Discussing prognosis with patients who have advanced cancer is associated with more realistic patient perceptions of life expectancy and more advanced care planning, and it doesn't appear to harm the patient-physician relationship or patients' emotional well-being. That was the conclusion of a recent analysis of the multi-site, prospective Coping with Cancer study, published in the November 10 issue of the Journal of Clinical Oncology (2015;33:3809-3816).
The findings reinforce the importance of conversations about life expectancy, according to the study's authors and other experts asked to comment for this article.
The analysis also suggests, however, that quality prognostic conversations are not occurring frequently enough: While 71 percent of the 590 patients in the study wanted to be told their life expectancy, only 17.6 percent recalled such a conversation with their physician.
“When we care for patients, we try to protect them from a lot of this news because it's devastating. I don't think our findings should be taken to mean that each and every patient should be told how long they have to live, but these data show that we grossly underestimate how many of our patients want this information, and how powerful these conversations can be,” said the study's lead author, Andrea C. Enzinger, MD, Attending Physician with the Adult Palliative Care Program at Dana-Farber/Brigham and Women's Cancer Center and Instructor in Medicine at Harvard Medical School.
In this sense, the findings also highlight an increasingly appreciated dichotomy between the desire of many patients to understand their prognosis and physicians' level of comfort and skills for having effective conversations, the researchers said.
The National Cancer Institute's Coping with Cancer study—various aspects of which have been analyzed in other published studies—enrolled patients at the outpatient clinics of eight cancer centers from 2002 to 2008. Patients had metastatic cancer and disease progression after one or more courses of palliative chemotherapy. They underwent a 45-minute interview at baseline and then were followed until they died. The researchers used the National Death Index to record deaths that occurred after the study site closures.
This most recent analysis looked at responses to the questions “Have the doctors talked with you about how much time you have left to live?” and “If your doctor knew how long you had left to live, would you want him or her to tell you?” Patients were also asked to estimate their own life expectancy, and they were asked whether they trusted and respected their physician.
Various tools were used to measure mood, anxiety, and mental illness, and patients were asked if they had completed advance care planning and if they preferred end-of-life care that is focused on life extension or palliative care.
Numerous studies have shown that patients with advanced cancer overestimate their life expectancy and are overly optimistic about the benefit of chemotherapy. The findings from this study were no different: Approximately half of the patients were willing to estimate their life expectancy, and most of those patients who shared estimates were indeed overly optimistic. Yet there was a significant difference in expectation between the patients who recalled receiving a diagnostic disclosure and those who did not.
“Patients who recalled a prognostic conversation estimated they had about a year to live, on average. Patients who didn't recall a conversation estimated four years, on average, which is extremely unrealistic... These were patients with very advanced cancers—on average, they lived only five months,” Enzinger said.
“So when patients had conversations with their doctors, they were still optimistic, but it was much more realistic,” she said. “I found this effect really powerful.”
Patients who had a more realistic understanding of their life expectancy were significantly more likely to complete a DNR order, living will, or health care proxy, and significantly more likely to prefer palliative care over life-prolonging care. Moreover, they were no more likely to have a worried or anxious mood or to meet the criteria for major depression or generalized anxiety disorder than patients who did not recall a prognostic disclosure. And there were no differences in the proportion of patients reporting a strong relationship with their physician.
The study's senior author, Holly G. Prigerson, PhD, principal investigator of the larger Coping with Cancer study and the Irving Sherwood Wright Professor in Geriatrics and Director of the Center for Research on End of Life Care at Weill Cornell Medical College, said that physicians should feel reassured about the results:
“The data show that prognostic disclosures won't harm the physician-patient relationship, and that in most cases they'll actually be helpful. ... There are few downsides,” she said.
The problem is, physicians are sometimes hesitant to discuss life expectancy. Researchers have described a variety of reasons over the years, from concern about damaging patients' well-being and concern about the accuracy of their estimates, to uncertainty about whether patients want specific prognostic information and uncertainty about how to have such conversations.
Added to this unease is a layer of differing perception and perspective of physicians and their patients. “Patients and physicians often come away from conversations with very different perspectives,” said Christopher Daugherty, MD, Professor of Medicine at the University of Chicago Medicine, who has studied doctor-patient communication.
“The oncologist will come out and say ‘I had a prognostic discussion because I always do.’ And the patient will come out of that conversation and, most of the time, will say ‘we didn't talk about prognosis.’”
Longitudinal Studies Needed
Does this mean that oncologists' messages are too vague or too nuanced? Or are patients misinterpreting or missing what doctors say, or not getting the information they need? Or are various factors at plan? Right now, it's hard to say exactly, said Areej El-Jawahri, MD, an oncologist specializing in hematologic malignancies at Massachusetts General Hospital.
“What's missing in the field are longitudinal studies that look specifically at the conversation or conversations taking place between patients and physicians,” El-Jawahri said. “What we're seeing in this study is recall of a conversation. But we don't know the content of the conversation and how the information was conveyed.”
The study also did not capture prognostic conversations that might have occurred after the baseline assessment. This is important to note because “helping people understand their prognosis is a process... it may actually involve multiple conversations over time,” she said. “Most patients want to know [their life expectancy], but they want information “on their own terms and in their own timeframe.”
Also asked for his perspective, Anthony L. Back, MD, Professor in the Department of Medicine, Division of Oncology, at the University of Washington School of Medicine, suggested asking patients how they would like to talk about the future and whether they want statistics on life expectancy at that point in time: “Patients often think of [prognosis] as, what can I expect in the future? What kind of plans should I make?
“It's not always prognosis in the sense of median survival,” said Back, who also has researched patient-physician communication. “It's up to us as oncologists to figure out what patients are really asking for, and what they need at that point in time.
“Some patients will say “I don't' want to know the numbers.' But virtually everyone wants to know about their future in some way, and the challenge for me is to help them talk about it... so they can do even a little planning,” he said.
Prigerson said she has questions about the delivery and characteristics of prognostic conversations, and plans to publish more insight soon. With an NCI Outstanding Investigator Award, she is leading a series of studies examining how various types of communication influence patients' understanding and end-of-life choices.