The best, most likely, and worst-case prognostic framework is a helpful tool for discussing median survival with patients in a way that enables them to make sense of the data.
Listen to an interview with Dr. Levin in which he goes into greater depth on how to discuss prognosis with cancer patients.
“How much time do I have left?” seems an inadequate way of asking your doctor about prognosis for a number of reasons. First, the question suggests that it can be answered with an absolute number-eg, 1 year-which patients often interpret literally. They feel cheated if they are short-changed, and if they surpass the deadline, they feel as though they are just waiting to die. Statistically, it confounds average and median survival. Survival is commonly measured in terms of median survival, but patients are rarely familiar with the concept of “median,” and understand it to mean “average.” Perhaps more importantly, it does not account for the fundamental statistical principle of range, an essential part of understanding survival curves. Finally, prognosis should be tailored to the individual in order to account for variables such as fitness, comorbidities, newer treatments, and better delivery of care. Prognosis should also offer hope, which an absolute number does not do.
For the patient, pragmatic planning and figuring out how to cope with cancer are essential outcomes of a prognostic discussion. Defined classically by Lamont and Christakis as a “prediction about a patient’s future,” prognosis is a multifaceted construct that includes curability, lifespan, response to treatment, and quality of life. In the same way that checking the weather forecast helps one to plan a weekend trip, understanding the prognosis allows a cancer patient to plan for the days, weeks, months, and years ahead. In other words, the patient has to translate the prognostic statistics by asking, “What do these numbers mean to me and my life?” For this reason, prognostic discussions may occur in a variety of clinical contexts as cancer treatments unfold, and are often a series of discussions and realizations rather than one comprehensive discussion. Prognostic awareness is associated with less psychological distress,[3-5] better end-of-life planning, and improved bereavement outcomes.
First, consider how, in each of the three classic approaches to communication of prognosis-realism, optimism, and avoidance-miscommunication can easily result.
When communication is too realistic, it can be overly detailed and overwhelming, often sounding pessimistic. One patient nicknamed his oncologist Dr. Death because he routinely described every potential disastrous side effect in graphic detail. Without accompanying empathic strategies, a realistic prognosis can seem “brutal”; not infrequently, patients complain of being “hit over the head” by the bad news.
While the optimistic approach is supported by surveys of patients, who say they want hopeful clinicians, an overly optimistic outlook may result in resentment later when reality bites. Blinded by optimism, patients already in the dying phase of their illness may elect to have “more chemo” rather than pursue a more appropriate palliative care approach.
Clinicians who avoid prognostic discussions say things such as, “My crystal ball is broken,” “Everyone is different,” and, “It is the responsibility of the primary oncologist, not the ICU, to discuss the cancer’s prognosis.” Emphasizing outliers such as the “one patient who was cured” is also considered avoidance if that case cannot be extrapolated to the current one. Avoidant communication may leave patients feeling unsupported because without an understanding of their prognosis, they cannot plan.
The approach that I use combines the first two stances into one called “realistic optimism,” which balances hope and realism, and which can pragmatically inform better coping. Coping is classically defined by Folkman and Lazaraus as the cognitive and behavioral efforts used to regulate negative emotions, manage the problem causing the negative emotions, and foster well-being. By discussing prognosis in a realistically optimistic way, the clinician is likely to foster better coping. Improving coping should be seen as a major goal of prognostic discussions.
It is well established that patients have a bias towards optimism when it comes to understanding their prognosis. For example, 4 months after diagnosis, 69% of stage IV lung cancer patients and 81% of stage IV colorectal cancer patients (N = 1,193) believed that the chemotherapy that they were receiving was potentially curative.
Even in terminally ill patients, clinical prediction of survival is overly optimistic and, in one systematic review, was overestimated by at least 4 weeks in 27% of cases. In another study of patients admitted to hospice, for whom median survival was only 24 days, 20% of predictions were accurate (within 33% of actual survival), 63% were overly optimistic, and 17% were overly pessimistic. The better the doctor knew the patient, the less accurate the prognostication. In fact, each year the doctor had known the patient worsened the prognostic accuracy by 12%. Perhaps we clinicians try to shield people we know and like from perceived harm. Physicians’ false optimism may also be linked to the false optimism that is a prevalent attitude in modern society-sometimes called the “tyranny of positive thinking.”
To illustrate the harm that false optimism can do, consider this scenario: a patient wonders aloud if he will die. Family and friends unanimously express their conviction that he will not die; they tell him he needs to “think positive.” Such a patient learns that discussing prognosis is not helpful, and he is left to deal with his fears about death and dying alone.
The PROG-S model for discussing prognosis was developed in our communication training laboratory at Memorial Sloan Kettering Cancer Center. It has five steps (Table 1). Additional helpful concepts are summed up by the acronym NOSI, which is explained below.
Kiely, an oncologist whose research focuses on prognosis, showed that median survival can be translated into best, worst, or most likely outcomes. To illustrate, a 12-month median survival means that half of patients will live longer than 12 months and half will live less than 12 months. The most likely outcome is that the middle 50% on the survival curve would live for 6 months to 2 years (half to double the predicted median). The best case, occurring in about 10% of patients, would represent an excellent response to treatment, with survival beyond 3 years (about 3 to 4 times the predicted median). The worst case would occur in about 10% of patients, with rapid progression and death in a matter of a few months (one-sixth of the median survival).
Using this model, the spread of the median survival data is graphically portrayed with the example of a patient who receives the “good prognosis” of a 5-year median survival (Table 2, boldfaced row). However, the worst case is that 1 in 10 patients with such a 5-year median survival will have a precipitous decline and die within 10 months. Patients who only “see” the 5-year number are at a disadvantage in terms of end-of-life planning and may well pursue inappropriately aggressive care in the face of likely death.
The message from Kiely’s data is that hope is based on statistical underpinnings: 1 in 10 patients will statistically do very well. Even with a 1-year median survival, 1 in 10 patients will live for 3 to 4 years, by which time there may well be newer and more effective treatments-so there are good statistical reasons for the stance of realistic optimism.
One of the ICU attendings at Memorial Sloan Kettering Cancer Center, Louis P. Voigt, MD, expressed hope in the setting of a downward spiral this way: “I think that he is dying but he is very strong-willed. If he proves me wrong and we can take him off the ventilator, I will be very happy to have been proven wrong.”[personal communication, 2010] The point here is that the message of hope and realism can-and should be-put into your own words, but is based on the solid statistical notion of spread.
When discussing prognosis, patients and families will become emotional. What is the rationale for using empathic strategies, other than just being nice?
Empathy builds trust. Without trust, it is impossible for a patient to engage in collaborative treatment decisions about life and death with a clinician that he or she hardly knows. Empathy signals prosocial collaboration, framing the environment as nonthreatening and turning off the “flight, fight, or freeze” reaction that is deleterious to problem solving. The clinician’s aim is to have the patient’s mind in rational learning mode, because a mindset of bracing for threat impedes learning. The “bracing for threat” mindset is easily recognized: The patient hears but does not listen, and asks multiple questions but is not reassured by the answers.
Decatastrophizing, a classic technique used in the treatment of panic and anxiety, helps the patient develop an action plan for dealing with the worst-case scenario. Without this action plan, the feared situation is frozen in time, and the anxiety is sustained because the threat is never processed rationally.
A statement of nonabandonment follows: “If the worst-case scenario comes true, we will do all we can to help you, even if you are facing death and dying…” One study showed that such reassurance can reduce anxiety and uncertainty and improve self-efficacy. However, if you make such a statement, you have to mean it. Promising to help a patient and family through death and dying but being absent when it counts may seem hollow when viewed in retrospect. Articulating that you will do your best with the resources that you have available is reassuring and reflects an ethic of caring.
The “NOSI” acronym stands for the following four additional points, which are helpful to keep in mind when discussing prognosis with a patient:
Numbers, not percentages: Percentages are vulnerable to cognitive distortion. A 5% pay increase may be perceived to be an insult by one worker and a compliment by another. A 30% discount voucher may have one person lining up before the store opens while another throws the voucher straight into the garbage. Instead of percentages, use wording such as, “If there were 100 patients with your type and stage of cancer, then we might expect that 80 would respond to this drug.”
Offer both sides of the coin: One patient was told that she had a greater than 80% chance of being cured, and she made plans to take an out-of-state job as a result. This communication omitted the fact that 1 in 5 similar patients (20%) would relapse and require a stem cell transplant. Mixed framing presents the chances of living/remission as well as the chances of dying/relapse in order to provide a more accurate topographical map of the prognosis. Here is an example of mixed framing: “If there were 100 patients like yourself, in 5 years’ time, 80 would be cured, and 20 might have relapsed lymphoma…”
Summarize in writing: Consider the following statistic about how difficult it is to remember what the doctor says: half of all prognostic information given to cancer patients is not recalled. The more data presented to a patient, the less is remembered. Further, highly charged negative emotional states worsen recall.[16,20,21] Health literacy, the ability to understand medical information, is basic or worse in 36% of Americans. The average reading level is eighth grade, and because of this, the American Medical Association recommends that patient reading material be written at a fifth- or sixth-grade level. Low health literacy and low heath numeracy-although the evidence for the latter is not as strong-are both associated with worse health outcomes. These data make a strong case for presenting a written summary of the best, most likely, and worst-case scenarios, as well as the plan of action for the latter, which reinforces nonabandonment.
Individualize: Extrapolating prognostic data from large cohort studies that may be several years old has numerous limitations, giving the clinician a chance to personalize the data and offer hope. Newer treatments, better delivery of care, fitness, age, family support, education and resources, and fewer comorbid illnesses can all be seen as factors that can improve the prognosis. Folkman, an expert on coping, notes that meaning-focused coping can be used to promote hope. In meaning-focused coping a patient can draw on religious or spiritual beliefs (eg, “God is purposeful and there is a reason for my cancer…”), values, and existential goals (such as finding purpose in life: “I want to heal the rift with my son before I die”) to motivate and sustain coping and well-being during illness. This allows the clinician to personalize the prognosis with less tangible but very meaningful variables such as prayer, meditation, diet, love, altruism, music, and miracles. It is an opportunity for the clinician to draw on his or her own creativity, experience, and values as they intersect with those of the patient. Often, personalizing prognosis with meaning can be folded into the best-case discussion.
The best, most likely, and worst-case prognostic framework is a helpful tool for discussing median survival with patients in a way that enables them to make sense of the data. The PROG-S mnemonic is a useful way of teaching the strategies and skills necessary to negotiate this communication challenge. Patients, instead of asking, “What are my chances?” should be taught to ask, “What are the best, most likely, and worst-case scenarios? What will our plan be in case the worst-case scenario comes to pass and we are facing recurrence or death?”
Financial Disclosure: Dr. Levin has no significant financial interest or other relationship with the manufacturers of any products or providers of any service mentioned in this article.
1. Lamont EB, Christakis NA. Some elements of prognosis in terminal cancer. Oncology (Williston Park). 1999;13:1165-70.
2. Applebaum AJ, Kolva EA, Kulikowski JR, et al. Conceptualizing prognostic awareness in advanced cancer: a systematic review. J Health Psychol. 2014;19:1103-19.
3. Lichtenthal WG, Nilsson M, Zhang B, et al. Do rates of mental disorders and existential distress among advanced stage cancer patients increase as death approaches? Psychooncology. 2009;18:50-61.
4. Chochinov HM, Tataryn DJ, Wilson KG, et al. Prognostic awareness and the terminally ill. Psychosomatics. 2000;41:500-4.
5. Innes S, Payne S. Advanced patients’ prognostic information preferences: a review. Pall Med. 2009;23:29-39.
6. Bradley EH, Hallemeier AG, Fried TR, et al. Documentation of discussions about prognosis with terminally ill patients. Am J Med. 2001;111:218-23.
7. Hebert RS, Schulz R, Copeland VC, Arnold RM. Preparing family caregivers for death and bereavement. Insights from caregivers of terminally ill patients. J Pain Symptom Manage. 2009;37:3-12.
8. Lamont EB, Christakis NA. Prognostic disclosure to patients with cancer near the end of life. Ann Intern Med. 2001;134:1096-105.
9. Campbell TC, Carey EC, Jackson VA, et al. Discussing prognosis: balancing hope and realism. Cancer J. 2010;16:461-6.
10. Folkman S, Lazarus RS. The relationship between coping and emotion: implications for theory and research. Soc Sci Med. 1988;26:309-17.
11. Weeks JC, Catalano PJ, Cronin A, et al. Patients’ expectations about effects of chemotherapy for advanced cancer. N Engl J Med. 2012;367:1616-25.
12. Glare P, Virik K, Jones M, et al. A systematic review of physicians’ survival predictions in terminally ill cancer patients. BMJ. 2003;327:195-8.
13. Christakis NA, Lamont EB. Extent and determinants of error in doctors’ prognoses in terminally ill patients: prospective cohort study. BMJ. 2000;320:469-72.
14. Kiely BE, Tattersall MH, Stockler MR. Certain death in uncertain time: informing hope by quantifying a best case scenario. J Clin Oncol. 2010;28:2802-4.
15. de Vignemont F, Singer T. The empathic brain: how, when and why? Trend Cogn Sci. 2006;10:435-41.
16. Portnoy DB. Waiting is the hardest part: anticipating medical test results affects processing and recall of important information. Soc Sci Med. 2010;71:421-8.
17. Beck AT, Emery G. Anxiety disorders and phobias: a cognitive perspective. New York: Basic Books; 1985.
18. van Vliet LM, van der Wall E, Plum NM, Bensing JM. Explicit prognostic information and reassurance about nonabandonment when entering palliative breast cancer care: findings from a scripted video-vignette study. J Clin Oncol. 2013;31:3242-9.
19. Jansen J, Butow PN, van Weert JC, et al. Does age really matter? Recall of information presented to newly referred patients with cancer. J Clin Oncol. 2008;26:5450-7.
20. Schwabe L, Wolf OT. Learning under stress impairs memory formation. Neurobiol Learn Mem. 2010;93:183-8.
21. Schwabe L, Wolf OT, Oitzl MS. Memory formation under stress: quantity and quality. Neurosci Biobehav Rev. 2010;34:584-91.
22. Kutner M, Greenberg E, Jin Y, Paulsen C. The health literacy of America’s adults: results from the 2003 National Assessment of Adult Literacy [NCES report 2006-483]; in US Department of Education NCFES (ed). Washington, DC. 2006. Available from: http://nces.ed.gov/pubs2006/2006483.pdf. Accessed January 8, 2015.
23. Ad Hoc Committee on Health Literacy for the Council on Scientific Affairs of the American Medical Association. Health literacy: report of the Council on Scientific Affairs. JAMA. 1999;281:552–57.
24. Berkman ND, Sheridan SL, Donahue KE, et al. Low health literacy and health outcomes: an updated systematic review. Ann Intern Med. 2011;155:97-107.
25. Folkman S. Stress, coping, and hope. Psychooncology. 2010;19:901-8.