Article Text

Download PDFPDF
Problems with precision and neutrality in EOL preference elicitation
  1. Yael Schenker,
  2. Robert Arnold
  1. Division of General Internal Medicine, Section of Palliative Care and Medical Ethics, University of Pittsburgh, Pittsburgh, Pennsylvania, USA
  1. Correspondence to Dr Yael Schenker, Assistant Professor; Director, Palliative Care Research, Division of General Internal Medicine, Section Palliative Care and Medical Ethics, University of Pittsburgh, 230 McKee Place, Suite 600, Pittsburgh, PA 15213, USA; schenkery{at}

Statistics from

The article by Kamm identifies ‘potential pitfalls’ in current approaches to providing patient-centred advanced illness and end-of-life care (EOLC). Specifically, Kamm cautions against ‘nudging and framing effects’ in discussions to elicit patients' preferences that could produce ‘skewed or biased results’. Kamm further argues that precision and neutrality in questions designed to elicit patient preferences may help to ensure that preferences are ‘reliably elicited’ and avoid mistakes in EOLC decisions. In response, we highlight the problems with a focus on precision and neutrality in discussions about end-of-life (EOL) preferences and advocate an alternative approach that we believe is more likely to achieve patient-centred care.

The problems with precision and neutrality

Kamm describes eliciting patients' values via precise and balanced trade-off questions, with increasingly specific follow-up questions used to pinpoint patients' wishes and ‘avoid mistakes in EOLC decisions’. This approach does not resonate with our experience conducting EOL conversations with seriously ill patients and, we would argue, it risks biasing preference elicitation for different reasons.

First, Kamm's model assumes that people have preformed preferences for EOL treatments. She suggests that the goal of the interview is to elicit these preformed preferences in a way that lets the ‘truth’ come out. In contrast, we argue that preferences are discovered in conversation, a process in which patients reflect on priorities, evaluate their importance and consider their implications. A truly autonomous decision requires an open exchange; it is more like counselling than a biopsy.1 To promote autonomy, the clinician's job is to help patients reflect on and explore their priorities. The questions Kamm criticises as ‘imprecise’ (such as those included in resource guides for the Conversation Project and the Serious Illness Conversation Guide) were developed based on clinical experience to help people feel comfortable and reflect on these issues when facing serious illness.2

Second, Kamm ignores the role of emotions in thinking about serious illness and the medical choices patients and families much make. Conversations regarding disease progression, goals of care, and death and dying elicit strong emotions. Much has been written about the role of emotion in influencing decision-making.3 Attending to emotions requires flexibility and empathy, not precise, neutral questions about values. And, there is a deeper problem. Recent work in decision-making has shown that emotions both reflect and influence patient values.4 Conversations with patients thus need to illuminate patients' emotional statements in an effort to elucidate the values that underlie them. This requires an inductive, interpretative conversational style.

Third, more precise questions impose a pre-existing medical framework on values elicitation. For example, asking about time frames for valued outcomes (as Kamm suggests) presumes that time is an important factor and makes assumptions about how that time should be measured. This specificity may not capture what is most important to the person answering the question. Suppose a woman cares most about being alive for her daughter's wedding. If she is asked about valuing time in units that do not include mention of her daughter's wedding, there is a similar risk of failing to elicit her true values.

Kamm also highlights examples of ‘nudging’ in values elicitation questions, giving examples of non-neutral language such as ‘What are you afraid of about your illness?’ and ‘What kinds of medical care do you not want?’ that make presumptions about people (having fears and not wanting certain kinds of medical care, in these examples) and, Kamm argues, may bias their responses. Kamm recommends neutral questions such as ‘Are you afraid of anything about your illness?’ and ‘Are there kinds of care that you do not want?’ to ensure accuracy in EOLC decisions. This view raises additional concerns.

First, it seems to presume that it is possible for clinicians to be completely neutral in discussions about EOL preferences. We would argue that this is neither possible (because clinicians are humans) nor desirable (because it is not what many patients or family members want).5 Furthermore, every choice of words involves framing (and therefore nudging) of some kind.6 Recognising this, thoughtfully constructed questions to elicit peoples' values cannot hope to be truly neutral. Finally, existing biases in EOL decision-making have been well described.7 For example, optimism bias influences people to express hopes, not fears; default options to continue receiving life-sustaining care (with active choice required to receive comfort-oriented care) influence people to opt for medical treatments, not to forego them. Given the inevitability of bias (and the known direction of many biases with relevance to EOL discussion), nudging, as in the questions Kamm references, may be helpful in counteracting such biases and allowing people to express their true wishes.6

An alternative approach

Our initial approach to discussions about EOL treatment preferences is an effort to elicit the patient's story using simple, open-ended questions. Who are they as a person? What do they hold dear? What gives their life meaning? What do they most want to avoid? This approach is grounded in recognition that every patient's story is unique; that empathic listening strengthens therapeutic relationships; and that values cannot be easily quantified or categorised a priori. The desired outcome of this approach is not an accurate accounting of patients' preferences, but rather a rich and nuanced understanding of patients' values and goals, as well as the relationships and experiences that shape them. Such an understanding promotes patient autonomy by prioritising the patient's voice, counteracting known biases (that may interfere with true expression of a patient's wishes) and illuminating the factors that must inform subsequent treatment choices.

In our experience, these factors provide insights rather than answers. That is, patients' values and preferences do not map perfectly onto a single treatment option. Translating patients' wishes into a treatment plan that prioritises autonomy involves a subsequent shared decision-making process.8 We ask people what is most important to them not to get at a ‘true’ belief, but to help understand them as people. The goal is an authentic decision,9 meaning a decision that is most compatible with a patient's goals. This approach strives to achieve patient-centredness as defined by the Institute of Medicine, that is, ‘providing care that is respectful of individual patient preferences, needs, and values, and ensuring that patient values guide all clinical decisions’.10 The precise, neutral approach described by Kamm seems unlikely to achieve this.



  • Competing interests None declared.

  • Provenance and peer review Commissioned; internally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles

Other content recommended for you