Survival Expectations and Hope Among Cancer Patients at End-of-Life (SHAPE)

Cancer burden is increasing in Singapore. As Singapore's population increases, so too does the incidence of cancer. A total of 61,519 incident cancer cases were diagnosed in Singapore between 2010-2014. High disease incidence has resulted in a greater fiscal burden for patients and the health system as cancer patients consume a disproportionate amount of healthcare resources.Despite improved treatments, cancer related mortality is high; cancer has overtaken diseases of the circulatory system (including ischemic heart disease) to become the leading cause of death in Singapore. It now accounts for nearly 30% of all deaths. Given these sobering statistics, policy makers are struggling with how best to cost-effectively accommodate the health care needs of this vulnerable population.

Many Patients Want to Know Their Prognosis. A recent systematic review showed that the majority of cancer patients (and their caregivers) expressed a desire for prognostic information and how it varies by treatment choice. However, the preferred style for receiving that information varied. One study showed that nearly 80% of patients wanted a qualitative prognosis (i.e. will they die from the disease) whereas only half wanted a quantitative prognosis (i.e. how long they will survive). Regardless of how the information is conveyed, studies have shown that patients who discuss prognosis with their physicians have a better understanding of the disease trajectory, are more likely to pursue comfort-oriented over life-prolonging care, and are better able to cope with their illness and clarify priorities and goals.

Physicians Often Withhold Prognostic Information. Despite many patient's preferences for prognostic information, the literature reveals a reluctance among physicians' to disclose information about prognosis when it is poor.8 Physicians often find it easier to offer aggressive treatments rather than engage in challenging end-of-life discussions.In Singapore, as with other Asian countries, physicians often collude with caregivers to hide prognostic information from patients. Even when discussed, patients may fail to understand and recall the information presented because of unfamiliar terminology, such as "median survival" or "relative/absolute risk reduction".In many cases, patients are reluctant to ask additional questions about prognoses even though they may desire this information.

Systematic bias in Prognosis. If lack of communication were the sole cause of poor prognostic information, then patient's beliefs about their prognosis would be associated with high variance, but not necessarily biased in a particular direction. Yet, most patients with advanced cancer state unrealistic expectations about the benefits of treatment, believe that their cancer is curable and over-estimate their predicted life expectancy.This bias results from several factors. First, patients who are managing their symptoms well, have strong social and family support, and/or are happy with other aspects of their care may suffer from a form of halo effect, which is a cognitive bias where their chances of a cure are positively influenced by these other properties.Second, patients diagnosed with cancer at early stages are often (rightly) presented with optimistic chances of survival. For those whose cancer progresses, even when informed with the new information, they are less likely to fully update their prior beliefs as compared to someone first diagnosed at a later stage. This lack of complete updating is consistent with confirmatory bias where people give disproportionately less consideration to new information that does not conform to their prior expectations.However, even those newly diagnosed at a later stage are likely to be overly optimistic about their chances for a cure.Although there are negative consequences from this form of optimism bias, such as overtreatment, it can also be part of a successful coping strategy.In fact, the benefits of 'forcing' patients to understand their prognosis may not be worth the costs due to the additional emotional distress it may inflict. This may hold true even if hope leads patients to undergo costly and burdensome treatments with a low probability of success. However, it is still important to understand the extent of divergence between hope and beliefs, the level and type of information that patients would like to receive and how best to convey that information.

Discerning Hope from Beliefs. To discern whether or not patients have an accurate understanding of their prognosis requires disentangling what they believe will occur from what they hope will occur. There are two promising strategies that allow for producing these estimates. The first strategy is to ask patients about their own beliefs, and then to compare their responses to how they believe their physician would respond. If their physician has provided them with accurate prognostic information, then eliciting the physicians' frame should allow for a more accurate assessment of beliefs less likely to be confounded by hope. Consistent with this hypothesis, a study of advanced cancer patients in the US reveals that the degree of prognostic discordance between physicians and patients decreases when patients are asked to respond using the physician's frame of reference.Yet, a second study using a similar approach still found very high discordance, which suggests either poor patient-physician communication or that using the physician's frame does not entirely reduce the confounding of hope and beliefs.

A second approach is to offer what economists term 'an incentive compatible strategy. Consistent with the idea that 'ignorance is bliss', even if patients are aware of their prognosis, many may prefer to express false hope rather than state the truth about their condition. In economic terms their utility (a measure of happiness) is higher when they express false hope. To encourage these individuals to state their true beliefs, they can be offered a reward such that the utility of expressing their true beliefs plus the reward is greater than the utility when they express false hope. Because the reward makes it in the best interest of the participant to reveal his/her beliefs truthfully, such incentive compatible approaches are sometimes known as truth serums.It is often operationalized by stating that the reward will be provided if the responses match the views of a professional (i.e., their physician) or the best evidence from the literature. For example, in one US study, college students were asked to rank the mortality risk from 12 different causes of death. Those offered a reward for correct rankings were more likely to rank the mortality risks accurately for their own age group.