This Symposium has focused on how physicians and patients interpret treatability statements (e.g., “This is a treatable condition…,” “There are things we can do…,” “We have treatments for this…”). Physicians turn to treatability statements to introduce a potential intervention for further deliberation. Treatability statements seem necessary because physicians are ethically obligated to discuss available treatments that a reasonable person might desire, according to contemporary accounts of informed consent and shared decision making.Footnote 1 While it is true that physicians might fulfill their ethical responsibilities without using treatability statements,Footnote 2 anecdotal clinical experience tells us that physicians tend to rely on treatability statements. As Kaufman, a medical anthropologist with decades of experience observing physician-patient interactions notes in her commentary, treatability statements remain “default medical practice.”Footnote 3
Several commentaries on our article illuminate the social context of treatability statements, clarifying why treatability statements are worth our attention. First, treatability statements are spoken in high-stakes situations, such as advanced cancer or critical illness.Footnote 4 Second, in the fragmented contemporary healthcare system, multiple treatment teams care for and communicate with the same patient, especially in the context of serious and critical illness.Footnote 5,Footnote 6 Practically, this multiplies the odds that a treatability statement will be used, as each clinical interaction with a physician presents a further opportunity to deliberate about potential interventions. Lastly, the modern biomedical industrial complex creates incentives that promote the creation of more and more medical interventions; therefore, as medical industry progresses, we can expect that more treatment options will be available that require deliberation, introducing more opportunities to use treatability statements.Footnote 7
More to the point, all of the commentators discuss the significant miscommunication that can arise from treatability statements. In our empirical work, we have summarized these miscommunications by noting that treatability statements falsely convey good news for prognosis and quality of life, as well as encouragement to hope and pursue further treatment.Footnote 8 This miscommunication is especially challenging to address because it stems from ostensibly everyday language—“treatable,” “treatment,” or “things we can do,” appear to be ordinary words, and are used in ordinary contexts (not just medical). As Weinfurt points out, “one would assume” that these words “are interpreted uniformly.”Footnote 9 This is the basis for the common teaching that physicians should use everyday language, and avoid clinical jargon. But as our empirical results show, and our commentators emphasize, everyday language can lead to profound miscommunication. This greatly magnifies the challenge that physicians face when discussing potential medical interventions with patients.
There are several important themes that emerge from the commentaries that potentially extend and expand our argument. First, Xafis and Wilkinson, Kaufman and Lynn all raise the possibility that physicians may sometimes be using “treatable” language to intentionally mislead patients. Lynn suggests that the “obfuscatory phrasing” of treatability statements “serves to buffer what is usually a very painful narrative to all concerned.”Footnote 10 In other words, physicians use treatability statements to avoid suffering—to avoid facing terminal illness and all of its limitations, pains, and losses. In the midst of these grim clinical scenarios, treatability statements provide a solitary positive point of focus: that treatments are still available. Xafis and Wilkinson develop this idea further, coining the term ‘intentional ambiguity.’Footnote 11 They note that, while treatability statements allow physicians to be technically truthful, they simultaneously allow physicians to avoid the reality that further curative treatment is unavailable, and to sidestep difficult emotions and challenging family dynamics. Kaufman notes the discomfort and lack of training that physicians typically receive in discussing death and dying with patients, and that retreating into the more comfortable language of treatability statements is a good way of avoiding these difficult conversations.Footnote 12
Of note, our empirical data supports these claims. Physicians showed a wide range of responses regarding potential misunderstandings by patients. Some seem to have assumed patients’ shared understanding with them, while others assumed that patients would read too much into the language used. Others highlighted the “intentional vagueness” of the language, much in the way suggested by Xafis and Wilkinson. Interestingly, at least some patient misconceptions revealed in our data (e.g., that physicians say ‘treatable’ to encourage further treatment, or that physicians say ‘treatable’ to convey hope) were not anticipated by any physician participants we interviewed.
Second, our analysis can be expanded or enhanced by other related and complementary tools for the study of communication. Weinfurt discusses how speech act theory (particularly the work of Grice’s mentor, JL Austin) can used to account for some of the possible misinterpretations that take place.Footnote 13,Footnote 14 For example, when an older, frail patient tells his nurse that he’s lived long enough and it’s time to stop, he could be making an informed refusal of life sustaining measures, or he could be complaining. In Austin’s terminology, these are two distinct illocutionary speech acts. But as Weinfurt points out, our article is primarily concerned with the perlocutionary effects of treatability statements—that is, the actual impact of the speech act on the behavior, belief, and emotional affect of the conversational participants.
Brindley’s commentary highlights some of the myriad factors beyond the literal meaning of words that impact the perlocutionary effects and inferences made about the intention of the speaker.Footnote 15 Tone, body language and other nonspoken aspects of communication encounters can play a crucial role, along with contextual features of the communicative act, including “genre.” As Brindley points out, when communication occurs between individuals with a great deal of shared knowledge, background, and experience, all of these features of language have the potential to make it possible to communicate more efficiently. But, as our research shows, when that shared background is missing, when one party to a conversation thinks they are in one “genre” (one type of speech act) and the other party thinks they are in another, communication has the potential to go awry.
Sources of Miscommunication
The commentary that most significantly challenges our interpretation of our data is the commentary by Xafis and Wilkinson.Footnote 16 Drawing on dictionary definitions of “treat” from six English language dictionaries, they show that British English tends to define “treat” as including reference to cure or restoration. In contrast, American English has a narrower definition of “treat,” focused on the delivery of medical care “with no reference to outcome or intention.” According to their proposed account, patients who falsely hear “good news” from treatability statements do so because they are operating on a definition of “treat” that more closely resembles the British English definition, and therefore semantically (literally) conveys ideas of cure or restoration. Xafis and Wilkinson point to the concept of “semantic shift” to explain the shift in how words are defined over time and in different communities. It is likely true that differing definitions of “treat” lead to some of the miscommunication we have observed in our empirical research.
However, many of the inferences physicians and patients make about treatability statements are not contained in any dictionary definitions of the word “treat.” For example, there is no dictionary which defines “treat” as indicating that treatment is a good idea, or as signaling that the physician is encouraging further treatment. Nevertheless, several of our patient participants discussed ‘treatable’ in this manner. For example:
‘Treatable’ would mean that, yes, it’s a very positive suggestion I think. Let’s do something and let’s see if it helps and just, you know, look at the side effects and continue with doing something. - Patient
Similarly, there is no dictionary that defines ‘treat’ as meaning or implying incurability. On this basis, Xafis and Wilkinson claim that there is “no linguistic evidence” for this possible meaning of ‘treatable.’ Nevertheless, several extant pieces of literature use the term in this wayFootnote 17 and our extensive interviews show that some physicians interpret treatability statements in this way:
When I imagine a physician using the term ‘treatable’, it means he’s trying not to use the term ‘incurable’. So I would think that the physician’s probably saying well, we can help manage your disease, we have drugs, options, procedures, that can make your disease better and in the lethal disease world that I deal in, ‘treatable’ means that it’s usually not curable. - Physician
The inference that ‘treatable’ implies ‘incurable’ is not drawn from the dictionary (semantics), but rather from the relationship between the words and their conversational context (pragmatics). Physicians who inferred that ‘treatable’ implies ‘incurable’ made these inferences only for diseases that slowly progress over time and are ultimately fatal (e.g., cancer, progressive neurologic disease). In these contexts, ‘treatable’ is intended to convey the sense that treatment can still be provided, even though the disease is incurable.
All of these inferences—that ‘treatable’ signals good news or encouragement of various sorts, or that ‘treatable’ signals that a disease is incurable—are examples of pragmatic inferences. In contrast to semantic meanings (e.g., dictionary definitions), these pragmatic inferences are generated by the listener, based on the relationship between words and their conversational context. As we have discussed elsewhere, these inferentially-enriched meanings often go far beyond the literal, semantic meanings.Footnote 18
It is important to emphasize that, ultimately, Xafis and Wilkinson’s proposed semantic mechanism and our proposed pragmatic mechanism are not mutually exclusive. Both likely contribute to the miscommunication generated by treatability statements, and both may even occur simultaneously in the same communicative episode. Further, semantic shift may be attributable to changing patterns of pragmatic inference across a population. As the same pragmatic inference is made repetitively over time (e.g., when American physicians persistently use ‘treat’ to refer to incurable diseases, as above), the dictionary definition shifts over time to capture these widespread patterns of use. In other words, what begins as pragmatic inferences may “harden” over time to become fixed semantic meanings, ultimately captured by dictionary definitions.
Avoiding Miscommunication
How can we avoid the miscommunications caused by treatability statements? Is there any way to discuss available treatment options and be properly understood? One strategy suggested by some of our commentators is simply to explain more of what is meant by treatability statements.Footnote 19,Footnote 20 Xafis and Wilkinson further suggest that explanatory adjectives (e.g., active treatment, curative treatment) will allow for semantic clarity that will solve the problem.Footnote 21 Our empirical data indicate that physicians themselves endorse these methods for avoiding miscommunication.
We are skeptical. Of note, there is no empirical evidence that supports the claim that explanatory clarification works. There is indirect evidence of its failure, in the literature that demonstrates widespread misunderstanding that palliative treatments are curative.Footnote 22 These misunderstandings occur in spite of what we and others have observed to be common physician practice: explaining more. As has been pointed out elsewhere, physicians are “explainaholics.”Footnote 23
We are also skeptical that physicians can simply avoid treatability statements, as Lynn suggests.Footnote 24 Kaufman’s excellent summary lays bare why this would fly in the face of modern medical history.Footnote 25 Nevertheless, our empirical data indicate that physicians endorse this strategy as well.
We suggest that a better strategy is for physicians to clarify what they do not mean—in other words, they should explicitly cancel, or contradict, the false inferences that patients derive from treatability statements. This refers to an important property of pragmatic inferences: that they are cancellable, or defeasible.Footnote 26 In Grice’s words: “conversational implicature can be canceled in a particular case…by the addition of a clause that states or implies that the speaker has opted out [of the inference].”Footnote 27 In the case of treatability statements, physicians should explain they do not intend to signal good news for prognosis, quality of life, or encouragement to hope or pursue further treatment. By so doing, they can cancel the miscommunication in a targeted manner.
In theory, we have no reason to question that treatability statements can be cancelled. In practice, however, we suspect that even this process may not be a reliable method for addressing the false pragmatic inferences generated by treatability statements. First, physicians are not perfect communicators. They may simply be unaware of the extensive analytic maneuvers their speech requires of their listeners. And even if they believe they ought to explain clearly what they do not mean, they may not always have the time, emotional bandwidth, or cognitive capacity to properly identify and cancel all implicature misfires.
Second, patients are not perfect listeners, especially in the context of serious illness. It is broadly accepted that in these circumstances emotion floods cognition and prevents the understanding of otherwise clearly stated, explicit information. In other words, patients may not even hear clearly stated semantic information, much less be able to correctly infer a physician’s intended meaning. In fact, in our own empirical work, the false inferences that patients drew from treatability statements led them to ignore previously stated negative information.Footnote 28
Lastly, and most importantly, cancelling a false inference requires knowing what that inference is. Without empirically exploring how patients tend to understand and interpret physician language, solutions that are grounded solely in physician intentions to state things more clearly, are likely to fail.
This is because physician intentionality may not matter much when it comes to addressing the miscommunications generated by everyday language such as treatability statements. One of the major upshots of Grice’s theory of conversational implicature is that the received meaning—that is, what patients actually understand from treatability statements—is constructed by the patient, not the physician. It is therefore the patient’s assumption about what the physician intends to communicate that determines meaning, not the physician’s intention. To our knowledge, virtually no attention has been paid to this issue in the large corpus of literature focused on physician-patient communication.
