• © 2011 by American Society of Clinical Oncology

It Takes a Village

I recently started a patient with recurrent colorectal cancer on first-line palliative chemotherapy. Despite an initial restaging evaluation that showed stable disease, he developed obstructive jaundice with cholangitis shortly afterward and then, while in the hospital, had a large pulmonary embolism. We had previously discussed his overall prognosis and his personal goals in clinic, but we had not addressed every aspect of his advance directives, thinking that we had more time to discuss all of those questions. Unfortunately, his disease was on the more aggressive end of the spectrum. While in the hospital, he and his family were served by outstanding palliative care physicians who had initiated discussions early on in the admission around resuscitation and intensive care use. He ended up dying during that hospitalization with his family around him.

When this sort of scenario arises, I often feel that perhaps I have failed. Is it not my job to have these discussions with my patients? The answer is yes but perhaps not exclusively. Maybe the scenario I outlined above is not inappropriate in some cases. Contrary to our self-image, the evidence suggests that, in many (or even most) situations, we are not always able to provide patients with cancer with all of their primary care, psychosocial support, symptom palliation, or possibly even end-of-life discussions. This is not to suggest that we see our role as merely giving chemotherapy but rather that we need to view ourselves as functioning as part of a team of providers that includes primary care, mental health, nursing, pharmacists, palliative care, and other professionals.

It is worth considering the landmark study by Temel et al¹ published in the New England Journal of Medicine last year. Patients with metastatic non–small-cell lung cancer were randomly assigned to receive usual care with or without an early palliative care intervention. Unexpectedly, those assigned to early palliative care lived longer with a better quality of life. The authors did not overstate the survival benefit, given that it was not the planned primary outcome of the trial, and indeed it needs to be confirmed. However, regardless of whether or not it was a chance finding, this has quieted arguments that palliative care means giving up on a patient to his detriment.

In the article that accompanies this editorial, Greer et al² report on measures of the aggressiveness of care received by patients in that trial. In this study, patients with early palliative care were not denied anticancer treatment—they received the same number of chemotherapy regimens as those in the control group—but they were less likely to have chemotherapy continued close to death and were more likely to enroll in hospice for a longer duration. It is important to note that this group also had better ratings of overall quality of life and mood compared with the control group.

As with all studies, there are some limitations. All patients had advanced lung cancer and were treated in a single tertiary care institution. The practice patterns observed may not be generalizable to other patient populations. Providers and patients in different settings may not respond in the same way to the intervention of early palliative care. In addition, some of the observed differences in care near the end of life may be more a function of these patients living longer beyond the end of treatment rather than of changes in patient treatment preferences or provider practices.

Still, why would patients receiving palliative support end up receiving less aggressive anticancer intervention when the two were independent, parallel processes within the trial? The study reported by Greer et al² confirms observations that aggressive treatment does not prolong survival in metastatic solid tumors³ and that hospice and palliative care do not shorten survival.⁴ The inverse relationship between supportive measures and aggressive chemotherapy use has been reported in other settings,⁵ but prior studies have been observational and consequently prone to confounding influences: if a patient is receiving chemotherapy (at least in the United States), he cannot also be in hospice. The randomized design of this study is a better confirmation of the inverse association. Previous studies have found that living in a region in which there is less hospice availability leads to more aggressive chemotherapy use.⁶ This suggests the possibility that, when we do not have support in providing end-of-life care, oncologists tend to do what we were trained to do: give chemotherapy.

But are not oncologists trained to deal with end-of-life and palliative issues as well? In a study in need of updating, the end-of-life content of 50 textbooks from multiple medical specialties was evaluated a decade ago.⁷ The investigators found that hematology-oncology textbooks ranked 10th among 12 specialties. The only specialties with less end-of-life content were infectious disease and surgery. Family medicine, geriatrics, and psychiatry had the highest percentage of helpful end-of-life content. The implication is that we've been trained to see ourselves in the company of curative specialties.

This is clearly an overgeneralization about the culture of our specialty, but our failure to address end-of-life issues has been documented many times in surveys of patients⁸ and physicians⁹ and continues to be a deficiency in our training.¹⁰ An ethnographic study of the care of patients with terminal small-cell lung cancer described a collusion between patients and their oncologists.¹¹ Focusing on treatment activities allowed everyone to ignore the long-term picture and led patients and their families to develop a false optimism about recovery. Patients in this study eventually got more accurate information about their disease trajectory by observing what happened to other patients in clinic than they got from their oncologist. Unfortunately, both patients and family members often later expressed regret, believing they would have made different choices about treatment and spending time with loved ones had they better appreciated the situation.

I would have thought that more conversations between oncologists and patients about the patients' values and end-of-life wishes are better than fewer,¹² but studies have actually shown that, surprisingly, a majority of patients prefer to have discussions about advance directives with physicians that they do not know, such as an admitting doctor at the time of hospitalization.^13,14 Patients explain this by characterizing their relationship with their oncologist as one that is about optimism: “You go to an oncologist to be cured not to be buried.”^13(p32) Patients report feeling that their advance care preferences are outside the purview of their oncologists and that they do not want their oncologists to face a double-bind of working simultaneously to extend life while planning for death as well. Indeed, patients receiving chemotherapy were more likely to feel this way than patients who were not.¹³

Perhaps, then, an important aspect of having a comprehensive care team with different providers—such as palliative care, primary care, and others—is that our colleagues can serve a role of treatment brokers or providers who are trusted by both us and our patients who can help both of us clarify and communicate our assumptions and preferences about care at the end of life.¹¹ Maybe this is what we are seeing among patients randomly assigned to have early palliative care in the report by Greer et al.¹ Oncologists need to accept the possibility that our patients might be better off if we do not try to do everything ourselves. It takes a village to help our patients through this journey. The quality of end-of-life care can be better when we take a team-based approach to managing incurable cancer right from the start.