"He will definitely not survive" - I do not think it means what you think it means

Photo Copyright Annie Levy, 2008

Sorry, stuck on the Princess Bride references.

At our most recent geriatrics journal club we discussed a terrific article by Lucas Zier (first author, UCSF internal medicine resident), Doug White (senior author, Pitt), and colleagues.  The article is titled Surrogate Decision Makers' Interpretation of Prognostic Information, published in Annals of Internal Medicine.

Here is the setup.  You're working in the ICU.  You want to communicate a prognosis to the family of a patient who is so ill he cannot make decisions.  You sensitively state the facts: the patient has less than a 5% chance of survival.  Or perhaps you say "he will definitely not survive."  The family confers, and decides that they want to focus on keeping him alive as long as possible.  You wonder to yourself, "I know what I said. What did they hear? Do they understand that he's dying?"

In this study Zier and colleagues surveyed surrogates of seriously ill ICU patients.  Researchers presented surrogates with a range of possible prognostic statements for a hypothetical patient  - NOT their loved one -  and asked them to mark what they interpreted the statement to mean on a scale from 0% to 100% chance of survival.  Results were astounding:

• For optimistic statements, such as "He will definitely survive," or, "He has a 90% chance of surviving," almost all surrogates generally indicated agreement with the statement's intent, interpreting the statements to mean >90% chance of survival.  
• For pessimistic statements, such as "He has a 5% chance of surviving," or, "He will definitely not survive," surrogates were much more optimistic, with mean interpretations in the 25-30% survival range, and some surrogates indicating 95% chance or greater likelihood of survival!
• When these findings were shown back to surrogates, they were at a loss for how to explain their optimistic responses to the grim prognoses.

What to make of this?  The authors point out that there is much more to communicating accurate information than misunderstanding of numerical risks.  Psychological factors play a huge role in how patients hear and interpret information.

I remember my first day rotating on the palliative care service at the Brigham and Women's Hospital.  Susan Block was holding forth about how dying patients and their surrogates, when confronted with a grim prognosis, sometimes simultaneously believe contradictory information: most of the time they hope for survival, but at some level, they recognize that they are dying.  The study by Zier and colleagues seems to hint at such a complex psychological picture.  My guess is that forcing surrogates to assign a single number to such a complex psychological phenomena is where the study ran up against the limitations of its methods.

What to take away from this?  Telling a surrogate that the prognosis is grim is not the same thing as surrogates understanding that the prognosis is grim.  They may take away a substantially more optimistic message than  intended.  This is important information to understand and teach.  The question is: what to about it?

Here is where I disagree with the conclusion of the article: that these psychological biases are an obstacle to be overcome, or remedied, because "not all optimism is ethically benign."  I worry that undermining the optimism at a vulnerable time could be harmful to the surrogates, and potentially detrimental to the relationship between ICU providers and surrogates.  Ultimately, this sort of approach may hinder rather than facilitate communication.

Rather, as Susan Block taught, our job is to prop up those psychological mechanisms that are supporting the surrogate, helping them to cope with a dreaded event, even as we're preparing them in case things don't go as they hope by providing realistic information.  Even if these goals seem at cross purposes.  A complex response to a complex psychological state.

by: Alex Smith