The fragility of trust between patients and oncologists: A multiple case study

Objectives. This study aimed to explore in a naturalistic, real-life setting the dynamics of trust in oncological consultations. Methods. Cases to study were purposively selected from a data set of audio-recorded and transcribed consultations between oncology physicians and patients with advanced cancer, and analyzed qualitatively. The analytical approach was deductive, relying on a thematic framework of dimensions of trust


Introduction
Trust can be defined in the medical setting as "the optimistic acceptance of a vulnerable situation in which the truster believes the trustee will care for the truster's interests" (Hall et al., 2001). The essential role of trust has been recognized for a long time, but systematic empirical research has only emerged in the 2000s (Hall et al., 2002a).
Different scales to assess trust have been developed, such as the most widely used Trust in Physician Scale (Anderson and Dedrick, 1990), Wake Forest Physician Trust Scale (Hall et al., 2002b), and Trust in Oncologist Scale (Hillen et al., 2013). They focus on different dimensions of trust. Among them are (i) fidelity, defined as the will to do whatever is possible in the interest of the patient, (ii) competence, relating to the medical and interpersonal skills of the physician, (iii) honesty, which requires truthfulness, (iv) confidentiality, which assures to handle sensitive information carefully, (v) caring, which implies to be attentive to the patient's needs, and (vi) global trust, which unifies all these dimensions (Hall et al., 2001(Hall et al., , 2002aHillen et al., 2012bHillen et al., , 2013. To our knowledge, no previous study has assessed the dynamics of trust (formation, erosion, deconstruction, breakdown, etc.) throughout real medical consultations. We found one study examining the "interactional accomplishment" of trust, but it focused on a single, and very specific consultation (second opinion in the surgical setting) and evaluated trust with regard to a professional discipline and not to the physician (O'Grady et al., 2014). The aim of our study was to explore in a naturalistic, real-life setting the dynamics of trust by taking into account the whole oncological consultation and by considering the context in which trust develops. We privileged an in-depth exploration of a restricted number of consultations (case study approach).

Methods
The study was designed as a collective or multiple case study with a qualitative method of data analysis. Data for the study were from one source: audiotaped consultations between oncology physicians and advanced cancer patients.

Material
The material is part of a data set consisting of 134 consultations between 24 oncology physicians and 134 patients with advanced cancer. The consultations were audio-recorded and transcribed verbatim in the context of a naturalistic multi-center observational study (De Vries et al., 2017); this study received approval by the ethics committee of the participating Swiss hospitals, and all patients signed an informed consent form. Patients were informed that they have advanced cancer and that they receive treatment without curative intent. The objective of the consultations was to discuss the results of investigations (e.g., computed tomography scans, histopathological examinations, or tumor marker levels), documenting the evolution of the disease.

Selection of the cases
Four consultations were purposively selected, based on the identification of trust-related issues (manifest or underlying). In other words, we selected the consultations based on their own merits, and their genuine interest for us, in accordance with sampling considerations in the case study approach (Crowe et al., 2011). The cases show situations where trust was challenged, already eroded, or could be undermined because it was built on a fragile base. Trust is thus present in the four consultations. It may be apparent, but based on a collusion of silence (case 1), it may depend on the treatment strategy and hopes placed in it (case 2), it may be difficult to build due to a mutual distrust between the oncologist and the patient (case 3), and it can be in a latent way the central issue at stake in a consultation (case 4). In other words, the cases illustrate the fragile balance between mutual trust and distrust, and the analysis reveal factors that play a role in the dynamics of trust.
Communication between oncologists and patients in the four cases can sometimes appear inadequate or in need of improvement. In this regard, it is important to keep in mind that these are real-life consultations and that they also reflect the reality of clinical practice. We provide nevertheless some guidance on how communication and interpersonal trust could have been improved following each case. We refer for these guiding comments to the position paper based on the third consensus meeting among European experts on communication training of oncology clinicians (Stiefel et al., 2018), and to our clinical and academic experience in teaching clinical communication (Stiefel, 2006).

Data analysis
Based on the scales most frequently used to systematically analyze trust in the medical setting, especially the Trust in Physician Scale (Anderson and Dedrick, 1990), the Wake Forest Physician Trust Scale (Hall et al., 2002b), and the Trust in Oncologist Scale (Hillen et al., 2013), and the resulting dimensions mentioned above (e.g., fidelity, competence, honesty, see the Introduction), we developed a sensitive framework upon which analyses were based. The analytical approach was deductive, relying on the thematic framework of dimensions of trust, and inductive, not restricted by the framework. Thematic analysis (Braun and Clarke, 2006) was used to identify factors playing a role in the dynamics of trust (our themes) and affecting dimensions of trust recognized in the literature (the framework). The analysis consisted of iterative listening to gain a comprehensive view on the cases, generating themes, and defining and describing them in detail. The analysis was conducted by a multidisciplinary team consisting of physicians (TF and FS), and a social science researcher embedded for years in the medical setting (CB). TF coded the four transcripts. The two other investigators independently reviewed the coding, and differences were reconciled by discussion. Team discussions were held for the definition and description of the themes, and to obtain in-depth understanding of the dynamics of trust in the cases.
The case study as a research approach "allows in-depth, multifaceted explorations of complex issues in their real-life settings" (Crowe et al., 2011). In this study, it enabled a contextual exploration of the phenomenon of trust and comparison between situations to generate a broader appreciation.
In the results section, the four consultations are shortened, while keeping the flow of interaction between the patient and the oncologist. This allows to grasp the dynamics of trust. The speech turns to which we refer in the following are numbered in square brackets, which relate to Tables 1-4 (transcripts).

Results
We detail in the four cases the dynamics of trust and distrust between patients, their relatives, and the oncologists. Dimensions from the thematic framework (based on the scales measuring trust) are written in italics.

I trust you
This consultation is characterized by interpersonal trust between the patient and his oncologist (see Table 1). While the literature stresses the importance of the truster's relational vulnerability (Hall et al., 2002b;Hillen et al., 2012a;Gabay, 2015), the patient seems here to obtain all he demands, and one might question if he is vulnerable. However, by entrusting his health to the physician, he accepts a certain dependency and takes a risk, which can be conceived as vulnerability. Many dimensions of trust come repeatedly into play, such as the oncologist's competence (Hillen et al., 2013) [39-41, 186-8, 206, 210], trust related to confidentiality (Hall et al., 2001) [190-9], and global or holistic trust (Hall et al., 2001).
However, the "capital of the past" seems also to play an important role. It nurtures trust by means of past positive effects: the physician did what the patient expected, and it worked [39-41, 186-8, 210, 236]. In this regard, the literature distinguishes between trust (oriented toward the future) and satisfaction (oriented toward the past) (Thom et al., 2004;Hillen et al., 2012a). Yet, Gabay (2015) observed that satisfaction with outcome nurtures trust, and Hall et al. (2001) consider that trusting attitudes direct as much to motivations and intentions as to results. In other words, while trust is an interpersonal phenomenon, factors independent from the physician and the relationship with the patient also play an important role in its development.
In addition, there seems to reign an atmosphere of cronyism in this encounter, which might indicate a collusion (Stiefel et al., 2017). Cronyism is manifested by the demonstrative consideration and politeness 395] or the negotiation concerning the timing of chemotherapy . Collusion appears when the patient relates that he hides the disease from his wife , an attitude, which remains unquestioned by the physician, who seems to be more preoccupied about maintaining the therapeutic alliance, than exploring the patient's stance. Collusion, resulting in a conspiracy of silence between the physician and the patient and between the patient and his wife, can be considered as an indicator of distrust. The former indicates that the physician has not enough trust in the solidity of the relationship with his patient and the later that the patient has not enough trust in the relationship with his wife. The view that trust and distrust are mutually exclusive is not tenable. Trust is issue-related. Indeed, only a trusting relationship allows to share problems (Hupcey et al., 2001), and solidifies the working alliance (Fuertes et al., 2017).
Since trust seems here to be based on satisfaction with past results and a collusive relationship, one might question the solidity of this trust. For instance, the issue of the patient's loss of hair [403-13], a danger for the "disease secret", seems to trouble the harmony. This issue pops up at the end of the consultation and provokes a certain embarrassment, as indicated by the laughter of the oncologist [410], and the fact that the topic is quickly put aside. The relationship can be characterized as pseudo mutual because it is challenged by a disturbing topic. What would happen if the treatment stopped working, if the disease progressed rapidly or the patient had to admit to his wife that he has cancer? Therefore, underlying factors of trust formation have to be taken into account when assessing this phenomenon in the medical setting.

Comment on case 1
First, it is always tempting to focus on the success of past treatments, and by explicitly recalling it to increase the past capital of trust. However, one also has to remind that treatments may not always work. A more nuanced attitude may be the best prevention against an erosion of trust in moments when anticancer treatments are no longer beneficial. Such an attitude might be more difficult to adopt, but in the long run trust-building. Second, one can consider that it is the patients' right to not inform their significant others about their medical condition. However, having acknowledged that, the physician could also have investigated what makes it so difficult for the patient to openly discuss with his wife. Such a stance would not have undermined trust related to confidentiality, but enhanced trust related to fidelity (acting in the best interest of the patient) as well as trust related to honesty (by expressing a desire for transparency).

Did Prof. M* say something?
Trust is here directed from the patient and his wife toward two individuals: the oncologist (a chief resident) and his supervisor (a professor in oncology) (see Table 2). The patient and his wife seem to trust the professor, but to a lesser degree the chief resident. While they do not distrust him, since this would imply the expectation of harmful behaviors (Hillen et al., 2011), the relationship with the chief resident is characterized by low trust (Hall et al., 2001;Hillen et al., 2011). The patient and his wife consider the chief resident not to be completely capable to make the (right) decisions. The patient presents himself more nuanced than his wife, as illustrated by choice of words, such as "[the professor] might have changed his mind" [167] or "the results are good, that's already something!" [346]. The prudent stance may indicate that he is preoccupied to maintain a good relationship with the chief resident. This indicates that verbal manifestations of trust, which might result from insecure attachment or dependency, cannot always be taken at face value, and have to be contextualized. The core of this consultation is the current treatment, chemotherapy, conducted in Switzerland, and the possibility of a targeted treatment qualified by the professor as the "miracle molecule", available in the USA. Despite the fact that current chemotherapy is beneficial [21,104], the patient's wife rapidly shows interest for the targeted treatment [99], which she associates with enhanced life expectancy [255] and superior efficacy [201,251]. The chief resident repeatedly confirms her view: "Well, it's amazing, it [the targeted treatment] has nothing to do with chemotherapy" [202,256]. This leads to an incomprehension: there seems to exist a treatment, more efficient than the current one, but the patient is denied to obtain it. The dimension of fidelity, "pursuing a patient's best interest" (Hall et al., 2001), appears to be questioned by the wife. In addition, the oncologist's communication, another dimension of trust (Hall et al., 2001), also seems to fail, since he does not explain the rationale of his stance.
Another determinant of trust, articulated to the two previous dimensions, favors low trust: the lack of coordination between the physicians. The chief resident endorses the "miracle" treatment but prefers not to provide it for the moment; and the professor seems to encourage the patient to get the treatment in the USA. The patient and his wife deplore this lack of coordination .
The competence of the chief resident is constantly compared with the competence of the professor. While the chief resident's competence is never questioned explicitly, it is tacitly challenged by repeated references to the professor [99, 160-9, 177, 182-4, 339], and even by the chief resident himself [21,117,126,148,152]. One could thus consider that the chief resident also lacks trust, trust in himself. This last issue illustrates that trust can also be undermined by the trustee himself. Lastly, maybe even more relevant than the professor's competence is the trust in the "miracle drug"; the professor is maybe just entrusted to provide access to it.
In conclusion, developing trust toward the chief resident would imply "a willing dependency on his actions" (Hupcey et al., 2001), which does not occur. The consultation, therefore, ends as it started, with a negotiation concerning the treatment and a reference to the professor: Comment on case 2 From a communication perspective, it might be worthy to explicitly address and meta-communicate on the relational dimension, which operates within this consultation. The chief resident could for example have stated: "I observe that it is difficult for us to reach a trustful relationship, I have the impression that you prefer to rely on the professor's advice, and I believe that he and I should first coordinate our propositions and then speak with one voice." This consultation shows that to maintain trust when different physicians are involved, it is important to avoid any splitting and to coordinate both care and communication.
Are there any other therapies?
In this consultation, a lack of interpersonal trust of the physician in the patient is observed, but also of the patient in his physician and other physicians, the hospitals, conventional medicine, and the healthcare system (see Table 3). This "global" distrustcontrary to low trust in case 2indicates that the patient considers the physician to act against his best interests (Rose et al., 2004;Hillen et al., 2011). While the literature focuses on patients' trust in physicians, trustor as in this consultation, distrustcan also be directed from the clinician toward his patient. As stated by Thom et al. (2011): "[ … ] patient and physician trust are closely linked to expectations of behavior with respect to complementary roles." In this consultation, the patient fails to be entrusted by the physician, who considers that he does not accept the patient role and does not provide the necessary information for medical care [13,19,47,53,109]. The patient is not considered trustworthy, as these remarks from the physician seem to indicate: [ While it is not completely clear what the physician wishes to express by this last sentence, it appears that she compares the patient to other patients, who seem to be more entrustable.
On the other side, the patient seems to be very wary of his medical care; wariness being a characteristic of distrust (Hall et al., 2001 To trust implies to willingly transfer discretionary power (Grimen, 2009) to a person or an institution. This is not what the patient shows, and the discussion with the oncologist, but also with the group of physicians, ends up in a mutual struggle for power. In this context, relational elements appear in the conversation, for example, when the patient passes from the polite and adequate "vous" (in French) to the familiar and inadequate "tu" [78]. This might indicate that he fights a certain asymmetry in the relationship, which he attempts to flatten.
The patient's personality seems to play an important role in the interaction. He obliges the oncologist to contain his distrusting feelings [20, 74-8, 94, 222, 242-4, 334-6, 378-86]. This might be conceived from a psychological perspective as projective  [The oncologist explains that the only curative treatment is surgery. The patient is not satisfied, he thinks that it is possible to do more than just "cutting one piece" (2x identification (Gabbard, 2001). The patient provokes feelings of helplessness in the physicianbeing accused of not caring for the patient -, which the patient might himself experience (helplessness with regard to how to manage the disease). The underlying and unconscious motivation of the patient might be that he attemptsas with the resort to the transgressive "tu"to establish a more symmetrical relationship (both feeling helpless), at the cost of trust. [77] O: Huh?
[78] P: When you [P uses "tu"] said "there's something wrong", but you didn't say what! However, when the other physicians enter the room, they establish a more asymmetric relationship, particularly Dr.*: "We have to accept the PET as it is [ … ]" [461]. This sentence indicates that the physician asserts his power by relying on scientific knowledge and facts (the PET scanner), sending at the same time an implicit message to the patient: your distrust is related to your inability to accept your situation. Finally, the physicians as a group decrease the patient's "combativeness," who only issues very few comments or objections [e.g., 769].

Comment on case 3
One often observes that physicians restrict the discussion to medical problems, even when other issues take over, such as emotions, existential difficulties, or relational elements, as in this case. Instead of justifying herself, the physician might rather address the difficulty of the patient to develop some trust toward medicine and healthcare professionals, and invite him to look for ways to improve, together, the situation.
The issue of trust is never explicitly addressed, only remarks of the patient implicitly indicate her difficulties to trust her oncologist. For example, the email to which the oncologist is said not having replied, the sick-leave certificate, which the patient had asked for but never received, or the MRI, which had not been done in the right way. Just as satisfaction seemed to nurture trust in case 1, dissatisfaction seems here to nurture distrust.
To trust implies to be vulnerable and to accept one's own vulnerability, without having a guarantee of a benefit for the truster (Goudge and Gilson, 2005;Bachinger et al., 2009;Grimen, 2009;Hillen et al., 2011). In this case, the patient questions whether the physician took her vulnerability (life-threatening situation) seriously and thus does not entrust her [e.g., 234-46, 297-98, 337-81, 427-34, 471-72]. To trust requires coping with uncertainties, but here uncertainties, which naturally exist in oncology, motivate distrust. [The oncologist explains that the disease was stable on the MRI from February] 234. P: Yes, but it's also because you have booked an exam with the 1.5 tesla magnetic field and it's true that, even the radiologist's report says that, it's written at the end, "the cavernous angioma has a reduced aspect, reduced dimensions, because the magnetic field is less powerful" so with a 3 tesla we can infer that they were able to assess more precisely In addition, the patient shows low trust in the hospital: she criticizes specific aspects of its functioning, such as the communication between healthcare professionals, and a lack of continuity of care [297,313]. Her observations may be facts, but they do not automatically have to create low trust. The status of the oncologist is also questionedis she really the right hand of the chief of service? [line 297], is she a qualified oncologist or "only" a general practitioner? [line 427] -, indicating low trust.
Here, the capital of the past is a capital of distrust, which has accumulated and can thus be considered as long-term distrust (Hillen et al., 2012a).

Comment on case 4
The oncologist, instead of dealing with distrust, attempts to affirm her competence and defends herself [46-87, 297-98, 313-25, 337-404, 427-34], which leads to an increasing confrontation of protagonists (Watzlawick et al., 1967), nurturing distrust. Containing negative emotions, avoiding the temptation to enter into symmetry and, again, addressing the relational components might be a more constructive way to attempt to restore trust instead of defending oneself or justifying one's actions.

Discussion
While we observed different dimensions of trust previously described in the literature, we have also identified factors, which appear to play a role in the dynamics of trust and affect trust or nurture distrust. These factors are the number of treating physicians and how they communicate, continuity of care and the capital of trust, the hierarchical position of the physician and the physician's self-trust, and the patient's personality.
In the cases, factors related to the setting seem to be relevant for trust in the patient-physician relationship. Among them are the number of treating physicians. Since oncology is interdisciplinary, care relies on different treatment modalities, and often needs specialists' advice; patients may thus encounter multiple physicians. However, with an increased number of physicians, a certain diffusion of the sense of responsibility for patientsdecreasing their trustmay develop. This phenomenon, already identified by Balint as a "collusion of anonymity" (Balint, 2000), is even more potent when communication between physicians is not coordinated, or when physicians provide different or even contradictory information.
Second, continuity of care also plays an important role with regard to trust. This is especially difficult to maintain in tertiary care centers, in which residents undergo rotations. The dangers of a lack of continuity of care has been observed by Hillen  reported to be strong, there is concern that this solid trust is eroding, due to changes in health-care organizations that might pave the way to less continuity of care and less personal attention to the patient" (Hillen et al., 2011). The possibility of developing a capital of trust may therefore suffer. On the other side, longstanding relationships with physicians may also harbor a risk that physicians lack a critical stance toward the patient, as revealed by the above identified unhealthy collusion (case 1), which may mimic trust.
Inter-and intrapersonal factors were also identified as affecting trust. Indeed, hospitals are hierarchically organized, and patients know that. Therefore, clinicians who occupy lower hierarchical rankswith regard to clinical responsibility or academic achievementsmore often face doubts with regard to core competencies. Trust might thus be hampered by reasons independent of the physician or the relationship he has established with the patient. To know this is important to question the origins of distrust and to avoid to attribute it immediately to oneself (and to not feel self-trusting). The same holds true with regard to the patient's personality. Distrust might be a general characteristic of the patient, such as in patients with paranoid personality traits, and thus develops independent of the clinician's behavior.
Besides these factors, we observed indicators of eroding trust in the low trust and distrust consultations. What is striking is that patients and physicians did not explicitly address trust or distrust. While it is difficult for patients to address this issue, since they are in a situation of dependency, it is also difficult for physicians, given their prosocial motivations. However, addressing the issue can reveal origins of eroding trust, which might be related to the setting (e.g., continuity of care) or to the patient's anxiety, attachment difficulties and uneasiness (e.g., manifested by introduction of third party agents). Distrust merits to be addressed, since it might be attenuated when expressed; confusion, erroneous interpretations, and projections may diminish after clarification and its causes might at times be eliminated.
Trust is at the core of the medical encounter. This case study reveals that trust is a dynamic phenomenon, affected by contextual, interpersonal, and intrapersonal factors, which all can enhance or erode trust. The Gestalt of trust can only be approached and grasped by examining the singular situation. In this respect, our study is original because trust issues were examined throughout the consultations/cases, demonstrating how trust is both a dynamic and fragile phenomenon, and that trust and distrust can co-exist closely in the same consultation. Our findings also show that trust in oneself, in others and in the world is not only necessary for patients to find their way to care, it is also a constitutive element of the patient-physician relationship and a challenging and invested issue for the physician.