In the traditional “Marcus Welby” model of the physician-patient relationship, the wise physician cared for the passive patient. Of course, physicians would ask for consent before doing a procedure, but there was no expectation that the patient would decline. In the last several decades, however, it has become widely accepted that some medical interventions involve complicated trade-offs, such as a better quality of life if you survive the heart surgery. Although there may be a “medically best way” to do a specific procedure, the statistics on immediate and long term survival well-known, and the surgeon willing to operate, whether to have surgery requires patient input.
Bioethicists and physicians have been discussing for many years what is called shared decision making, (SDM). It entails having patients directly engaged in choosing the care they will receive, to whatever extent they wish to be involved. Exactly how patients are, or should be, involved in such choices, however, is the subject of both professional debate and ongoing research, including some work by my colleagues. For some people, the notion of SDM implies that patients be fully informed about the available clinical options. Because the clinical literature can be difficult to comprehend for a lay person, patients are implicitly expected to ask their physicians, or physicians to proactively offer, to discuss the pluses and minuses of each treatment option based on what they know about the patient’s particular condition. Some SDM interventions even use a visit reminder suggesting that the patient think of three questions to ask. That notion of SDM, however, is like a consumer deciding on the purchase of a new personal computer—the consumer finds an expert to explore their needs (e.g., word processing, heavy video game use, portability) to help inform the "purchase" decision. The consumer and computer geek engage in an unemotional conversation about features and preferences. This model might be appropriate for decisions about a knee replacement to improve mobility in which the patient slowly becomes aware of the problem, there is time to meet with various physicians, explore options, and quality of life, not life itself, is at stake. Other people who study SDM in practice, however, see a more complicated and nuanced set of interactions, even for such choices about which open discussions might be expected. In some clinical situations, such as advanced cancer, conversations may be even less straightforward.
Previous posts focused on how in our oral narratives both Lori and I unconsciously re-shaped information that we had clearly noted at the time. Those posts highlighted how we unintentionally focused on some facts and ignored others. It is one thing, however, to mis-remember or forget information received. It is another not to have the information. In this post I describe instances in which information was not conveyed, questions were not asked, information was either withheld or not sought, or words were not used.
Information Not Conveyed Immediately
Lori’s narrative was that she knew something was seriously wrong when she met with Dr. B, the internist she saw about her ear problem in advance of our trip to India. Her story was that he told her during the visit that the sudden explosion of seborrheic keratosis she experienced was known as the “sign of Leser-Trélat,” associated with an underlying systemic illness like cancer.
In fact, Dr. B’s note in Lori’s medical record reads: “I told her that, in rare cases, this can be due to chemical signals related to far more serious systemic problems (without specifically mentioning malignancies or the sign of Leser-Trélat). I recommended she see her PCP for a follow-up visit to discuss this.” His note made explicit that he did not mention cancer or Leser-Trélat.
I don’t know exactly what he actually told her, but early the next morning Lori sent a non-urgent e-mail message to her usual PCP, Dr. C. “Yesterday I saw Dr. [B] for a minor problem with my ears. As he was describing that it is a form of dermatitis, I mentioned to him that I've had a huge explosion of seborrheic keratoses all over my body in the last couple months. He suggested that there could be an underlying systemic cause and might be good to followup with you. I'm in town this next week, and as you know, we are going to India Jan. 11-25.” Lori had an excellent memory—she used the word “systemic” just as he did. Had Dr. B actually mentioned the sign of Leser-Trélat, I am sure Lori would have included that in her e-mail.
Early Monday morning, Lori’s e-mail was read by a staff member at the clinic. The staff member messaged Dr. C who immediately replied to Lori that she’d be happy to see her. Lori then scheduled an appointment for Wednesday, the next time she would be home. Wednesday’s appointment led to lab work, an X-ray and then a CT scan Friday morning. When the results of the CT were forwarded to Internal Medicine mid-day Friday, Dr. C was not in. Another physician in the office reviewed the results and tried reaching Lori, leaving messages on both of her phones to call back. Lori responded within the hour and that physician noted, “She prefers to have results discussed over the phone.” That conversation led to the PET CT to finally determine that the CT findings were cancer.
Both the above examples reflect clinicians showing significant concern for how “bad news” is delivered. In the first example, Dr. B thought Lori might have the sign of Leser-Trélat, but that is just an indication, not an accepted test for cancer. With no prior connection to Lori, he was hesitant to give her what might be devastating news. Instead, he provided enough information so she would follow up with her usual internist. His message had the intended results. In the second example, when the very concerning findings from the initial CT scan came back, the internist who received them first checked with Lori to determine if she wanted to come in to hear in person about the CT results. Some people would have wanted to receive what would likely be bad news in a face-to-face meeting.
Sharing the News—or Not
How physicians deliver news is one thing; how patients share it is another. Every person is different, as is their response to receiving a diagnosis of cancer. Lori’s career as clinical psychologist shaped at least one aspect of her response. During those three decades she treated at least four patients who came to her after their therapists suddenly died or could no longer care for them. Feelings of “abandonment” exacerbated whatever problems the patients had been addressing with their previous therapists. Lori was passionate that every psychotherapist should have a professional will outlining what emergency steps should be taken in the event of sudden death.
Lori knew her death wasn’t imminent. Chemo, however, would make it difficult for her to give her patients the full attention she felt they deserved during every visit. Thus, during that first weekend after getting the news, she decided to transfer her patients to other therapists and close her practice. Complicating this plan was that some of her patients were connected indirectly with friends of ours. Sharing the news with friends could risk some patients discovering Lori was ill before she could meet directly with them—and that might cause them harm.
Lori shared her (still tentative) diagnosis with only our immediate family and a handful of friends who questioned why we suddenly cancelled Friday night social plans. At gatherings on Saturday we cheerfully accepted best wishes for our India trip scheduled to begin the next day. For much of the next two weeks, as Lori identified therapists willing to take on each of her patients and then met with each patient, we maintained e-mail silence, sneaked into medical appointments, and otherwise avoided our large circle of friends who thought we were half-way around the world.
While triggered by a very specific and quite unusual reason, this ten-day period allowed us to begin processing the information, some of which was changing day by day as we got new test results. Not sharing much immediately meant we did not have to repeatedly correct people’s understanding of changing facts, as well as update them on how we were dealing with the totally shocking news. During the peri-diagnosis period the typical patient and immediate family face enormous challenges in acquiring and processing information. Broad-based community support is wonderful, but for some it may be easier to accept and “manage” such support after the initial shock has worn off and all the facts have been assembled.
Questions Not Asked
When our internist, Dr. C, called Lori late on Friday after receiving the PET CT results, she delivered the news that the scans showed metastases. She also talked about the seborrheic keratosis being the “sign of Leser-Trélat,” but cautioned us to “not go on the Internet.” At the time, I understood this to mean that without biopsy results and other tests (still to be scheduled for the next week), whatever we would find on the Web would be relatively meaningless. Having done many Web-searches for my own research, I know how confusing those can be unless one is an expert in the field and even experts are handicapped without having all the information. We had doctorates, but not in cancer.
What is interesting, however, is that I didn’t do any searches for articles on the “sign of Leser-Trélat” (LT) until when I began these blog posts long after Lori died. There actually isn’t very much in the literature. A few authors question the scientific validity of the “sign.” Others are less circumspect: “The presence of LT is quite ominous suggesting a highly aggressive tumor. The incidence of metastatic spread has been reported to be 57%, and average survival times have been estimated between 1 to 2 years.” http://ascopubs.org/doi/full/10.1200/jco.2008.17.9143 Another source is even less optimistic, “Most of the malignant tumors associated with this sign behave in an aggressive manner. Thus, the prognosis of patients with the sign of Leser-Trélat is not good; the survival rate averages about 10.6 months.” http://misc.medscape.com/pi/iphone/medscapeapp/html/A1097299-business.html Reading those articles underscores the wisdom of Dr. C’s admonition to not go on the Internet that weekend; it would have been very concerning.
A second incident, moreover, offers some insight into why I never did searches until after Lori’s death. When we saw Dr. E at UCSF on January 23rd he was somewhat more optimistic about life expectancy for upper GI cancer (he reported it as 16 months instead of the 12 months mentioned by our oncologist). Importantly, according to my notes for that visit he said, “but bet on the long side…based on how [Lori] looks.” It was clear to me that he was referring to the fact that she had no symptoms, other than the LT, and had been generally in excellent health. His assessment made perfect sense and it reassured us that Lori would be able to enjoy Jana and Jason’s wedding which was 7 months away.
Even though Dr. E was an expert in GI cancer and had a huge referral practice, he had never personally seen a patient with LT—a clear indication to us of how rare it is. I remember that while sitting there I wondered whether Lori’s presenting with LT should be factored into his prognosis. It seemed relevant to me, but Dr. E hadn’t explicitly mentioned that in his optimistic response. I decided, however, not to ask the question. The researcher in me had quickly reasoned that given how rare the condition is, any oncologist treating a patient with LT who survived longer than the published median would certainly warrant publication of a “case report” in the cancer journals. Case reports are typically used to bring to the profession information about an unusual occurrence. For example, although LT was a sign of a “bad cancer” (GI cancers are usually discovered late and have poor prognoses), if a particular patient with LT lived months longer than would be expected for that type of cancer, the LT might be seen as a relatively good sign. Like “the dog that didn’t bark,” however, the absence of such case reports suggested to me that LT is likely to indicate average or worse than average prognoses for a GI cancer. I thought that asking Dr. E about the LT risked a response of “I don’t know” or “well, perhaps we shouldn’t be so optimistic.” Either answer would increase our anxiety about whether Lori would make it to the wedding. As I didn’t think the answer would impact any treatment choices, I never asked the question.
Words Not Said
During our second visit with Dr. D we discussed Lori’s life expectancy of about a year. We then focused on the differences between two possible chemotherapy regimens that differed largely in how Lori would get one of the two drugs in the “cocktail”—via pills or a “take home” pump worn for two days. It was clear neither regimen would cure the cancer. Any doubt as to their purpose was dispelled by the answers to Lori’s e-mailed question that night about what would happen if she didn’t take the chemo. Without chemo, her symptoms would worsen within a six-month timeframe, i.e., before the wedding. My extensive and detailed notes during the meeting repeatedly refer to the treatment regimens, their names, etc.
Dr. D’s notes, however, included something my notes did not: “We had long discussion re natural history of disease. Explained that goal of treatment is palliative. Median life expectancy as short as 12 mo. Recommended palliative chemotherapy. Many chemotherapy regimens appropriate. Due to relatively favorable adverse effect profile, recommended XELOX.” If she used the term “palliative chemotherapy,” I’m sure I would have noted and remembered it. (Dr. D has subsequently confirmed she didn’t use the term.) Discussions with others indicate that many oncologists are reluctant to use the term “palliative” even if the purpose of treatment is clear, as was the case with us. During our visit to UCSF with Dr. E we also discussed the various regimens, but he also never used the term “palliative chemotherapy.”
Why is this important? We knew the cancer was incurable and that the intent of the chemo was to postpone the debilitating symptoms of the cancer. We would have agreed that is the definition of a palliative treatment. For the next six months, however, Lori and I spoke frequently about how she was doing, always using the term “chemo.” Nearly everyone knows people whose cancers have been cured or gone into remission after chemo. Lori and I weren’t expecting that to happen, or even hoping that would be the case—we knew the trajectory. Friends and relatives hoping for the best, however, could be lulled into thinking this chemo was like that which worked so well for someone else.
Had Dr. D used the term “palliative chemo” we might have begun to use those words, thereby reminding others about Lori’s expected course. (I’ll address the role of expectations in a future blog.) It might have even been as effective had Dr. D (and then we) called it “blue chemo, that is, the name used for a treatment regimen intended to manage symptoms of the cancer, but not cure it.” Perhaps we just need a new term.
One Question, Two Answers
To date, these postings have focused on the first month of Lori’s Journey where much happened in a highly compressed time period. Near the end of the Journey, however, another example of a communication issue occurred. Lori had formally transitioned into palliative care in early November; a month later she shifted to home hospice care. She had a routine monthly appointment scheduled with Dr. D on December 23rd. I checked whether we should meet and Dr. D thought we should do so. I saw this as our “good bye” visit.
We had had multiple discussions with the physician and social worker in palliative care, as well as with the physician and nurse from hospice. We had a reasonable understanding of Lori’s expected final trajectory—probably continued slow worsening and then a relatively sharp terminal decline to death. Dr. D noted Lori’s weight loss and confirmed that this was usual for cases such as this. I asked when we might roughly expect the rapid decline. Dr. D said, “It could be two to three months.” Somehow, this didn’t seem consistent with Lori’s increased weakness, her falling weight, and that we had just finished discussing ordering a walker. I said nothing, but looked hard at Dr. D. Over the months we had developed an excellent relationship. “It could be two to three weeks” was her second answer.
I had a practical reason for the question. Lori‘s out-of-town friends and family wanted to see her and I was trying to manage schedules and expectations. Knowing it might be less than a month (Lori died just over 3 weeks later) allowed me to plan appropriately. I know that prognoses are far from precise or guaranteed. The issue here wasn’t accuracy—Dr. D could easily have said, “it might be as soon as a few weeks, or it might be as long as several months.” The problem was that Dr. D did not know whether we wanted to hear an optimistic or a conservative answer.
There is a substantial body of research on communications between physicians and patients. Much of this is focused on interactions in primary care settings. Other work focuses on interactions that occur in the hospital, especially around end-of-life decisions in intensive care units. Most such research requires informed consent, but being enrolled in a “research study” may change the way communication occurs. Not being part of a research study, I offer these reflections recognizing they may be colored by looking backwards as well as by my personal involvement.
Physician-patient communications are more than just transmission of facts, as if around a computer purchase. In a caring health care system, clinicians transmit information in a compassionate way. Dr. B purposefully did not tell Lori he thought she had the sign of Leser-Trélat and what that implied to him. He wanted her to get that information after an appropriate work-up from the internist who had known her for years. He conveyed just enough to lead her to follow up without her becoming too anxious. Lori actually did not share anything with me until after seeing Dr. C, probably to avoid my being anxious.
Our personal medical care “system” worked quite well. As soon as a staff member saw Lori’s request for an appointment and Dr. B’s internal note in the medical record, she alerted Dr. C about Dr. B’s concerns. (I can only speculate on what would have happened had Lori not followed up on his advice to see her internist. Since the electronic medical record is shared, my guess is that he might have sent Dr. C an e-mail about the concerning signs he observed in one of her patients.) When Lori’s CT results indicated possible metastases that required a PET CT for confirmation, the internist getting the results first asked whether Lori wanted to come in to discuss the report, or preferred to get the results immediately over the phone. In the absence of specific knowledge about how a person wants to receive very concerning news, it is best to balance sensitivity with timeliness. Weeks of delay before active engagement begins may be clinically problematic; taking an extra few hours to communicate devastating news in the way each patient wants is both compassionate and may help subsequent processes go far more smoothly.
For shared decision-making discussions regarding cancer care, the information needed is far more complex, important, and emotionally fraught than most non-medical situations in which expertise is needed, e.g., before buying a new computer. Initial interactions may impact the patient’s connection to his or her physicians who will be guiding, and sometimes delivering, whatever interventions are chosen. Typically, communications with a primary care physician can build on relationships that may have developed over years. If a patient is in the ICU, communications may occur with many different physicians with little time for building a relationship.
When a patient is newly diagnosed with advanced cancer, s/he needs an expert to discuss the various treatment options, which may include surgery or radiation, as well as chemotherapy. Those discussions will typically depend on not just imaging, but also biopsy results and sometimes genetic testing which may take several weeks. Answers are rarely clear—our two oncologists disagreed on whether median life expectancy for Lori’s cancer was 12 vs. 16 months. They also disagreed on the best specific chemo regimen, but agreed it was largely a matter of preference because there were no “head-to-head” trials comparing the two.
In our case, the decisions were relatively easy; Dr. D was willing to use FOLFOX, the regimen preferred by Dr. E, rather than her preference, XELOX. (XELOX involved taking multiple large pills for two weeks each cycle—Lori disliked taking any pills, especially large ones.) This, however, is not always the situation. Patients appropriately seek advice from various clinicians who may have markedly different, and sometimes strongly held, beliefs as to the superiority of certain treatment strategies. From the patient’s perspective, however, there is much more to a treatment choice than the objective numbers in published studies, which often just focus on survival. Unlike many surgeries, both chemotherapy and radiotherapy are expected to entail repeated visits over months. How should one weigh the convenience of seeing a local oncologist with one opinion vs. a distant oncologist (perhaps in an academic center) with a differing, but also strongly held opinion about the “facts”? There are important clinical tradeoffs once one moves beyond the simple metrics on life expectancy. How does one weigh a treatment offering an extra month of median life expectancy but with substantially worse side effects?
More subtle issues may also be at play, especially with respect to interactions with oncologists. When choosing one surgeon over another, a patient rarely needs to be concerned about the feelings of the “rejected” surgeon. One would hope this is also true for oncologists. Not knowing the expectations for professional behavior, however, patients may draw from their own experiences with interpersonal interactions. Many of us have had friends or family who asked for advice and then didn’t follow our recommendations. This may have led us to resent their wasting our time, and even be reluctant to repeat the experience. Patients wanting to rely on a convenient community-based oncologist may be concerned that even seeking a second opinion could lead the local physician to be a less strong and reliable advocate during the rest of his or her cancer care. (Work by my colleagues indicates that even well-educated patients are reluctant to ask questions they feel might impact the socially expected roles of physicians and patients. Frosch et al, 2012 http://content.healthaffairs.org/content/31/5/1030.abstract)
Aside from my expectation that the Leser-Trélat implied a more pessimistic prognosis, I also thought that simply asking the question might affect our potential long-term relationship with Dr. E. Although we hadn't met previously, Dr. E had very graciously offered to see Lori immediately—even later on the same day I made the request. We were likely to have treatment locally, but I anticipated requesting additional consultations and relying on Dr. E for potential referrals. UCSF is a teaching institution, so a trainee was present in the room during our visit. My asking about something potentially relevant to the prognosis that Dr. E. hadn’t mentioned might have risked our newly developing relationship. This concern may have been totally unfounded, but given how vulnerable we felt after getting the diagnosis, I was reluctant to take on any additional risks. We were well-educated and well-connected, yet such concerns impacted our communication.
Unlike my conscious decision not to ask Dr. E. about the implications of the sign of Leser-Trélat, such concerns may not be recognized by most patients. Living with a clinical psychologist probably accounts for my awareness of conscious and unconscious factors in interactions. If issues of physician-patient hierarchy unconsciously influence the questions asked and thus the ways patients in general communicate with their physicians, then an additional professional on one’s care team might facilitate productive interactions. Having a Cancer Journey Companion (CJC) or similar person present with the patient and immediate family from the very first meeting with the oncologist may help.
Our experience, as well as a small, but growing body of literature, suggests that oncologists are in a very difficult position in seeking to help their patients. The clinical choices are rarely clear cases of “for this tumor treatment A is the best in dimensions—survival, side effects, convenience, etc.” Even if the only thing that mattered to the patient was survival, choices are not always clear. Suppose Treatment A has a median survival figure of 18 months, but 20% of the patients die within 6 months. Treatment B’s median survival figure is 12 months, but only 1% die within 6 months. Treatment A has statistically superior survival statistics and would be the “right” choice for most people. Since Lori cared so much about being present at Jana and Jason’s wedding, in this hypothetical case she might have chosen B, with its shorter median survival but much greater likelihood of being alive in August.
The treatment options presented to us differed primarily in convenience—going home after infusions with a pump that would then be removed after two days vs. taking fairly large pills for 14 days. The choices were fairly straightforward and not laden with emotion—Lori didn’t have to decide between a shorter life expectancy to increase the certainty of surviving a relatively short period. When we were having those discussions, barely two weeks from getting the diagnosis, we were still in shock. Because our choices were “simple,” I didn’t ask about more than the minimum necessary for making decisions, trusting our oncologists to tell us what we needed to know. Asking for information that was not “actionable” didn’t make sense to me. It risked jeopardizing both our own emotional well-being (e.g., potentially learning that Leser-Trélat is associated with particularly poor prognosis, even for GI cancer) and our potential relationship with oncologists upon whom we would have to depend. That reflects, however, my own preferences on how much information I want. Some people want to be as informed as much as possible about their condition and all the treatment options. They reach out not only to all the available physicians, but search the Web for new treatments and clinical trials. Others may ask no questions at all and implicitly or explicitly leave all decisions up to their physician. Each person may have a different preference, and preferences may change over time. Preferences may also differ across the patient, partner, and other family members.
Communication is critical for engaging patients in decisions about their care, but communications in medical situations are complicated. If there had been a cancer journey companion listening to the conversation about treatment options, the CJC might have offered a way (either in the moment, or afterwards) to describe to others the rationale for Lori’s chemo or to sort through possible questions or answers. Not being the clinical decisionmaker, the CJC is likely to have been seen as someone who could be asked whether Dr. D might be upset about certain questions. Likewise, by being able to focus on the emotional state of the patient and immediate caregiver, the CJC could let the oncologist know how much, or how little, information they wanted. I might have reached out privately to the CJC with my question about Leser-Trélat potentially impacting the prognosis and she might have helped me explore whether the question should be posed. The CJC might have even gently asked the reason for my question in December concerning how much time before we could expect a rapid decline to death.
These conversations are also difficult for oncologists. They are skilled in identifying treatment options in a rapidly changing field, but unlike mental health professionals, often have little formal training in “reading between the lines” when communicating with patients and family. The oncologist is typically dealing with multiple patients during a day, perhaps going from needing to quickly adjust medications due to an allergic reaction by one patient, to attempting to arrange for a referral for another, to having an initial visit with a patient to whom bad news needs to be conveyed, to simply being with another nearing the end of her journey. In an ideal world, every oncologist would be well-trained and at each visit be “fully present,” have ample time, and able to immediately engage with patients, learn what they want—and don’t want—to hear, and help them over time to be as involved in decisions about their care as they want to be. In the real world, however, we may need additional resources and people to make this happen. The CJC could potentially benefit the oncologist, the patient and the patient’s family. They have role well-worth exploring.