A Shift of Focus


Combining End-of-Life Care & Acute Care in the ICU


By Mitchell M. Levy, M.D.
The main point of this conference is to discuss the challenges of providing adequate end-of-life (EOL) care in prisons. These challenges are engaging other sectors of the medical community, of course, including acute care providers in intensive care units (ICUs). Perhaps our experiences in the ICU will cast some helpful light on the situation in prisons, where correctional health care professionals are also devising strategies for adequate EOL care. I cannot offer an account of our solutions. One might think we would be skilled in dealing with death in the ICU; unfortunately the reality is otherwise. But the nature of these challenges has recently become much clearer for us. I would like to present to you some recent findings, together with some considerations of my own about how we might respond to them.

I want to review with you some data which indicate the dramatic increase in the degree to which life support is now withdrawn in the acute care setting (1). In this study, the researchers looked at deaths that occurred in the medical-surgical ICU in hospitals associated with the University of California at San Francisco in 1987-88 and 1992-93. They found that ten years ago only half the patient deaths (51%) were preceded by any limitation to life support (i.e., withdrawal of life support, withholding of life support, or withholding of resuscitation); but five years later the number had risen to 90 percent. I think these results are consistent with our general experience, which now for the first time has been rigorously documented. You can see how important it is for acute care clinicians to have some kind of training for dealing with EOL issues. Unfortunately, formal training relative to EOL issues and withdrawal/withholding of care is almost universally lacking in medical schools and certainly in internal medicine and surgical residencies.

What are the problems and obstacles which confront caregivers in acute care settings? A key aspect of this questions was addressed at the International Conference of the American Thoracic Society in May of last year (2,3). Researchers from the University of Toronto had interviewed residents, staff physicians, and nurses about what they thought were the barriers to proper discussions on EOL issues between caregivers and patients. Their responses provide excellent insights into the challenges faced by acute care clinicians in care for the dying. Several problem areas were identified.

The nature of the care system itself presents problems. Medical education is largely specialty-based; it ill equips caregivers to see the need for such discussions. In large impersonal hospitals, patients are often lost in the shuffle and rarely form long-term relationships with caregivers. And consensus is often difficult to obtain about the need and the timing for discussions. One of the biggest problems involves the role of family physicians after their patients are transferred to the ICU, where they come under the care of staff doctors. Although we pay lip service to the need for including primary care physicians, they are all very busy; after a while they take advantage of the fact that they don’t have to be responsible for the care of their patients in the ICU. As a result, primary care physicians are often not included in important EOL discussions. Thus we overlook the invaluable input of someone who has been caring for our patient for twenty years or more.

The environment of the institution can be a significant obstacle.Often caregivers are unable to have appropriate EOL discussions even when they wish to. Scheduling discussions is difficult when the caregiver is too busy or the family are not available until after the caregiver has left. In any case, very few facilities are available for such discussions. My ICU, for example, has eighteen beds and one grieving room; often we have three or four patients who are active in EOL discussions at once. So we are often looking for a little corner where we can take a family–and usually end up in the hallway, because the grieving room is occupied.

Most important, I think, is the question of expertise. Interns and residents, nursing staff and attending physicians are often not trained to conduct EOL discussions. It would be unthinkable behavior for a resident to put a pulmonary catheter into someone’s heart without obtaining appropriate supervision. Yet it is common for interns with no formal EOL training to initiate complex and difficult EOL discussions with grieving families, undisturbed by their lack of training and the absence of supervision. This lack of expertise has become so institutionalized in acute care settings that we don’t feel badly about our lack of proper training for such discussions. Worse yet, we don’t even identify training itself as necessary for the provision of skillful EOL care. I might add that this situation creates a further barrier: because training is not universally viewed as essential, caregivers lack incentive to learn more. Furthermore, the Toronto study points out, caregivers are not evaluated in this area. If we are not being held accountable for EOL discussions, we simply don’t keep up our skills.

The lack of remuneration for holding such discussions may also be considered a significant obstacle. This dilemma is much discussed at professional meetings. Critical care physicians generally want to spend more time with families, but the fact that an hour’s discussion with a family does not have a billing code will inevitably affect a doctor’s priorities when deciding whether to hold those discussions. Another barrier comes from the fact that hospital policies regarding discussion of EOL issues are not standardized: it is not routine or mandated, and adequate cues for when to do it are not clear in the minds of many caregivers.

What are some of the more personal deterrents to these discussions for caregivers?

Many caregivers–both physicians and nurses–feel uncomfortable about pushing families or addressing EOL care issues in a very direct way for fear they will not be supported either legally or adminstratively. Risk managers in acute care settings tend to look for the path of least resistance. So in case of conflict the clinician feels that the hospital administration will of necessity support patients and families.

In the matter of family dynamics, the family’s lack of sophistication in working the care system and their inability to understand medical diagnosis and prognosis present great difficulties in acute care settings. Often caregivers are asked by family not to talk to patients. Families may be quite large, as you know: every day different family members identify themselves as the ones who need to be talked to, and are very upset if they are not seen as primary spokespersons. In addition, subplots surrounding family finances, as well as dysfunctional relationships, make direct and simple discussion about EOL issues very difficult.

One of the primary challenges faced by all acute care clinicians is to determine the appropriate time for starting EOL discussions. Prognostic uncertainty is one of the most important issues. At one time EOL discussions were initiated only when a patient was terminal. Do we need to wait until the severity score indicates that this patient has 95-per-cent-plus likelihood of dying in the ICU? Do we start discussions early with a chronic lung patient, when she is admitted to the ICU, although she has a high likelihood of being discharged early? The concept of parallel streams of care, introduced by Joanne Lynn and her colleagues (4), is now being talked about increasingly in critical care units. This approach urges us to start EOL discussions early to get people used to talking about these issues. As a result, when the time comes for them to make their decisions, they have had some preparation and are more comfortable. Finally, the Toronto study notes, both cultural differences and value differences between caregivers and patients create real barriers.

So the real issues in EOL care, at least for me, in an acute care setting are Who? When? and How? In my institution, several people feel they have responsibility for EOL discussions. Part of the problem is to decide who has responsibility in a given case. Conflict often arises between the acute care clinicians, chronic care clinicians, hospice, continuing care workers, discharge planners, social workers, and the palliative care service. The situation is very confusing for interns and residents. When they have a family in crisis about EOL care issues, they are expected to call in the continuing care and palliative care services. If they have a patient who is ready to deal with the inevitable, hospice workers feel that they are the ones to be called in. Since the lines between specialties tend to be blurred, it becomes more important that we EOL caregivers come together and develop a working consensus on how to address the situation and who answers what questions. When is a critical illness “terminal”? When does end-stage disease become terminal? The terminology often gets in the way of knowing when to initiate discussions. Determining when EOL issues begin is very important for us, but we don’t have very clear answers at present.

At least part of the problem is that we still maintain a strict separation between aggressive medical care and comfort-measures-only. Families think that when they authorize a DNR Order for their loved one, all therapy will be withdrawn. For many acute care clinicians, especially in the ICU, the DNR Order calls for less aggressive care. There is an institutionalized feeling that a patient with a DNR Order does not even belong in the ICU. In progressive institutions, however, a DNR Order does not imply less aggressive care until such time as the DNR Order becomes applicable.

I would like to offer some suggestions on how we might do EOL care, and conduct appropriate discussions, at least in an acute care setting. We need to develop a routine approach to EOL, and to understand that EOL care does not mean “terminal” care and that acute care is compatible with EOL care at all stages of an illness. Patients can receive both aggressive care and EOL care–even in the ICU. We should be able to call in a hospice consult, a palliative care consult, an EOL consult, on the day the patient arrives in the ICU–before we have a feeling for whether or not he is going to survive. Such a policy would relieve families and acute caregivers of the need to make prognostic decisions; EOL care discussions would simply be routine in critical care settings. Instead of labeling the EOL discussion as an implied failure, we must understand these discussions as a way of healing people during their illness, during the dying process. Usually a hospice consult is called as an act of desperation, when the acute care clinician has exhausted all possibilities. Instead, we could see EOL issues as good news for patients: we could create a healing environment during a very difficult period which might result in survival or death. The issue, however, is not survival vs. death, but appropriate EOL care.

In conclusion, the need to develop adequate programs for training caregivers in EOL skills in acute care settings is now more imperative than ever. Our challenge is to separate hope for our patients from our need to make patients better. Many acute care clinicians, who identify their job as making people better, need to be reminded that, as often as not, we don’t make people better. We can have good news come out of not making people better, but for this to happen, we must become comfortable with a feeling of hopelessness about the process of life and death in our patients. We need to learn EOL skills for patients who are likely to survive. We need to learn not to be embarrassed by EOL issues, so as to be able to develop personal and institutional strategies not only for withholding and withdrawing care, but also for applying the widest possible range of clinical skills–based on frank and open discussion between patient, family, and caregivers.


1. Prendergast, T.J., & Luce, J.M. (1997). Increasing incidence of withholding and withdrawal of life support from the critically ill. American Journal of Respiratory and Critical Care Medicine, 155, 15-20.

2. Guest, C.B., McLean, R.F., Palda, V.A., Vachon, M.L.S., Kelner, M.J., Lam-McCulloch, J. (1998). Barriers to discussion of prognosis, death and resuscitation: Perspectives of care givers. Department of Anaesthesia, Sunnybrook Health Science Centre, University of Toronto. [Printed summary of study results distributed at the convention.]

3. Guest, C.B., McLean, R.F., Palda, V.A., Vachon, M., Kelner, M.L., Lam-McCulloch, J. (1998). Barriers to end-of-life discussions: the physician’s perspective (abstract). Sunnybrook Health Science Centre, University of Toronto (Toronto, ON, Canada). American Journal of Respiratory and Critical Care Medicine, 157 (3, part 2), A303.

4. Lynn, J., Harrell, F.E., Cohn, F., Hamel, M.B., Dawson, N., Wu, A.W. (1996). Defining the “terminally ill”: Insights from SUPPORT. Duquesne Law Review, 35, 311-336.

  • Mitchell Levy is associate professor of medicine at Brown University’s Rhode Island Hospital in Providence. The following is an edited version of a presentation at the conference on “Finding Common Ground for Hospice Care in Connecticut Prisons,“sponsored by the Connecticut Chapter of NPHA at Choate Rosemary Hall School, Branford, CT, May 1st, 1998.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s