.” Sometimes, these heroic interventions seem like they cause more harm than good.

Perhaps physicians decline CPR and other “heroic” measures because of anecdotal experiences and emotions. This isn’t randomized, placebo-controlled data. However, anecdotal experiences and emotions are still data.

Furthermore, there is no true “informed consent” with CPR. When patients are able to consent to CPR, they are not truly informed. They cannot fully appreciate and understand what CPR entails because they have never experienced it.

By the time patients are truly informed about CPR—when someone is pushing on their chests, when a second person is manually inflating their lungs, when a third is injecting medications into their blood, when a fourth is trying to stick a breathing tube down their throats—they are unconscious. They cannot offer or withhold consent.

(This is true with many things in medicine: No one can give true informed consent for general anesthesia, surgery, or even medications. We often only know all the information after the fact. Patients often give consent based on hope and faith.)

Physicians see and treat patients who have undergone CPR. Those patients are usually paralyzed, swollen with fluid, and unconscious. Upon witnessing that, physicians might wonder what the differences are between “living” and “existing”.

This could explain why their end-of-life care preferences differ from that of the general public.

Maria Yang is a psychiatrist who blogs at In White Ink.