Friday, May 24, 2013
Verb Selection in Code Status Discussions: A Potentially DisruptiveHospital Innovation
Modal verb (n.): a verb such as can, must, may or will that is used with another verb (not a modal) to express possibility, obligation, permission, intention, etc.
Quick: You see a patient who has progressive hypoxia and the nurse hands you the results of the arterial blood gas while the patient is on a non-rebreather. The partial pressure of oxygen is 50 mmHg. What should you do? Pull out your Washington Manual of Medical Therapeutics if you must.
Now, you see a patient who has metastatic cancer. He has exhausted all chemotherapy options and has had extensive discussions about prognosis with his oncologist. He has expressed a goal to remain at home surrounded by family during his final weeks with a focus on comfort. He develops sudden dyspnea and instead of calling the home hospice, his wife calls 911 and he is transported to the emergency department. An ABG on a non-rebreather reveals a Pa02 of 50 mmHg. What treatment course might be most consistent with the patient's goals?
 |
| Mindmap of Modal Verbs via http://balanbalanbalan.blogspot.com/ |
Your answer to each question might have been different because of how the questions were framed. For the second question, you hopefully answered, "I need to talk with the patient and/or patient's surrogate about this, ASAP."
In a study presented at the American Thoracic Society meeting, researchers from the University of Pittsburgh recorded a simulated encounter between an emergency medicine physician, hospitalist, or critical care physician and a 78 year old man with metastatic gastric cancer in a similar situation. The researchers evaluated the frequency of five rhetorical strategies which the physicians used in the conversation, including: The next step ("we will"), a necessary action ("we must"), a convention ("we usually"), an option ("we can"), a treatment preference-eliciting question ("ask"). They evaluated how frequently each rhetorical strategy was used when discussing a) Intubation and mechanical ventilation and b) palliation (non-intubation and comfort measures). The results:
- Most physicians discussed life-sustaining treatments and only discussed palliation after the patient revealed his preference against intubation.
- 54% of physicians used a "must" statement when discussing life-sustaining treatments (e.g. "Medically, he needs to go on a mechanical ventilator") with another 6% using "will" statements and 4% using "usually" statements (e.g. "usually, we put patients like you on a ventilator").
- Only 86 out of 114 physicians reached a point in the conversation where palliation was discussed, in spite of the man's wish for comfort measures over intubation.
This study identifies two methods of framing. In 68% of encounters, life sustaining treatments featured as the first topic of conversation. The order in which options are presented represents an important frame and a nudge towards the first option listed.
The modal verb serves as the other vehicle by which the decision is framed. Have you ever heard a physician say something like, "If the mask you are on doesn't keep your oxygen level up, we might need to place a tube in your throat and hook you up to life support to keep you alive, but we want to know if you want that." Some patients will declare a longstanding preference against intubation and others might want it under any circumstance. I suspect there's a large group of patients in the middle who aren't sure, but say "well, if you think I need to do that, I guess that's what I have to do." You think you've identified the patient's preferences, but not really. Perhaps the patient has already inflated the chances of ICU survival with a good outcome, and now you've said that the patient needs to go on a vent. You've really just inadvertantly nudged the patient towards a decision, and a single word- the modal verb you have selected- might have sealed the deal.
Decision made. Frame, set, match.
How physicians use the modal verb in this circumstance likely plays a role in the determination of the patient's decision frame. More importantly, however, it may reflect the "inner frame" of the physician. Of the 54% of physicians who used a "must" statement when talking about LST, I don't believe that all of them would recommend intubation to the patient (if the patient asked). However, I hypothesize that a significant proportion of physicians have framed the decision in their head as "I have a hypoxic patient for whom I think mechanical ventilation is indicated, and I need to decide whether to intubate" rather than "I have a dying patient, and I need to decide what interventions are most consistent with his goals of care." In other words, in many cases the use of the modal verb "must" or "need to" may reflect the inability to consider how patient-centered goals of care might impact a decision.
A timeout can help. Think through the words you'll say during the conversation. Look at the modal verb for a second. If you think intubation is indicated medically and the vast majority of people would agree to it, then maybe it's ok to say "must" or "need to." Realize that you're nudging the person in that direction, but it may be OK in many circumstances (e.g. a 23 year old with status asthmaticus). If a reasonable person might not want the proposed interventions, at the very minimum, consider a change in your modal verb to the more permissive and frame-neutral "can" or "could" and closely follow it up with the alternative option of comfort measures. ("While we could place you on a mechanical ventilator, I'm afraid there's a high chance you would not come off the ventilator to be able to go home. Another option to consider would be to continue to focus on keeping you comfortable and making sure that your family can be at your bedside. Tell me what you think about these options?") Better yet would be to pause, make sure you understand a little about the patient's experience, values, and goals, and then frame the information about options, your preference question, or a recommendation based on that information. You may reorder the list of options, describing the palliative option first, using an "obligitory" modal verb to convey your strong support for the option ("We will focus on making you as comfortable as possible.") If you're absolutely clear that comfort is the overarching goal of care in a dying patient, I would even consider "negative" modals for ICU interventions. ("We should not" or even "we will not place you on a ventilator.")
I've focused a lot on the modal verb and have generated a plausible hypothesis about the effect of verb selection on the frame of the conversation (and thus the decision), but there's more to this study. In a separate analysis published recently in Critical Care Medicine, the researchers looked more broadly at variation in physician communication behaviors and decision-making roles. A minority of the physicians communicated short term prognosis (27% conveyed prognosis if patient were intubated and 38% conveyed prognosis without intubation). 39% sent the patient to the ICU and 9% decided to intubate (even though this wasn't the patient's preference--a medical error borne from inadequate communication). Only 36% explored the patient's broader values with 82% eliciting intubation preferences. So most talked about intubation preferences without the important contextual frame of prognosis and goals of care/values. One might say, "but only 9% decided to intubate, so most arrived at a plan honoring the patient's preferences against intubation without doing the values/goals of care or prognosis thing." What if the patient's preferences weren't so clear though? What other opportunities to honor the patient's goals of care might be missed? This patient probably shouldn't be going to the ICU 39% of the time.
For more on research about code status conversations in the hospital as well as how goals of care and prognosis frame code status discussions, see here.
This week, Geripal is having a "code discussion week." Alex Smith asked whether the default code status should be changed to DNR for patients with serious illness. (In fact, a comment on the post by Amber Barnato led me to the abstract for the abstract referred to in this post.) For many reasons, I doubt we'll get to the point of default DNR for anyone. We do, however, have control over the approach we take to the conversation, the language we use, and at the very minimum, the default modal verbs we use, or rather must use.
