Predicting a “Lazarus Effect” in Patients with Advanced Cancer Near the End of Life: Prognostic Uncertainty, Oncologists’ Emotions, and Ethical Questions

“I think if the crystal ball is telling me this guy is going to die in a week no matter what you do, you don’t treat him. Period. But it’s just not always that clear cut.”

– Oncologist (Bluhm et al. 2016, e1011).

Patients with cancer face multiple stresses near death. Diverse symptoms, contemplating mortality and accuracy of prognostication, loss of functioning, changing family dynamics, financial strain, and frequent interactions with the health care system are the norm (Committee on Approaching Death: Addressing Key End of Life Issues; Institute of Medicine 2015). Many patients experience emotional difficulty even in the absence of mental health comorbidities. Oncologists who treat patients near death often experience emotional difficulty themselves including sadness, anticipatory grief, shock, anger, and regret. They face the prospect of having frequent emotionally draining conversations, often with patients they have known for a long time. This commentary explores an important area of clinical practice introduced by the target article “The Inner Lives of Doctors: Physician Emotion in the Care of the Seriously Ill” (Childers and Arnold 2019): the unique challenges in caring for patients at this stage in the face of prognostic uncertainty. We present an overview of the role of prognostic uncertainty in oncologists’ emotions, why the landscape of prognostication for cancer is changing, and how certain medical advances that have the potential to induce a “Lazarus effect” for patients near death present daily ethical challenges for oncologists (Roeland and LeBlanc 2017).

Cancer is the second leading cause of death globally and in the United States (Committee on Approaching Death: Addressing Key End of Life Issues; Institute of Medicine 2015; World Health Organization 2019). Approximately 606,880 patients with cancer are expected to die in the United States in 2019, and more than 1.7 million new cases are expected (American Cancer Society 2019). Despite clinical research demonstrating substantial toxicity with late chemotherapy, (usually defined as chemotherapy within a defined period of time before death, e.g., 1 month), its use in patients with chemotherapy-refractory metastatic disease is common (Prigerson et al. 2015). Factors that influence provider decisions about administering chemotherapy late in the course of illness include a sense of responsibility for life-and-death decisions, acknowledging the emotionally draining nature of communication specific to cessation of chemotherapy (or not starting it), and prognostic uncertainty (Bluhm et al. 2016).

With the rapid development of new cancer therapies, prognostic uncertainty has made decision-making more complex for patients and providers (Roeland and LeBlanc 2017). The quandary of prognostic uncertainty near death, or the lack of a “crystal ball,” is not new (Bluhm et al. 2016). But exceptional responders with advanced cancer who exhibit a “Lazarus effect,” or a durable response that extends life and occasionally even cures the underlying illness with complete remission, are increasing in number (National Cancer Institute 2018; Roeland and LeBlanc 2017). Exceptional responders are the focus of a new National Cancer Institute Initiative and include patients who respond uniquely to treatments that are not effective for most other patients with the same illness, including patients in early-phase clinical trials in which <10% of patients respond to the medications being investigated and patients who respond to treatments in later-phase clinical trials (National Cancer Institute 2018).

The opportunity to achieve durable response for an illness when a patient is near death has raised the stakes of prognostication for oncologists, ironically at a time when explicit prognostication is often missing or disappearing from medical textbooks. In his book Death Foretold, sociologist and physician scholar Nicholas Christakis describes prognostication as stressful, taxing, and unpleasant (Christakis 1999). He notes, “Prognosis is also stressful because, more than diagnosis or therapy, it is based on probabilistic reasoning, which physicians often regard as problematic and for which they are typically also poorly trained. Physicians see a tension between aggregate, population-based probability statistics on one hand and the application of such data to identified, individual patients on the other” (Christakis 1999, 88). Exceptional responders amplify the tension from prognostic uncertainty.

Recent exploratory research is notable for elucidating factors that influence oncologists when they prescribe chemotherapy near death. In one analysis of 17 in-depth interviews with oncologists who practice in a variety of settings, participants noted dealing with two types of prognostic uncertainty. The first uncertainty was predicting the time of death, and the second was knowing whether chemotherapy will ultimately help. Notably, all “seemed to be able to call to mind cases of the rare patient who improved against all odds after late chemotherapy” (Bluhm et al. 2016, e1009). While it is possible that physicians may make decisions to avoid having difficult conversations or interactions with patients without realizing it (Childers and Arnold 2019), this research suggests otherwise (Bluhm et al. 2016). Oncologists acknowledged the emotionally draining nature of communication specific to cessation of chemotherapy or not starting it to begin with, and a staggering sense of responsibility for life-and-death decisions. In the words of one oncologist interviewed in the study, “It’s pretty emotional, it’s pretty intense in the room. You can kind of feel the heaviness, and it almost feels like every time you do this, a ton of bricks has fallen on top of you, if you let yourself go there” (Bluhm et al. 2016, e1011). Another states, “It’s a hell of a lot easier conversation to say, ‘Alright, let’s do your next round of chemo. I’ll see you in 3 weeks,’ rather than have a conversation to say I don’t think we should do more chemo.” Yet another describes dissociation (a classic trauma response) while delivering bad news: “There are times when I feel like I’m having an out-of-body experience because the things I’m saying are coming out so routinely that I almost feel like I’m listening to myself or watching myself” (Bluhm et al. 2016, e1011).

Patient factors that influence oncologists’ decisions around prescribing chemotherapy near death were also notable: some acknowledged chemotherapy may be used as a form of “palliation for emotional symptoms,” and others noted it could be used as “a means to preserve hope” (Bluhm et al. 2016, e1009). Patient wishes regarding the pursuit of treatment and also patient ability to weigh trade-offs were important patient-level factors identified in this and other studies. Speaking of the ability to weigh burdens and benefits in making decisions, one oncologist interviewed for a newsletter of the American Academy of Hospice and Palliative Medicine noted, “If there is even a 2% chance of a long-term remission with an immunologic, many patients and families are willing to take those odds, and I worry that more patients are spending their final moments undergoing aggressive therapy in the hospital and clinical setting rather than focusing on quality of the end of their life at home” (Roeland and LeBlanc 2017).

What happens to oncologists’ emotions when accurate determinations of clinical efficacy and prognostication become more difficult with the presence of newer therapies for previously incurable cancer and with the increasing presence of exceptional responders? Physicians must balance a host of factors that they may not have had to face using older prognostication guidelines with cytotoxic chemotherapy. For medical advances such as immunotherapy, which hold the promise that a small subset of the population may have a “Lazarus effect” near death, the question becomes wrought with ethical and emotional challenges.

Several important research and policy gaps in our knowledge base about the emotional and behavioral influence of prognostic uncertainty are present: How do oncologists communicate the risks and benefits of newer therapies or interventions when cancer is metastatic, and traditional cytotoxic therapies have failed? In the process of respecting autonomy, how does an oncologist ensure that patients’ preferences, values, and beliefs are honored in the decision-making process when the potential benefit of cure is uncommon but highly desired? What criteria do oncologists use in order to choose which patients with metastatic disease are eligible for newer experimental drugs? What role does a patient’s education have on the process by which interventions are requested by patients or their family members? Ethical issues involving justice and resource allocation include challenges with engaging marginalized groups of individuals who may not have knowledge about or be able to afford costly interventions such as immunotherapy. How do oncologists discuss use of expensive newer drugs when insurance doesn’t cover the cost, and patients may not be able to afford it? How do oncologists feel about the fact that billions of dollars of public funding contributed to published research associated with these drugs and yet their patients may not be able to afford them? (Cleary et al. 2018).

Additionally, a further complication is that patients’ desire to access newer therapies near the end of life may cause delay in accessing approaches with known benefits, such as palliative care and hospice. Because drugs like immunotherapies are considered potentially curative, the Medicare hospice benefit will not pay for them, which may lead to unintended delays of entry into hospice care. How do oncologists factor this knowledge into their decision making about when and how to offer new treatments to patients? The list of important practice-related ethical questions is seemingly endless, and yet oncologists must contend with these questions every day.

Dying in America, a landmark report published in 2015, states “The overall quality of communication between clinicians and patients with advanced illness is poor, particularly with respect to discussing prognosis, dealing with emotional and spiritual concerns, and finding the right balance between hoping for the best and preparing for the worst. Ample evidence documents structural and financial disincentives for having these discussions. In the absence of adequate documented advance care planning, the default decision is to treat a disease or condition, no matter how hopeless or painful. A result of inadequate advance care planning, therefore, can be more intensive treatment, as well as more negative impacts on family members” (Committee on Approaching Death 2015, 12). Perhaps the time is ripe for a more formal social science exploration of provider emotions and experiences, especially for those who experience emotional turmoil, grief, sadness, or regret when large numbers of their patients die receiving therapies that may cause harmful side effects while hoping for the highly desired but often elusive “Lazarus effect.”