There’s been a lot of talk about advance care planning lately, with Ellen Goodman’s Conversation Project, Atul Gawande’s book, Being Mortal, and very recently, Angelo Volandes’ book, The Conversation. The message: think about what matters to you and discuss with your physician and with your family the approach to medical care that makes sense for you if you become very ill. We know that these kinds of discussions may not lead to the patient’s directive being followed when he or she actually does get very sick—sometimes the advance directive does not travel with the patient to the hospital, often the directive is difficult to interpret in practice (what exactly does “no heroic measures” mean?), and in some circumstances physicians or families override the directive. We know that patients with a medical order such as a POLST (Physician Orders for Life-Sustaining Treatment) are more likely to get what they signed up for than those with a wish statement (for example a living will). What we don’t know much about is whether patients change their minds when faced with an actual illness. A new study examines how often patients admitted to an ICU with some kind of “treatment limitation” in effect nonetheless received the treatment they previously had said they didn’t want.
The answer: about one-quarter of the time.The most common treatment that people specifically said on admission to the hospital they didn’t want was attempted CPR (accounting for 77.4% of the limitations expressed), but 24.6% of them ended up having CPR initiated. Another 21.3% of people said they didn’t want specific therapies such as dialysis or artificial nutrition, but it is not clear from the study just what proportion of them received those treatments. Finally, 3.9% of those admitted said they wanted a focus on comfort. We don’t know what kind of treatment those patients in particular received. All we know is that of the 13,405 patients admitted to one of 141 ICUs at 105 US hospitals between 2001 and 2008 with some kind of treatment limitation request in place, 3123 ended up getting attempted CPR, 3841 got intravenous medicines to maintain their blood pressure, 2660 were put on breathing machines, and 283 got dialysis. Finally, we know that there was considerable variability between hospitals, both with respect to the rate of treatment limitation expressed initially (from 1% to 30%) and to the rate of treatment limitation reversal (2% to 76%). So what does this all this mean?
Since we don’t know what led to the changes that occurred, we can’t know for sure. What is clear that these were not cases where the doctors simply never saw the advance directive. These are all cases in which the plan was not to provide a specific treatment (principally CPR) but the reality was that the order was rescinded and the forbidden treatment was actually administered. So somewhere between entering the ICU and leaving it (whether alive or dead), a major change took place. It could be that when push came to shove, patients changed their minds. When confronted with their own mortality, they chose to have a shot, however small, at living longer. It could be that most of the patients were incapacitated by illness and the actual decision to reverse course was made by a health care proxy or family member—we know from other studies that fully 70% of patients in whom a decision about life-sustaining treatment needs to be made are cognitively unable to engage in decision-making at the crucial moment. It might be that the limitation-of-treatment directive didn’t really reflect the patient’s goals—perhaps what they meant was they didn’t want to spend their life as a vegetable, not that they didn’t want a trial of life-prolonging treatment—and they came to understand that their directive as written should be revised. Or it could be that patients and families were persuaded to change course because the environment of the ICU promotes life-prolonging treatment and it is very difficult to stick to a plan of care that violates the raison d'etre of the ICU.
In all likelihood, the answer is all of the above. This study does not imply we should abandon advance care planning. It’s worth pointing out that 75% of patients with treatment limitations in place retained those restrictions. But it does raise questions about whether patients really know what they are signing when they complete an advance directive—whether a traditional living will, a detailed instructional directive, or a POLST form. It makes me even more convinced than ever that we need to focus more strongly on ascertaining the patient’s goals of care when we engage in advance care planning, and leave the translation of those goals into medical treatment to the moment when an actual decision must be made. And it does remind us that hospital culture in general and ICU culture in particular are very powerful. When an institution is structured and staffed to provide life-sustaining care, then treating some patients without using life-sustaining care creates cognitive dissonance. If patients truly want a different approach to care, we should provide an alternative environment in which to deliver that care, either the home or an intermediate facility such as a skilled nursing facility.