When optimism isn't called for
How doctors can counteract surrogate bias
Keywords: Hospital study shows optimism bias in communication leads to overtreatment.
by Deborah Gesensway
Published in the May 2012 issue of Today's Hospitalist
CONVENTIONAL WISDOM holds that family members who are overly optimistic about the prognosis of a critically ill loved one may simply not understand what doctors have been telling them.
But research in the March 6 issue of the Annals of Internal Medicine reveals a more complicated communication challenge. It's not just about doctors giving—or not giving—patients clear information. Instead, family members filter the dialogue through an "optimism bias" when hearing bad news.
For Douglas B. White MD, MAS, senior author of the study, the takeaway message is that improving discussions with surrogate
decision-makers takes more than a straightforward strategy of teaching doctors to communicate more clearly. It also requires doctors to spend time with surrogates, learning more about their emotional and psychological factors, then customizing how and when to deliver bad news so that families can process it.
|"We need to mention a poor prognosis earlier than we are comfortable with."|
–Douglas B. White, MD, MAS
University of Pittsburgh
Dr. White is associate professor of critical care medicine and director of the program on ethics and decision-making in critical Illness at the University of Pittsburgh. He spoke to Today's Hospitalist about the findings and what they mean for hospitalists.
What's wrong with surrogates being optimistic?
It's entirely understandable. There is something heartening about the fact that people respond to adversity with optimism. But there are specific situations where this optimism is problematic. In the ICU, surrogates should generally make decisions that comport with the patient's values and preferences. You can imagine how an optimism bias may threaten that.
Say it was clear that a person would never want to be in a nursing home, and the doctor says that is unavoidable because the patient is now so impaired. The right answer in those circumstances, according to the patient's values, would probably be to not undertake aggressive measures.
If surrogates don't have a clear perception of the likely outcome, they may be more likely to make a treatment decision contrary to what the patient wants.
This is probably one of many reasons we have so much overtreatment at the end of life. It takes a lot of emotional work to try to overcome a bias that can push us toward decisions that may not align well with patient values.
The study found that when doctors expressed a high risk of death, surrogates tended to hear that the patient had about a 50% chance of survival. What does that tell you about how doctors should phrase grim prognoses?
Most people in the past have said this represents a misunderstanding: either the doctors are not talking about the poor prognosis or they are talking about it in vague ways. If that was the only problem, we could just develop ways to clearly communicate risk information.
These data suggest that the problem is much more complex. There are issues of understanding and also of whether surrogates believe doctors' prognostic estimates and are willing to accept them. Those issues are a lot harder, but not insuperable, to overcome, but you need to understand the cause of the problem to design a solution.
For the doctor to just say the poor prognosis louder and more frequently is probably not the answer. Maybe the family has heard what he's said and they do understand it, but they are not yet at a place where they accept it. Or maybe they have some legitimate doubts. In such cases, doctors can spend their efforts exploring the reasons that underlie those doubts.
After giving bad news, what should doctors do to see if surrogates understand it?
Deal with the cognition issue first. I often say something like, "Who will you go to and talk to about this?" Then I say, "Can you tell me what you are going to say to them? How will you summarize this conversation?" That is a nice way to get someone to teach back to you what you just said. It feels more like a rehearsal than a quiz.
In their response, we often see certain issues come up. Surrogates might say, "I heard you say that you don't think the prognosis is very good, but we are hoping for a miracle." All of a sudden, you have an insight into the family. Or they say, "I'm going to tell them that mom is really sick but that she is going to pull through." Then the doctor can realize that maybe there is an issue of misunderstanding and he needs to gently reinforce what he said.
Or the family could say, "Yes, doc, we've heard this all before. Mom was in an ICU 10 years ago and they said the exact same thing—and she was home doing well a month later." Then the doctor has to decide whether his prognostic estimate is accurate and explain to the family why now might be different than 10 years ago.
At what point in their interaction with surrogates should hospitalists raise concerns about a grim prognosis?
This runs counter to our human nature, but we need to mention a poor prognosis earlier than we are comfortable with. These conversations require time, and families need time to process new information.
I like to think of it as a preliminary mention. A fairly common thing I hear doctors say in the first meeting with the family is, "She is really sick, but we have her on the ventilator and we are doing everything we can, and we'll see how things go." Instead, say, "Although she is on a ventilator and doing OK right now, I'm concerned because there is a real chance that she won't survive this. I am hopeful that we will get her through this, but I wanted to share with you the possibility that this may not go well."
That shares the uncertainty of outcome. If you haven't had the conversation on day 1 and you get to day 3 or 4 and the patient is doing badly, those are days lost for the family to start to prepare for the outcome. There has to be time built in to let the families first process difficult news before asking them to make hard decisions.
Also, I think doctors should have a default expectation that several conversations will be required for families to come to terms with a poor prognosis. I think this would prevent some of the frustration I see among physicians that these conversations are hard and the decisions take so long.
Is there a difference between how you talk to family members to help them accept a prognosis and start grieving and how you talk to a surrogate who first must make decisions that reflect the patient's values?
To get decisions that reflect the patient's values, doctors have to attend to families' emotions. In so doing, you are helping them as individuals cope with a very difficult thing, and you are helping them be better surrogates. It's a challenging two-for-one, but it is a two-for-one.
Deborah Gesensway is a freelance writer who covers U.S. health care from Toronto.