Last month on the wards, I unilaterally told two patients’ families that we were not going to resuscitate their loved ones. My residents were horrified – this violated the DNR playbook – but the alternative seemed both immoral and absurd.
What do you think?
In deference to HIPAA, I’ve altered a few clinically irrelevant factoids, but here’s the gist: In both cases, the patients were elderly Asians. One was a 92-year-old nursing home dweller with severe dementia; prior to admission he was bed bound and could not recognize his family members. He was admitted with pneumonia and ARDS, and had been intubated for nearly a month when I assumed his care. My predecessor had held a number of family meetings; in each, the family insisted on doing “everything,” and that’s precisely what we did. The family continued to demand the full court press, even as his organs shut down, one after another. On his final day, he required high doses of pressors and as much respiratory support as his mechanical ventilator could muster. Nevertheless, he barely maintained a blood pressure and oxygen saturation compatible with aerobic survival.
I described the situation to the patient’s family, who continued to want to “do everything,” including CPR. I told them that the time had come to stop. “At this point,” I said a bit sternly, “the chances of his surviving are nearly nil, and I can’t be party to torturing him during his death. So we won’t do it.”
I saw my resident’s jaw drop. “What would you have done if the family had demanded that we continue to push forward?” she asked me later. I would have refused, I told her, and offered to try to find another doctor. But instead, they appeared relieved as we shifted to comfort care. The patient died a few hours later, surrounded by his loving family.
The second case also involved a patient in her early 90s, this one riddled with metastatic cancer. She was bed bound and in chronic pain. I approached the family to ask them what they wanted us to do if she deteriorated. “Everything,” was the predictable answer. “We will do everything we can,” I said, “to keep her comfortable and get her home. But CPR in 90-year-olds with widely metastatic cancer is all-but-futile, and I think it would be wrong. So we won’t do that.” Again, I heard my resident gasp. But again, the family didn’t fight the DNR order.
As recently as few years ago, my approach would have been different. Being a bit of a student of ethics and having written about DNRs and CPR in the past (my early research was on ICU care for AIDS patients), I generally support the modern ethical paradigm, which emphasizes patient autonomy over all else. In essence, it says, “Doctor, you can choose not to offer therapy – CPR, ICUs, transfusions, dialysis – only when the therapy is futile. If it is not, then the patient’s judgment rules the day – in essence, all life-sustaining treatments are informed consent procedures. And, if the patient is unable to participate in these discussions, we use proxy decision-makers (a formal DPOA or a close family member) to tell us what the patient would have wanted.”
I embraced this model for most of my career, and still believe that it has much to commend it. It prevents we doctor-types from being too paternalistic – from substituting our opinions for those of patients and family members. It also prompts us to give patients and loved ones the information they need to make informed choices. Finally, it explicitly excludes considerations of resource use, a noble bit of cognitive insulation, since in an ideal world patients and families wouldn’t have to worry about the expense of CPR or dialysis when deciding whether to accept life-sustaining care.
Because physicians have retained paternalistic control over only the sliver of terrain called “futility,” many scholars have struggled to define what this word means. I like the Schneiderman definition as a starting point: when physicians conclude that there have been no successes in 100 similar cases, that’s futile. (The definition hinges on a convenient statistical shortcut: the 95% confidence interval of 0 successes in n trials is 0, 3/n. In other words, 0 successes in 100 cases yields a 95% confidence interval of 0-3%, a range that many people are comfortable calling “futile”).
I once described the modern ethical algorithm to one of my residents and simplified it by saying that the world could basically be divided into two buckets: futility (which is very rare), and everything else. If care was truly futile, we could paternalistically not offer it – we have “no obligation to offer futile care.” But all non-futile care (the vast majority of situations) should generate a discussion with patients or proxies in which we blandly give them prognostic data and perhaps a weakly-worded recommendation, and simply wait for them to decide.
The resident was a smart guy. “Seems to me that we need a third bucket,” he said tartly.
And we do. In the case of my nonagenarian with sepsis who was dying despite maximal ICU support, one could argue that his care was physiologically futile: we were doing everything we could but it wasn’t working. But the family “had faith” and – had they pressed me – I’m not sure I could have said with 1000 percent certainty that our treatments has no chance of working. Rather, it was crystal clear that the burdens of continued aggressive care outweighed the benefits. Ditto for the nonagenarian with metastatic cancer: CPR (had she needed it) might not have been futile, but it just seemed so very wrong in every way.
But why resurrect the dread paternalism – which violates American ideals of self-determination and risks the substitution of our own biases and even conflicts of interests? Wouldn’t families agree to withhold aggressive care if you just gave them the facts, accompanied perhaps by a bit of crepe-hanging spin?
Why indeed? Because our ethical model takes us to places (quite often, I’m finding) that are both absurd and morally wrong. Absurd: how can we justify CPR on demented 90-year-olds? Or those with metastatic cancer? Morally wrong: I have come to believe that our reluctance to exercise thoughtful paternalism – to say, “this would be wrong and I won’t do it” – leaves families in the gut-wrenching position of being forced to shoulder a lifetime of guilt for having endorsed actions (or lack of actions) that allowed their loved one to die. Many families, particularly those of certain cultures (in my experience, Asian and Russian families are the least likely to consider limiting care), simply cannot tolerate this burden, or have belief systems that cannot countenance any limits.
Which brings me to the final absurdity: money. Each day in an ICU costs someone $5,000-20,000. There are 50 million uninsured people in our country, the public schools are terrible, there are homeless people outside my window, the bridges are falling down, and GM is bankrupt. Can anyone really argue with a straight face that we still have the luxury of offering “everything” to 90-year-olds with dementia? It is a cop-out to say that doctors shouldn’t do “bedside rationing” – yes, I know the arguments: the patient needs to know that I am advocating for him (“The Doctor’s Master” can only be the patient, not “society”, famously wrote Levinsky a generation ago), there is no guarantee that the money saved won’t be spent on Iraq, Bridges to Nowhere, B1 Bombers, yada yada. But these are straw men. Squandering precious resources because we can’t say no is just plain wrong, and I believe everybody knows it. We just don’t know how to create that third bucket without doing violence to an ethical framework that has generally served us well.
And, of course, through the magic of our healthcare insurance system (in these situations, even patients who don’t have insurance are covered because the hospital picks up the bill), patients and families have the luxury of being entirely protected from the cost consequences of their decisions. It would be interesting to see what would happen if, once doctors recommended no further aggressive care, patients or families faced a co-pay of even 10 percent of the extra cost. I’m guessing that the vast majority of decisions would be different. It’s easy to want “everything” when it is paid for out of somebody else’s checkbook.
Don’t get me wrong. I still buy into the general ethical model for decisions about life-sustaining treatment. For every patient in whom I unilaterally refuse to offer care that I consider to be wrong (it’s not like I go around making everybody a DNR without consulting with them), I find that I overturn the housestaff’s preliminary DNR orders at least twice. “I asked her if she wanted to be resuscitated, and she said no” is a common refrain I hear on morning rounds. And then I speak to the patient and learn that yes, the 68-year-old with moderate COPD would not want to be “stuck on a ventilator” or be “a vegetable.” On the other hand, when I ask, “if you developed an infection and needed to go on a breathing machine, and we thought you had a 50-50 chance of recovering, leaving the hospital, and getting back to where you were last month, would you want that?” they say, “Of course.” That’s not a DNR patient; that’s one whose preferences should be noted in case they do later “get stuck” on a vent.
For now, I feel like what I did in the cases of these two 90-year-olds is fully in sync with my Hippocratic oath. It is the usual practice – doing everything just because families couldn’t or wouldn’t say no – that is feeling increasingly unethical to me.
But this is tricky stuff. Am I the only one who has become more paternalistic in these matters? The only one who is bothered by the absence of that third bucket? I’d love to hear your thoughts and practice.