Thursday, July 8, 2010
American Journal of Respiratory and Critical Care Medicine (AJoRaCCM!) has published a randomized trial of two methods of prognostic disclosure about critically ill patients. The exclamation points represent artifacts of my slack-jawed shock when I read the article's title: a controlled trial evaluating how we should talk with patients is a very rare thing. Although while after reading the trial, my reaction is more of a (!) - more questions than answers dammit! - why does it always have to be this way! - one wasn't of course expecting big answers, and the trial highlights some important questions.
The study has an interesting design, which tries to balance the desire to study actual family members/decision makers of critically ill patients, while acknowledging the difficulty (near impossibility) of actually randomizing those people, involved in real patient situations, to an experiment which may influence important medical decision-making about a patient. The compromise here is that they show real family members of critically ill patients videos of a family conference for a fictional patient - and this occurs during the patient's actual illness (e.g. they recruited these subjects right from the waiting room). The videos, which I'd really like to see, demonstrate a family conference regarding an incapacitated ICU patient with low likelihood of survival. They note that they specifically designed the video using multidisciplinary input about best family meeting practices, including palliative care (the transcript is included in the paper). Subjects were randomized to one of two videos, which differed in how the physician disclosed the patient's chances of survival:
- “I would say it’s very unlikely that he will survive. Saying it another way, that means it’s very likely he’s going to die.”
- “I would say he has about a 10% chance of surviving. Saying it another way, that means there’s about a 90% chance that he’s going to die.”
Subjects, when asked their own estimation of the patient's chances of surviving, estimated it to be ~23%, with no real differences between the quantitative and qualitative disclosure groups. That is, even those who were given a numeric disclosure reported they believed the patient's survival to be more than double what they were told. Both groups as well reported their understanding of the physician's prognostication to be a ~16% chance of survival (these were all mean estimates). Only half the subjects who heard the numeric estimate (10%) reported that as their understanding/belief of what the physician thinks the prognosis is. 21% of subjects reported their belief the patient's chances of survival were 20 or more percentage points higher than what the physician disclosed. Subjects' self-reported trust in their loved one's physicians was associated with less discordance between estimates of prognosis.
So, numeric vs. qualitative-only prognostic disclosure: no difference! Well maybe.....
It's important to recognize that even with their efforts to find real-life subjects to do this on, these were still people watching a fictional video and answering questions about it, as opposed to reporting their beliefs about their own loved ones, next door, in the ICU. It would be a mistake, as well, to conclude this means numeric disclosures aren't more helpful (helpful in the sense of people understanding and believing what you are saying) than any-old non-numeric disclosure. The video was idealized, and (reading the transcript) probably does approximate best-practice of what's out there. That is to say, a numeric disclosure may not be better than what was said in this video, but it may still be better than saying things like 'His prognosis is poor.' 'She's not doing well.' 'She may not make it.' 'Things don't look good.' (At least in isolation of additional information.)
My take-home point from this is the all-too-mundane-but-probably-needs-to-be-said-again-and-again reality that the fine-grained details of what one says are far less important than the fact of taking the time to compassionately, honestly, with plain speech, and with keen attention to the emotions of those with whom you're talking, tell family members the truth about what is going on and what you think should be done.
I'm curious as to whether others are using numbers like this (percentages, rates of recovery, etc. - I'm not talking though about time-based estimates such as saying that someone likely will live 'another 3-6 months')? What has been your experience?
On to, however, the more even interesting question underlined by these findings - why is it that these subjects, overall, overestimated the patient's chances of survival by ~2x, even those who had been told 10%? The authors put it succinctly:
Our data raise the possibility that the physician-surrogate discordance about prognosis observed in prior studies may not be fully explained by poor communication. Other factors, such as the need to express optimism, skepticism about physicians’ ability to predict the future, different belief systems about illness, or distrust of physicians may also explain the discordance.Indeed. To me, this is the real message of the study. That you can do prognostic disclosure by the book - well-paced, attentive to emotion, plain-spoken, honest, all that good stuff - and that may not 'do the job.' I'd really like to hear readers' comments about what they think is going on here. I'd also like to ask what is 'the job' that we're trying to do here with having these discussions. It's obviously more than increasing a family's 'knowledge level' sufficiently that they agree with us on a multiple choice exam about the patient's condition, and once you've made sure you've communicated well about the medical facts, we are left with the challenges of trust and of engaging with grief-stricken family members' hope, beliefs, guilt, faith....
Thanks to Dr. Sean Marks for sharing this with me.
Image ('Hope') from Wikipedia.