You are here

  • Home
  • Online First
  • To what extent should doctors communicate diagnostic uncertainty with their patients? An empirical ethics vignette study

Article Text

Article menu

Download PDFPDF

Original research
To what extent should doctors communicate diagnostic uncertainty with their patients? An empirical ethics vignette study
  1. Caitríona Cox1,
  2. Thea Hatfield1,
  3. Matthew Parry2,
  4. Zoë Fritz1
  1. 1The Healthcare Improvement Studies Institute, The University of Cambridge, Cambridge, UK
  2. 2Department of Mathematics and Statistics, University of Otago, Dunedin, New Zealand
  1. Correspondence to Dr Caitríona Cox; clc70{at}cam.ac.uk

Abstract

Background/aims Although diagnostic uncertainty is common, patient-focused research examining its communication is lacking. We aimed to determine patient preferences for the communication of diagnostic uncertainty, and examine the effects of such communication on patients.

Methods We applied an empirical ethics approach, integrating the data collected with ethical analysis to form normative recommendations about diagnostic uncertainty communication. In this randomised crossover study, n=111 members of the public sequentially watched two video vignettes depicting either high or low communicated diagnostic uncertainty, in one of two clinical scenarios. After watching videos, participants completed online questionnaires. Primary outcome was preferred video (high vs low communicated uncertainty); secondary outcomes included satisfaction, trust, worry and understanding. Quantitative data were analysed using logistic regression and a linear mixed effects model; qualitative data were analysed thematically.

Results Quantitative analysis demonstrated that participants preferred greater diagnostic uncertainty communication, even though these vignettes were more worrying. Qualitative data revealed heterogeneous participant views justifying their communication preferences. These data raise issues relating to how doctors might balance harms versus benefits in diagnostic uncertainty communication and how doctors might communicate in the face of heterogeneous patient information preferences.

Conclusions We argue that doctors should err on the side of greater diagnostic uncertainty communication: to not do so (eg, based on benign paternalistic ideas about avoiding patient worry) or to do so variably (eg, based on unevidenced assumptions about patient information preferences) risks depriving patients of information they may value and may create or exacerbate inequalities.

  • Ethics
  • Decision Making
  • Ethics- Medical

Data availability statement

Data are available upon reasonable request. The participants of this study did not give written consent for their data to be shared publicly, only that the anonymised data collected may be used to support other researchers in the future. As such, the data that support the findings of this study are only available on request from the corresponding author, CC.

https://creativecommons.org/licenses/by/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution 4.0 Unported (CC BY 4.0) license, which permits others to copy, redistribute, remix, transform and build upon this work for any purpose, provided the original work is properly cited, a link to the licence is given, and indication of whether changes were made. See: https://creativecommons.org/licenses/by/4.0/.

https://doi.org/10.1136/jme-2024-109932

Statistics from Altmetric.com

    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.

    Introduction

    Ethical approaches to non-disclosure in the doctor-patient relationship

    Under what circumstances it is acceptable—ethically and legally—for doctors to not disclose information to their patients has been debated in various scenarios: regarding prognosis,1–3 risks of treatment4 and medical errors,5 to name but a few.

    Historically, doctors took a relatively paternalistic approach to doctor-patient communication: it was considered acceptable, for example, to withhold information such as a diagnosis of terminal cancer to avoid distress.6 Over the latter half of the twentieth century, there has been a gradual shift (particularly in Western medicine) towards a less paternalistic conceptualisation of the doctor-patient relationship,7 with emphasis on open communication to support shared decision-making.8 9 This shift has also been reflected in case law, which has scrutinised standards of disclosure: in the UK, Montgomery vs Lanarkshire10 examined the duty of the doctor to disclose information about material risks of a given treatment, and any reasonable alternative treatments.

    Despite these changes, there is still recognition that it is acceptable for doctors to withhold some information from patients. As Jones summarises, various justifications have been proposed:

    Patients will worry, and become too anxious about their treatment if they are given full information; they will not understand the information in any event, because it is too complex and technical…doctors do not have time to give full information to patients-there are too many patients and it would take too long…11

    Much of the ethical literature exploring medical non-disclosures employs a broad harm-benefit analysis: in determining the acceptability of a non-disclosure, there is consideration of its potential impact (both on the patient and healthcare system).4 12 13 Many analyses use Beauchamp and Childress’ ‘Four Pillars’14 as a scaffold, framing the fundamental ethical tension as arising from the need to balance competing ethical principles—particularly respect for autonomy (which would often favour greater disclosure of information) versus non-maleficence (which may favour non-disclosure of information which is distressing, overwhelming or unwanted).15 16

    Such ethical arguments often rely on claims about the effects of communication on patients. It is thus important to understand them; in the absence of such understanding, incorrect assumptions about what information patients want, or how they will respond to it, may result in flawed conclusions about the ‘right’ approach. An example of this can be seen in palliative care: although historically communication about poor prognosis was limited by the perception that patients did not want or would be harmed by such information, empirical research has repeatedly demonstrated that patients value honest communication of prognosis (even if the prognosis is poor).1 2 17 18

    Communicating diagnostic uncertainty

    One area of ongoing debate relates to communication about diagnosis.19 Diagnosis is not a single event, but a complex and collaborative process, which often involves considerable uncertainty.20 This uncertainty represents a communication challenge: to what extent, if at all, should doctors disclose diagnostic uncertainty to their patients?

    Both researchers20–22 and regulatory bodies23 have made recommendations that doctors should communicate diagnostic uncertainty: for example, UK General Medical Council guidance states that ‘[i]f you are uncertain about the diagnosis… you should explain this to the patient’.23 There is, however, evidence that diagnostic uncertainty is not always communicated. Our recent study, which used written vignettes to explore doctor communication, found significant variation in practice: doctors did not always explicitly disclose diagnostic uncertainty, even when they were not themselves certain about the diagnosis.24

    That vignette study,24 and two systematic reviews,25 26 explored reasons why doctors choose to (not) disclose diagnostic uncertainty to patients. As outlined in figure 1, we identified a range of considerations. Some of these (eg, the desire to avoid patient worry) have been used to justify both the disclosure and the non-disclosure of diagnostic uncertainty. Communication of uncertainty could plausibly increase patient worry (by acknowledging the possibility of unlikely but serious differential diagnoses) or decrease it (by explicitly addressing concerns patients might already have about such diagnoses). Such opposing ideas might explain some of the observed variation in communication: although doctors often have overlapping communication goals (such as reducing patient worry), different ideas about how to achieve these goals result in differing communication behaviours.

    Figure 1
    Figure 1

    Considerations identified in existing literature about whether doctors should communicate diagnostic uncertainty to their patients.

    A lack of patient-focused research in this area means that there is little empirical evidence to confirm or refute beliefs about the consequences of diagnostic uncertainty communication.25 26 While some evidence suggests that communicating uncertainty regarding treatment options may be associated with lower patient satisfaction,27 little research has directly examined the effects of communicating diagnostic uncertainty. One study in paediatrics found that explicit expression of diagnostic uncertainty was associated with lower perceived competence, less trust and confidence,28 while another study involving patients with endometriosis demonstrated a patient preference for clear communication of diagnostic uncertainty to facilitate informed decision-making.29

    Questions also remain regarding the role of communicating diagnostic uncertainty when ‘safety-netting’ (giving advice to patients to help them identify when to seek further medical help if their symptoms worsen/do not improve).30 Some have suggested that sharing uncertainty is an important component of effective safety-netting,31–34 and there is evidence that some doctors explicitly share diagnostic uncertainty to encourage appropriate attendance and protect against diagnostic errors.24 35 However, whether communicating diagnostic uncertainty actually makes safety-netting more effective in encouraging patients to reattend in the event of worsening/non-resolving symptoms remains unclear.

    What does the current study add?

    In summary, doctors have differing opinions about whether communicating diagnostic uncertainty to patients causes harm (eg, by causing worry or by overwhelming patients with complex details) or benefit (eg, by encouraging appropriate reattendance through effective safety-netting or by fostering a more trusting doctor-patient relationship). A dearth of relevant patient-focused research makes it difficult to establish which assumptions about the impact of diagnostic uncertainty communication are correct. As ethical analyses often rely on such assumptions, the lack of empirical evidence is problematic.

    This study begins to address this research gap. We aimed to (1) determine patient preferences for the communication of diagnostic uncertainty and (2) examine the effects of communicating diagnostic uncertainty on patients. By exploring the impact of communicating high or low diagnostic uncertainty on a range of outcomes, this study helps to address the normative question of how doctors should communicate diagnostic uncertainty to their patients.

    Methods

    Overarching empirical ethics approach

    Empirical ethics refers broadly to research which combines the use of empirical data with moral theory to reach normative conclusions.36–39 A wide range of methodologies exist within this growing field.36 These vary regarding their integration of the empirical and theoretical—different approaches use different ‘bridging methodologies’ to combine empirical and conceptual analysis to produce normative solutions to identified ethical problems.38

    The study presented here forms part of a wider programme of research, which uses an empirical ethics approach to explore issues relating to the formation, communication and recording of diagnoses. This strand of the project has examined the communication of diagnostic uncertainty, ultimately addressing the normative question of to what extent doctors should communicate diagnostic uncertainty to their patients.

    We applied Huxtable and Ives’ ‘Mapping-Framing-Shaping’ framework.38 Two systematic reviews helped to ‘map’ the area, identifying what was known and what remained unexplored regarding the communication of diagnostic uncertainty and identifying associated ethical issues.25 26 We then used vignette methodology to ‘frame’ relevant topics, collecting empirical data to explore: (1) how and why doctors currently communicate diagnostic uncertainty, (2) what the impact of such communication is on patients and (3) what patient preferences for such communication are. We finally aim to ‘shape’ the terrain by using wide reflective equilibrium40 to synthesise the empirical data with moral theories regarding information disclosure within the doctor-patient relationship, to suggest normative recommendations about how doctors should approach the communication of diagnostic uncertainty.

    In this paper, we present the empirical data from the patient-focused video vignette study. We discuss the results with reference to ethical theory and the existing literature, referring to work exploring therapeutic privilege and ethical aspects of non-disclosure within the doctor-patient relationship.

    Study design

    In observational studies of real consultations, specific communication behaviours cannot be isolated and manipulated: they can only explore correlations between communication behaviours and outcomes measures. In contrast, our use of vignette methodology in this study allowed controlled manipulation of the degree of diagnostic uncertainty communication (permitting consideration of causal relationships between uncertainty communication and relevant outcomes).

    This was a randomised crossover study. Participants sequentially watched two video vignettes depicting either high or low communicated diagnostic uncertainty. We used two separate clinical scenarios: participants were initially randomised to either group 1 (headache), or group 2 (change in bowel habit (CoBH)) and then were further randomised within group to the order of vignettes watched (figure 2).

    Figure 2
    Figure 2

    Study design.DU, diagnostic uncertainty; IUS, intolerance to uncertainty scale.

    We administered the vignettes using Thiscovery, an online platform developed by THIS institute. Participants engaged with the study alone using their own electronic devices, without researcher supervision. Videos could not be paused or rewatched.

    Development of vignettes

    The process of developing the vignettes, including pilot-testing, is reported elsewhere.41 In brief, we followed published guidance on the development of video vignette studies42 to produce two versions (high vs low communicated uncertainty) of two clinical scenarios. We drew on data from an earlier vignette study involving doctors to make the scripts as realistic as possible.24 The high communicated uncertainty vignettes included a range of verbally expressed implicit and explicit statements of uncertainty; the low communicated uncertainty vignettes did not include acknowledgement of diagnostic uncertainty. Other verbal and non-verbal elements of the consultation were kept constant between the vignettes.

    Outcome measures

    Participants completed questionnaires: (i) after watching each vignette individually; (ii) after watching both (asking which video was preferred, including an optional free text space to explain this preference); and (iii) to record demographic information, previous personal experience of the clinical scenario and a 12-item intolerance to uncertainty scale (IUS-12).43 We chose to ask participants to complete the IUS after watching the vignettes rather than before, to avoid priming them think about uncertainty.

    The primary outcome measure was preference regarding diagnostic uncertainty communication (ie, preference for video A or B, determined after watching both vignettes).

    Secondary outcome measures were identified via a combination of literature review, the results of a previous study conducted by the research team24 and patient public interest (PPI) group discussion. They included:

    1. Patient overall satisfaction with the consultation.

    2. Patient worry.

    3. Patient perception of doctor competence.

    4. Patient trust in the doctor.

    5. Patient understanding of information provided.

    6. Patient happiness with the amount of information provided.

    7. Patient desire for further investigations.

    8. Patient likelihood of seeking further medical attention if their symptoms persisted/worsened

    Secondary outcomes were assessed using single-item 0–10 numerical rating scales. We included verbal anchors at the extremes and midpoint to improve the reliability and validity.44 45 We used various sources to develop the items (including literature review and consultation with proposed respondents) to increase face validity.46

    The questionnaires can be found in online supplemental appendix A, alongside a referenced explanation for our decision to use single-item measures as opposed to multi-item scales.

    Supplemental material

    Participants, power and recruitment

    Participants were members of the UK public, aged 18 years and above. They all provided informed consent via an online form. Medical students and doctors were excluded. There were two separate recruitment approaches:

    1. General recruitment: advertisement via social media and cascading via individuals/networks.

    2. Partner organisations: four Healthwatch (an independent national champion for people who use health/social care services) organisations in different locations were contracted to each recruit ten participants from demographics that were proving harder to reach via social media recruitment (eg, non-university educated, those at extremes of age).

    In performing the power calculations, we assumed a significance level of 0.05, a minimum detectable difference in means of 1 and a within subject SD of 1. For 90% power, we would need a minimum of 48 subjects in total (ie, 12 in each group). We planned to recruit more than this, aiming for n=100 (ie, 25 subjects in each group), to capture a wider range of population characteristics.

    Data analysis

    We used a mixed methods approach: the final questionnaire combined quantitative and qualitative data (free text responses qualifying reasons for vignette preference). These data were integrated at the interpretation stage of the study and were given equal priority.47 48 Analysis of embedded qualitative responses was intended to augment and explain complex or contradictory quantitative responses.49 By combining statistics with thematic analysis, we hoped to avoid over-reliance on the former, better capturing experiences and subjective factors necessary to elucidate complex phenommena.48

    Quantitative analysis

    For the primary outcome, logistic regression was used to estimate the proportion preferring vignette A (the video communicating high uncertainty), allowing for a potential scenario effect (headache or CoBH) and a potential sequence effect (AB or BA). A goodness-of-fit test (deviance test) was used to assess model adequacy. Adjustment was also made for IUS score and subject history. In the analysis of the primary outcome, a significance level of 0.05 was chosen.

    Each secondary outcome was treated as a continuous response and a linear mixed effects model was used to estimate the (fixed) effect of vignette type, allowing for a possible ‘carry-over’ effect.1 The random effect allowed for between-subject variability but was constant for within-subject responses. The advantage of using a linear mixed effects model over a simpler model of within-subject differences is that adjustment can be made for IUS score and subject history of the condition. Graphical checks of the normality assumption for both random effects and residual errors were used to assess model adequacy. In the analysis of the secondary outcomes, a stringent significance level of 0.01 was chosen, to mitigate the effect of multiple testing leading to false discoveries (Type I errors).

    To determine whether individuals had been appropriately randomised into groups of scenario and sequence, Fisher’s exact test was used on the demographic data and a t-test was used on the IUS scores, with a significance level of 0.05.

    Microsoft Excel and R50 were used.

    Qualitative analysis

    Qualitative data were analysed using reflexive thematic analysis.51–53 This is a theoretically flexible analytical approach, in which codes represent researcher interpretations of patterns of meaning across the dataset.53 Our team consisted of clinical and non-clinical researchers; we used this analysis approach with the aim of both acknowledging and embracing the reflexive influence of varying researcher interpretations grounded in different backgrounds/perspectives. Using an iterative, open coding approach, we developed codes to capture the main themes interpreted from the data. TH and CC initially familiarised themselves with the data, before independently developing codes. The whole research team then met to review these codes and determine the ongoing coding approach. CC then coded the entirety of the data using this approach. Higher-level themes were identified and were compared with existing literature to highlight significant trends and gaps. NVivo 12 Pro software was used.

    Empirical results

    Data were collected in a continuous period from December 2022 to March 2023 (n=111 in total; table 1).

    View this table:
    Table 1

    Participant details

    Quantitative results

    The demographic data show the participants are representative of the UK population in terms of ethnicity. However, the study group is skewed in terms of age (67.5% are over the age of 50), sex (66.6% identified as female) and education (76.5% have a university or college degree, or higher). In terms of demographic characteristics and IUS score, participants were found to be appropriately randomised into groups scenario and sequence.

    The expressed preferences by scenario and sequence are shown in table 2. The estimated proportion (of the population) preferring videos with high communicated diagnostic uncertainty is 0.64 with a 95% CI (0.55, 0.73). There is no evidence for a difference in the order the videos were viewed, and, there is no evidence that IUS-12 score of previous history of the clinical scenario has an effect on preference. There is no evidence for a difference in preference based on education level.

    View this table:
    Table 2

    Number of participants preferring each vignette (presented according to sequence watched)

    Analysis of the secondary outcomes (table 3) indicates statistically significant results for levels of worry. The estimated mean change in worry from viewing videos with high communicated diagnostic uncertainty is an increase of 1.14 on the 10-point scale with a 95% CI (0.60, 1.67). There is no evidence for a difference between the vignettes (headache vs CoBH scenarios). The results are also robust if adjustment is made for IUS-12, personal history of the medical condition and education (university or college degree or higher; or not).

    View this table:
    Table 3

    Effect of high communicated diagnostic uncertainty (video A) on the secondary outcomes (using a significance level of 0.01 as the threshold)

    Of note, the level of understanding could not be analysed as normality assumptions were violated; this was most likely because the distribution of responses was heavily skewed: 76% of responses were 9 or 10. This indicates that all versions of the vignettes produced (both high and low communicated diagnostic uncertainty) were highly understandable for participants.

    Qualitative results

    After watching both high and low communicated diagnostic uncertainty vignettes, participants indicated their preference and were asked to provide reasons for this. These data revealed varied justifications. Common themes included: feelings about the detail provided, the perception of the communication as worrying versus reassuring, the impact of the communication on perceptions of the doctor and the impact of the communication on health behaviours.

    As there was little difference between the responses given by participants who watched the headache vignettes versus the CoBH vignettes, we do not differentiate the data and present the main themes below.

    Amount of information provided: desire for information versus overwhelming/unnecessary detail

    The most common reason participants gave for preferring the high communicated uncertainty vignettes was a desire to be better informed. Many responded in general terms:

    I like to know as much as possible when asking for advice about anything from medical matters to making decisions about plumbing. – P10

    Some elaborated that greater information was helpful in facilitating better understanding of the situation; others explained their preference with specific reference to the discussion about the differential diagnosis or the rationale behind investigations. Understanding the doctor’s thought processes—including the handling of uncertainty and the differential diagnosis—was helpful.

    More relevant information was given and I felt I had a clearer picture of the situation. – P1

    I like the thought process taken to come out with diagnosis. – P94

    This desire for more detail was not, however, universal: a minority felt that the amount of information provided in the high communicated uncertainty vignette was unnecessary.

    The [high communicated uncertainty video] included information I did not need, may not understand, created anxiety and expressed uncertainty about the quality of the investigation, diagnosis and treatment. – P30

    A few participants stated that they felt the level of detail provided should be patient-led—so if a patient wanted more information, they could ask for it.

    If I wanted more information about what they were specifically looking for, then I would ask. – P60

    Worry versus reassurance

    A prominent theme was that of worry versus reassurance: participants frequently referred to ideas about the communication causing or relieving worry (see table 4 for illustrative quotations). Importantly, responses around this theme were inconsistent. For some, communication of uncertainty, for example, through discussion about ‘what-ifs’ or unlikely but ‘scary’ diagnoses, was worrying.

    View this table:
    Table 4

    illustrative quotations regarding worry versus reassurance

    I feel the [high communicated uncertainty video] gives too much information and would be scary and cause me to overthink the issue. – P46

    A small number of participants explicitly stated that even though they found the information in the high communicated uncertainty video more worrying, they still preferred it. These participants acknowledged that receiving more information about diagnostic uncertainty might be anxiety-inducing, but felt that the benefits of a fuller explanation outweighed this worry.

    I would prefer to be aware of all of the information—what the doctor is sure of, and what can’t be ruled out…Although, of course, [this] made me feel more worried…I would always rather know the full truth. – P39

    [A]lthough knowing more about what the ‘worst case scenario’ might be (and that’s a bit scary and more worrying), I’d rather have the full picture. – P19

    In contrast, others described how the fuller explanation of uncertainties provided reassurance: communication of details about the differentials which were being considered helped to relieve worry, by making participants feel as though their concerns had been more thoroughly considered. Some noted that information can easily by looked up online, and they would rather be informed about uncertainties directly rather than coming across them themselves at home.

    I think that in the days of Google—many people could have an idea of what the symptoms could be. I would want it explicitly acknowledged that the Dr has considered…cancer for instance. It makes you feel like you are taken seriously. – P4

    In a similar vein, other participants stated that the omission of information in the low communicated uncertainty vignette was itself worrying. These participants explained that not being given all the information about possible diagnoses might cause anxiety.

    Impact of communication on perceptions of the doctor

    The impact (positive or negative) that the communication had on participant perceptions of the doctor was referenced. Acknowledging diagnostic uncertainty was sometimes perceived positively, taken as an indication that the doctor was honest and trustworthy:

    I preferred the fact that the doctor in [the high communicated uncertainty video] used words to reflect a (small) level of uncertainty like ‘likely’ or ‘unlikely’. That made me feel like he was being more honest. – P15

    By the doctor explaining the differential diagnoses…this will illicit trust from the patient. – P7

    A small number of participants described how the communication of diagnostic uncertainty had a negative impact on their perception of the doctor, as it gave an impression of lack of confidence.

    I didn’t like that there was the offer to do the Lumbar test on second return, as it added an element of lack of confidence on the Drs part. – P58

    Impact of communication on health behaviours

    Some participants mentioned the impact that the communication of uncertainty might have on future health behaviours. Some participants who preferred the high communicated uncertainty vignette suggested that it might make patients more likely to follow safety-netting advice. It is, however, important to note that some participants who preferred the low communicated uncertainty vignette also reported that they felt able to return after watching this video, suggesting that patients may be empowered to follow safety-netting advice based on factors other than the communication of diagnostic uncertainty.

    [The high communicated uncertainty video] would also make me more likely to look for help if I felt worse as I know there is uncertainty in the diagnosis. – P15

    If symptoms don’t go away I am definitely coming back to ED. After the [low communicated uncertainty video] I would have wavered about coming back.” – P19

    A small number of participants describe how the high communicated uncertainty vignette might make them want further investigations.

    In [the high communicated uncertainty video]…having some big names thrown around like cancer made me more worried, and would be more insistent on getting tests, which probably are not needed at this point. – P20

    Discussion

    Summary of empirical findings

    This study indicates that, overall, patients are likely to prefer greater communication of diagnostic uncertainty, even though this may be more worrying—but this preference is not universal. Although quantitative data did not demonstrate significant differences in secondary outcomes such as trust, understanding, perception of doctor competence or reported likelihood of returning to seek medical advice, patients did refer to these topics in their free-text responses. The varied nature of the reasons given for their preferred vignette in the free-text responses may explain the few significant findings in the quantitative secondary outcomes: different participants sometimes gave directly opposing justifications for their preferences. There were no significant associations between demographic characteristics (eg, sex or educational level) and communication preferences.

    The study examined two general medical presentations (headache and CoBH), with broad differential diagnoses. Although our findings are not necessarily generalisable to all other clinical scenarios, the principles discussed below (in particular the need to balance harms and benefits of information disclosure) are still widely relevant. Further research will be helpful in exploring diagnostic uncertainty communication in other kinds of clinical scenarios, for example where the symptoms are more indolent or where the investigations themselves are less conclusive (such as Parkinson’s disease or motor neuron disease).

    Our data cannot alone provide definitive answers to the normative question of how doctors should communicate diagnostic uncertainty to their patients. The insights provided are, however, important in empirically grounding such discussions. If incorrect assumptions are made about what patients want to be told or what the effects of communicating diagnostic uncertainty to them are, we risk building ethical arguments founded on flawed conjectures, which may not reflect the realities of doctor-patient communication.

    Balancing harms and benefits of disclosure

    Therapeutic privilege describes the right of the doctor to not disclose material information if there is a reasonable belief that it will result in serious harm to the patient.4 54 55 Ethical analyses of the therapeutic privilege typically consider how the potential harms of disclosure (eg, aversive patient responses to distressing information) must be balanced against the potential benefits (eg, the positive impact on patient autonomy).56 Our empirical data can help inform doctors in balancing the harms versus benefits of disclosing diagnostic uncertainty.

    Perhaps the most commonly identified benefit of greater information disclosure is that it will promote patient autonomy. Autonomy is widely recognised as a central concept in bioethics, but various conceptualisations exist.57 58 Beauchamp and Childress define autonomy as the right to self-governance, to act in accordance with a self-chosen plan with intention, understanding and non-interference.59 Many bioethicists draw on this in employing a procedural definition (according to which an agent is autonomous with respect to an action if it is (i) voluntary, (ii) intentional and (iii) informed).59–61 Such procedural accounts contrast with Kantian conceptions of autonomy, in which the focus is not on autonomous decisions but on autonomous persons—individuals whose actions are determined by impartial and abstract principles of reason.61 Stirrat and Gill, building on O’Neill’s work, put forward a version of patient autonomy which requires ‘the provision of sufficient and understandable information and space for patients, who have the capacity to make a settled choice about medical interventions on themselves, to do so responsibly in a manner considerate to others’.62

    Common to many of these conceptions of autonomy is the idea that it can be limited by inadequate provision of information.63 Ethical discussions around disclosure in the doctor-patient relationship commonly link enhanced information provision to greater patient autonomy: by being more ‘fully informed’, patients are better able to make decisions regarding their health. Of course, the provision of more information does not always result in greater autonomy—patients must be able to understand the information provided, so the provision of information which is overwhelming or overly technical may restrict, rather than promote, autonomy. Indeed, Rees has argued that in some in rare cases withholding information might actually promote a patient’s autonomy.64

    In our study, the most common justification for preferring the high communicated uncertainty vignette was an appreciation of the additional information provided. The majority of participants valued greater discussion, including about uncertainties in investigations and alternate diagnoses being considered; this often facilitated a deeper understanding of their medical situation, which was construed positively. Proponents of doctor-patient communication often emphasise shared understanding as important: to promote patient autonomy and foster more informed decision-making, but also to help build trust in the doctor-patient relationship and enhance patient emotional well-being.65

    Some of our participants explicitly linked having an increased understanding to the ability to make informed decisions, for example, seeking further medical attention in the event of worsening or non-resolving symptoms. In our qualitative data, a small number of participants stated that communicating diagnostic uncertainty might be helpful in facilitating appropriate reattendance; there was no significant difference in this outcome in the quantitative data. This discrepancy may be because only a small number of patients feel this way: although communicating uncertainty may influence some patients’ behaviours, others may be likely to follow safety-netting advice regardless. Another consideration is whether there is a significant a difference between participants’ reported and actual behaviours. Most participants in our study indicated that they would be very likely to return, suggesting a ceiling effect. Economic, logistical and emotional barriers to reattending hospital exist in real life, which our vignette methodology was unable to capture. As such, further research into real clinical practice would be helpful, examining whether communicating diagnostic uncertainty influences actual patient behaviours (as opposed to patient self-report of hypothetical behaviours).

    It is also important to consider the potential benefits of shared understanding beyond patient agency. In our study, many participants did not link their understanding to the ability to make better decisions: they simply stated they preferred having more information to better understand their medical situation. Providing information about diagnostic uncertainty may thus be of value in facilitating shared understanding even if patients do not themselves want to make decisions about treatment or investigations. This is in keeping with research demonstrating that patients often have a high desire for information concerning their care, even if they have a lesser desire to be actively involved in decision-making.66–68 In this context, the importance of understanding patient preferences for both information provision and involvement in decision-making has been emphasised: as Murtagh and Thorns summarise, ‘[e]stablishing preferences enables us to show respect for patient autonomy in a manner that is sensitive and timely for that patient’.69

    An oft-discussed potential harm of disclosing diagnostic uncertainty is that it will be distressing for patients. Our study validates some concerns raised by doctors in previous research24 70 that communicating diagnostic uncertainty may worry patients: the high communicated uncertainty vignettes were perceived as more worrying (although the effect size was small). Some of our participants explicitly stated that although they found the high communicated uncertainty video more worrying, they still preferred to receive this information. This aligns with a study of patients with endometriosis in which patients expressed a preference for the communication of diagnostic uncertainty, often conceptualising it as ‘the lesser of two evils’.29 Even though information about diagnostic uncertainty can be unsettling, patients may prefer it to ignorance on the matter. Politi et al have discussed the extent to which decisional dissatisfaction may be an unavoidable consequence of communicating uncertainty in scientific evidence to patients.71 This idea can be extended to the communication of diagnostic uncertainty: it might be that some anxiety is an inevitable consequence—a ‘necessary evil’—for some patients if doctors communicate diagnostic uncertainty information with the intention of promoting autonomy.

    In summary, our data highlight that patients may prefer to receive more information about diagnostic uncertainty, even if this information is worrying. This has implications for how doctors should weigh up the potential harms and benefits of information disclosure in this context—many patients value the communication of diagnostic uncertainty in facilitating better understanding of their medical situation, despite the potential harm induced by the creation of worry. It is, however, important to acknowledge that the preference for greater diagnostic uncertainty communication was not universal. As we discuss in the next section, this presents a significant challenge for doctors.

    Communicating in the face of heterogeneous information preferences

    We demonstrated variation in patient informational preferences—although most participants preferred greater communication of diagnostic uncertainty, a minority found the amount of information provided in the high communicated uncertainty videos unhelpful/unnecessary. This heterogeneity identified aligns with existing research examining doctor-patient communication more generally: studies have demonstrated variable patient informational desires in a range of clinical contexts (including in oncology,17 72–77 end-of-life care78 and surgery79–81). This presents a challenge for clinicians, particularly in acute settings where there may not be a pre-existing doctor-patient relationship: how can doctors determine whether the specific patient in front of them will appreciate additional diagnostic uncertainty information?

    One approach might be to use demographic characteristics to help guide communication. We did not find any significant associations between demographics and communication preference; this adds to mixed results in the existing body of literature exploring whether patient demographics can predict informational needs.82 Although some studies have found characteristics such as age,72 73 83 84 sex73 85 86 or educational level78 84 87 88 to be associated with informational preferences, as a review in oncology concluded: ‘demographics do not reliably predict individual informational preferences, and studies have found contradicting results’.74 Moreover, using demographic characteristics to predict information preferences raises its own ethical issues. Studies have demonstrated that black patients receive less information from their doctors compared with white patients89–91 and that clinician implicit race bias is associated with markers of poor communication.92 The use of patient characteristics such as ethnicity or perceived educational level to determine how much information to provide risks perpetuating such implicit biases and contributing to health inequalities.

    An alternative might be to explore patient attitudes towards risk/uncertainty to help guide communication (eg, using the IUS-12, a 12-item instrument measuring reactions to uncertainty/ambiguous situations).43 It is plausible that patients with greater intolerance to uncertainty might be less likely to value explicit communication of diagnostic uncertainty. In our study, we did not find any significant associations between IUS-12 scores and primary or secondary outcomes. This suggests that exploring patients’ tolerance to uncertainty in general terms may not helpfully predict their information preferences or how they will respond to uncertainty communication. The logistical challenges presented by using such scale must also be considered—even if a scale was to be developed that accurately predicts patient desire for uncertainty information, it would also need to be practically feasible for use. Such logistical barriers might account for why some such scales have been developed,67 68 93 94 but are not widely used.

    Finally, doctors could directly ask patients what they want to be told to help guide communication. Although this approach has some appeal, several problems arise in relying on patients to inform their doctors what information they would like to receive. First, it is difficult—arguably impossible—for patients who are ignorant about something to indicate whether they want to be told about it. For example, in the headache scenario, many patients may not have considered that a CT does not exclude a subarachnoid haemorrhage. They are therefore unlikely to ask for more information about this—but this is not to say they might not appreciate the information if provided. Second, relying on patients to guide information disclosure assumes that patients know, and are able to articulate, their own informational preferences.95 In practice, patients do not always voice their worries: a study in primary care found that patients commonly have unvoiced concerns regarding possible diagnoses.96 Relying on patients to articulate their informational preferences may perpetuate health inequalities: patients from minoritised groups are less likely to ask questions and tend to adopt a more passive role in the doctor-patient relationship.89 91 97 98 There is thus a risk that some patients will feel less empowered to ask for information and will be disadvantaged.

    Ultimately, the heterogeneity of views identified in this study presents a challenge for doctors in determining how much diagnostic uncertainty information to communicate to the specific patient in front of them. There are the problems associated with trying to base communication on demographic characteristics, scales or even patients’ own expressed wishes—not least the risk of perpetuating existing biases and inequalities.

    When deciding how to communicate diagnostic uncertainty, doctors should work with patients to elicit their informational preferences as far as is possible—we echo Nease and Brooks in their assertion that ‘[r]igid recommendations about how much information patients should receive and the degree to which patients should be involved in medical decision making may be inappropriate when applied to individual patients’. Yet it may not always be feasible to fully explore patient informational preferences in acute settings, and patients themselves may sometimes be unable to articulate such preferences. Pragmatically, we suggest that doctors could employ a utilitarian approach to address this challenge. If, as our study suggests, the majority of patients have a preference for greater disclosure of diagnostic uncertainty, then doctors should generally disclose this information (unless the patient has made it clear that they do not want this information)—as this will produce the best outcome for the majority of patients. When faced with a situation where it is not clear how much information a patient might want, we argue that the most ethical approach is to err on the side of greater disclosure of uncertainty information.

    Strengths and limitations

    Unlike observational studies of real consultations, this study’s vignette methodology permitted communication behaviours to be isolated and manipulated, allowing inferences to be drawn about causal relationships between diagnostic uncertainty communication and various outcomes. A strength of the crossover design was that participants were able to watch both the high and low communicated uncertainty vignettes and directly compare them, indicating their preference. The inclusion of a free text section on the questionnaire permitted the collection of qualitative data, which was helpful in illustrating and interpreting the quantitative findings.

    We chose to focus on two clinical scenarios, both of which are commonly encountered in general medicine. The fact that there was little difference between the responses given by participants who watched the headache vignettes versus the CoBH vignettes suggests some of our findings may be generalisable to other clinical scenarios, but further research is needed to explore this possibility.

    An inherent weakness of vignette methodology relates to the degree to which participants respond to vignettes as they would a real consultation. Previous research has suggested that analogue patients can be included as proxies for clinical patients in studies on communication,99 and our pilot-testing results demonstrated that the vignettes had adequate internal and external validity. The vignettes were developed using established guidelines,42 with input from experts and laypeople in their development to make them as realistic as possible. Additionally, we used data collected from a previous study to develop the scripts, meaning that they reflected what real doctors would say in communicating diagnostic uncertainty. Thus, although responses to the vignettes may not be wholly equivalent to real consultations, they were as realistic as possible, and it is likely that our participants were able to respond to them in way which reasonably reflects real patient reactions.

    As part of the pilot-testing process, we tested immersion and perceived realism of the vignettes—the final videos used in the main study performed well.41 We did not, however, test this in the main study (in part due to concerns about overburdening participants with multiple questionnaires); it is something we would consider doing in future vignette studies.

    The nature of the study design precluded any conversation between participants and the doctor—the vignettes all depicted a doctor monologue to camera. Real consultations typically involve at least some degree of conversation, with patients asking questions and discussing information with their doctor. This study was unable to examine the effect of communicating diagnostic uncertainty in the context of a dialogue between patient and doctor. Although attempts were made to ensure that the sample was as diverse as possible, we acknowledge that our sample had a higher level of university education than the general population, likely a result of self-selection of individuals who chose to participate. We did not, however, find any significant difference in the primary outcome (preferred vignette) when we compared those with a university degree or higher and those without.

    Conclusion

    Our study shows an overall patient preference for greater communication of diagnostic uncertainty, notwithstanding the potential for higher communicated uncertainty to be marginally more worrying. This provides some empirical support for recommendations to openly communicate diagnostic uncertainty to patients: although diagnostic uncertainty information might worry patients, they may still want to understand it.

    We note that patient preferences are not uniform, and it can be difficult for clinicians to determine what an individual patients might want to be told, particularly in the absence of an existing relationship – but we caution against using demographic data to make assumptions about informational preferences, as this may create or exacerbate health communication inequalities.

    Overall, we suggest that clinicians should err towards greater communication of diagnostic uncertainty to their patients. If healthcare professionals elect not to engage in dialogue with patients about diagnostic uncertainty (based on benign paternalistic ideas about avoiding patient worry) or do so variably (based on unevidenced assumptions about patient information preferences), they risk depriving patients of information they may value.

    Data availability statement

    Data are available upon reasonable request. The participants of this study did not give written consent for their data to be shared publicly, only that the anonymised data collected may be used to support other researchers in the future. As such, the data that support the findings of this study are only available on request from the corresponding author, CC.

    Ethics statements

    Patient consent for publication

    Ethics approval

    This study involved human participants and was approved by the University of Cambridge Psychology Research Ethics Committee (Ethics Application PRE.2022.065). Participants gave informed consent to participate in the study before taking part.

    References

    1. ↵
      1. Fallowfield LJ,
      2. Jenkins VA,
      3. Beveridge HA
      . Truth may hurt but deceit hurts more: communication in palliative care. Palliat Med 2002;16:297–303. doi:10.1191/0269216302pm575oa
      OpenUrlCrossRefPubMedWeb of Science
    2. ↵
      1. Surbone A
      . Telling the truth to patients with cancer: what is the truth? Lancet Oncol 2006;7:944–50. doi:10.1016/S1470-2045(06)70941-X
      OpenUrlCrossRefPubMedWeb of Science
    3. ↵
      1. Cox CL,
      2. Fritz Z
      . Should non-disclosures be considered as morally equivalent to lies within the doctor-patient relationship? J Med Ethics 2016;42:632–5. doi:10.1136/medethics-2015-103014
      OpenUrlAbstract/FREE Full Text
    4. ↵
      1. Johnston C,
      2. Holt G
      . The legal and ethical implications of therapeutic privilege - is it ever justified to withhold treatment information from a competent patient? Clin Ethics 2006;1:146–51. doi:10.1258/147775006778246450
      OpenUrlCrossRef
    5. ↵
      1. Chamberlain CJ,
      2. Koniaris LG,
      3. Wu AW
      . Disclosure of “Nonharmful” Medical Errors and Other Events. Arch Surg 2012;147:282. doi:10.1001/archsurg.2011.1005
      OpenUrlCrossRefPubMed
    6. ↵
      1. Novack DH,
      2. Plumer R,
      3. Smith RL, et al
      . Changes in physicians’ attitudes toward telling the cancer patient. JAMA 1979;241:897–900.
      OpenUrlCrossRefPubMedWeb of Science
    7. ↵
      1. Sokol DK
      . How the doctor’s nose has shortened over time; a historical overview of the truth-telling debate in the doctor-patient relationship. J R Soc Med 2006;99:632–6. doi:10.1177/014107680609901212
      OpenUrlCrossRefPubMedWeb of Science
    8. ↵
      1. Brody H
      . Transparency: informed consent in primary care. Hastings Cent Rep 1989;19:5–9.
      OpenUrl
    9. ↵
      1. Charles C,
      2. Gafni A,
      3. Whelan T
      . Shared decision-making in the medical encounter: what does it mean? (or it takes at least two to tango). Soc Sci Med 1997;44:681–92. doi:10.1016/s0277-9536(96)00221-3
      OpenUrlCrossRefPubMedWeb of Science
    10. ↵
      Montgomery v lanarkshire health board. UKSC; 2015.
    11. ↵
      1. Jones MA
      . Informed consent and other fairy stories. Med Law Rev 1999;7:103–34. doi:10.1093/medlaw/7.2.103
      OpenUrlCrossRefPubMed
    12. ↵
      1. Sokol DK
      . Truth-telling in the doctor-patient relationship: a case analysis. Clin Ethics 2006;1:130–4. doi:10.1258/147775006778246531
      OpenUrlCrossRef
    13. ↵
      1. Wu AW,
      2. Cavanaugh TA,
      3. McPhee SJ, et al
      . To tell the truth: ethical and practical issues in disclosing medical mistakes to patients. J Gen Intern Med 1997;12:770–5. doi:10.1046/j.1525-1497.1997.07163.x
      OpenUrlCrossRefPubMedWeb of Science
    14. ↵
      1. Beauchamp TL,
      2. Childless JF
      . Principles of Biomedical Ethics. Int Clin Psychopharmacol 1991;6:129–30. doi:10.1097/00004850-199100620-00013
      OpenUrlCrossRef
    15. ↵
      1. Edwin A
      . Non-disclosure of medical errors an egregious violation of ethical principles. Ghana Med J 2009;43:34–9.
      OpenUrlPubMed
    16. ↵
      1. Sarafis P,
      2. Tsounis A,
      3. Malliarou M, et al
      . Disclosing the Truth: A Dilemma between Instilling Hope and Respecting Patient Autonomy in Everyday Clinical Practice. GJHS 2014;6:128. doi:10.5539/gjhs.v6n2p128
      OpenUrlPubMed
    17. ↵
      1. Butow P,
      2. Dowsett S,
      3. Hagerty R, et al
      . Communicating prognosis to patients with metastatic disease: what do they really want to know? Support Care Cancer 2002;10:161–8. doi:10.1007/s005200100290
      OpenUrlCrossRefPubMedWeb of Science
    18. ↵
      1. Hagerty RG,
      2. Butow PN,
      3. Ellis PM, et al
      . Communicating With Realism and Hope: Incurable Cancer Patients’ Views on the Disclosure of Prognosis. JCO 2005;23:1278–88. doi:10.1200/JCO.2005.11.138
      OpenUrlAbstract/FREE Full Text
    19. ↵
      1. Berger ZD,
      2. Brito JP,
      3. Ospina NS, et al
      . Patient centred diagnosis: sharing diagnostic decisions with patients in clinical practice. BMJ 2017;359:j4218. doi:10.1136/bmj.j4218
    20. ↵
      1. Meyer AND,
      2. Giardina TD,
      3. Khawaja L, et al
      . Patient and clinician experiences of uncertainty in the diagnostic process: Current understanding and future directions. Patient Educ Couns 2021;104:2606–15. doi:10.1016/j.pec.2021.07.028
      OpenUrlCrossRefPubMed
    21. ↵
      1. Dahm MR,
      2. Crock C
      . Understanding and Communicating Uncertainty in Achieving Diagnostic Excellence. JAMA 2022;327:1127–8. doi:10.1001/jama.2022.2141
      OpenUrlCrossRefPubMed
    22. ↵
      1. Medendorp NM,
      2. Stiggelbout AM,
      3. Aalfs CM, et al
      . A scoping review of practice recommendations for clinicians’ communication of uncertainty. Health Expect 2021;24:1025–43. doi:10.1111/hex.13255
      OpenUrlCrossRefPubMed
    23. ↵
      1. GMC
      . Guidance on professional standards and ethics for doctors: decision making and consent. Manchester, 2020.
    24. ↵
      1. Cox C,
      2. Hatfield T,
      3. Fritz Z
      . How and why do doctors communicate diagnostic uncertainty: An experimental vignette study. Health Expect 2024;27:e13957. doi:10.1111/hex.13957
    25. ↵
      1. Cox C,
      2. Miller B,
      3. Kuhn I, et al
      . Diagnostic uncertainty in primary care: What is known about its communication, and what are the associated ethical issues. Fam Pract 2021. doi:10.17863/CAM.65222
    26. ↵
      1. Hart J,
      2. Cox CL,
      3. Kuhn I, et al
      . Communicating diagnostic uncertainty in the acute and emergency medical setting: A systematic review and ethical analysis of the empirical literature. Acute Med 2021;20:204–18.
      OpenUrlCrossRefPubMed
    27. ↵
      1. Johnson CG,
      2. Levenkron JC,
      3. Suchman AL, et al
      . Does physician uncertainty affect patient satisfaction? J Gen Intern Med 1988;3:144–9. doi:10.1007/BF02596120
      OpenUrlCrossRefPubMedWeb of Science
    28. ↵
      1. Bhise V,
      2. Meyer AND,
      3. Menon S, et al
      . Patient perspectives on how physicians communicate diagnostic uncertainty: An experimental vignette study. Int J Qual Health Care 2018;30:2–8. doi:10.1093/intqhc/mzx170
      OpenUrlCrossRefPubMed
    29. ↵
      1. Bontempo AC
      . Patient attitudes toward clinicians’ communication of diagnostic uncertainty and its impact on patient trust. SSM - Qualitative Research in Health 2023;3:100214. doi:10.1016/j.ssmqr.2022.100214
      OpenUrlCrossRef
    30. ↵
      1. Edwards PJ,
      2. Ridd MJ,
      3. Sanderson E, et al
      . Development of a tool for coding safety-netting behaviours in primary care: a mixed-methods study using existing UK consultation recordings. Br J Gen Pract 2019;69:e869–77. doi:10.3399/bjgp19X706589
      OpenUrlAbstract/FREE Full Text
    31. ↵
      1. Almond S,
      2. Mant D,
      3. Thompson M
      . Diagnostic safety-netting. Br J Gen Pract 2009;59:872–4. doi:10.3399/bjgp09X472971
      OpenUrlFREE Full Text
    32. ↵
      1. Edwards PJ,
      2. Silverston P,
      3. Sprackman J, et al
      . Safety-netting in the consultation. BMJ 2022;378:e069094. doi:10.1136/bmj-2021-069094
    33. ↵
      1. Jones D,
      2. Dunn L,
      3. Watt I, et al
      . Safety netting for primary care: evidence from a literature review. Br J Gen Pract 2019;69:e70–9. doi:10.3399/bjgp18X700193
      OpenUrlAbstract/FREE Full Text
    34. ↵
      1. Friedemann Smith C,
      2. Lunn H,
      3. Wong G, et al
      . Optimising GPs’ communication of advice to facilitate patients’ self-care and prompt follow-up when the diagnosis is uncertain: a realist review of “safety-netting” in primary care. BMJ Qual Saf 2022;31:541–54. doi:10.1136/bmjqs-2021-014529
      OpenUrlAbstract/FREE Full Text
    35. ↵
      1. Miao M,
      2. Dahm MR,
      3. Li J, et al
      . Managing Uncertainty During the Communication of Diagnostic Test Information Between Patients and Clinicians in Australian Emergency Care. Qual Health Res 2020;30:1287–300. doi:10.1177/1049732320913037
      OpenUrlCrossRefPubMed
    36. ↵
      1. Davies R,
      2. Ives J,
      3. Dunn M
      . A systematic review of empirical bioethics methodologies. BMC Med Ethics 2015;16:15. doi:10.1186/s12910-015-0010-3
      OpenUrlCrossRefPubMed
    37. ↵
      1. Frith L
      . Symbiotic empirical ethics: a practical methodology. Bioethics 2012;26:198–206. doi:10.1111/j.1467-8519.2010.01843.x
      OpenUrlCrossRefPubMedWeb of Science
    38. ↵
      1. Huxtable R,
      2. Ives J
      . Mapping, framing, shaping: a framework for empirical bioethics research projects. BMC Med Ethics 2019;20:86. doi:10.1186/s12910-019-0428-0
    39. ↵
      1. Dunn M,
      2. Sheehan M,
      3. Hope T, et al
      . Toward Methodological Innovation in Empirical Ethics Research. Camb Q Healthc Ethics 2012;21:466–80. doi:10.1017/S0963180112000242
      OpenUrlCrossRefPubMed
    40. ↵
      1. Daniels N
      . Wide Reflective Equilibrium and Theory Acceptance in Ethics. J Philos 1979;76:256. doi:10.2307/2025881
      OpenUrlCrossRefWeb of Science
    41. ↵
      1. Cox C,
      2. Hatfield T,
      3. Moxey J, et al
      . Creating and administering video vignettes for a study examining the communication of diagnostic uncertainty: Methodological insights to improve accessibility for researchers and participants. BMC Med Res Methodol 2023;23. doi:10.1186/s12874-023-02072-7
    42. ↵
      1. Hillen MA,
      2. van Vliet LM,
      3. de Haes HCJM, et al
      . Developing and administering scripted video vignettes for experimental research of patient-provider communication. Patient Educ Couns 2013;91:295–309. doi:10.1016/j.pec.2013.01.020
      OpenUrlCrossRefPubMed
    43. ↵
      1. Carleton RN,
      2. Norton MAPJ,
      3. Asmundson GJG
      . Fearing the unknown: a short version of the Intolerance of Uncertainty Scale. J Anxiety Disord 2007;21:105–17. doi:10.1016/j.janxdis.2006.03.014
      OpenUrlCrossRefPubMedWeb of Science
    44. ↵
      1. Beckstead JW
      . On measurements and their quality. Paper 4: verbal anchors and the number of response options in rating scales. Int J Nurs Stud 2014;51:807–14. doi:10.1016/j.ijnurstu.2013.09.004
      OpenUrlCrossRefPubMed
    45. ↵
      1. O’Muircheartaigh CA,
      2. Krosnick JA,
      3. Helic A
      . Middle alternatives, acquiescence, and the quality of questionnaire data: Irving B. Harris Graduate School of Public Policy Studies, University of Chicago; 2001.
    46. ↵
      1. Rattray J,
      2. Jones MC
      . Essential elements of questionnaire design and development. J Clin Nurs 2007;16:234–43. doi:10.1111/j.1365-2702.2006.01573.x
      OpenUrlCrossRefPubMed
    47. ↵
      1. Creswell JW,
      2. Plano Clark VL,
      3. Gutmann ML, et al
      . Advanced mixed methods research designs. In: Tashakkori A, Teddlie C, eds. Handbook of mixed methods in social and behavioral research. California2013.
    48. ↵
      1. Jogulu UD,
      2. Pansiri J
      . Mixed methods: a research design for management doctoral dissertations. Management Research Review 2011;34:687–701. doi:10.1108/01409171111136211
      OpenUrlCrossRef
    49. ↵
      1. Driscoll DL,
      2. Appiah-Yeboah A,
      3. Salib P, et al
      . Merging qualitative and quantitative data in mixed methods research: How to and why not. 2007.
    50. ↵
      R: a language and environment for statistical computing. Vienna, Austria R Foundation for Statistical Computing; 2023.
    51. ↵
      1. Braun V,
      2. Clarke V
      . Using thematic analysis in psychology. Qual Res Psychol 2006;3:77–101. doi:10.1191/1478088706qp063oa
      OpenUrlCrossRefPubMed
    52. ↵
      1. Terry G,
      2. Hayfield N,
      3. Clarke V, et al
      . Thematic analysis. In: The SAGE Handbook of Qualitative Research in Psychology. 2017: 17–37.
    53. ↵
      1. Byrne D
      . A worked example of Braun and Clarke’s approach to reflexive thematic analysis. Qual Quant 2022;56:1391–412. doi:10.1007/s11135-021-01182-y
      OpenUrlCrossRef
    54. ↵
      1. Hodkinson K
      . The need to know--therapeutic privilege: a way forward. Health Care Anal 2013;21:105–29. doi:10.1007/s10728-012-0204-5
      OpenUrlCrossRefPubMed
    55. ↵
      1. Richard C,
      2. Lajeunesse Y,
      3. Lussier MT
      . Therapeutic privilege: between the ethics of lying and the practice of truth. J Med Ethics 2010;36:353–7. doi:10.1136/jme.2009.033340
      OpenUrlAbstract/FREE Full Text
    56. ↵
      1. Berger JT
      . Ignorance Is Bliss? Ethical Considerations in Therapeutic Nondisclosure. LCNV 2005;23:94–8. doi:10.1081/CNV-200046398
      OpenUrlCrossRef
    57. ↵
      1. Schneider C
      . The Practice of Autonomy: Patients, Doctors, and Medical Decisions. Oxford University Press, 1998:307.
    58. ↵
      1. Slowther A-M
      . The concept of autonomy and its interpretation in health care. Clin Ethics 2007;2:173–5. doi:10.1258/147775007783560166
      OpenUrlCrossRef
    59. ↵
      1. Beauchamp TL,
      2. Childress JF
      . Principles of Biomedical Ethics. New York: Oxford University Press, 2001:454.
    60. ↵
      1. Pugh J
      . Autonomy, Rationality, and Contemporary Bioethics. Oxford University Press, 2020.
    61. ↵
      1. Varelius J
      . The value of autonomy in medical ethics. Med Health Care Philos 2006;9:377–88. doi:10.1007/s11019-006-9000-z
      OpenUrlCrossRefPubMed
    62. ↵
      1. Stirrat GM,
      2. Gill R
      . Autonomy in medical ethics after O’Neill. J Med Ethics 2005;31:127–30. doi:10.1136/jme.2004.008292
      OpenUrlAbstract/FREE Full Text
    63. ↵
      1. Cox CL
      . Patient understanding: How should it be defined and assessed in clinical practice? J Eval Clin Pract 2023;29:1127–34. doi:10.1111/jep.13882
      OpenUrlCrossRefPubMed
    64. ↵
      1. Rees M
      . Patient autonomy and withholding information. Bioethics 2023;37:256–64. doi:10.1111/bioe.13130
      OpenUrlCrossRefPubMed
    65. ↵
      1. Street RL,
      2. Makoul G,
      3. Arora NK, et al
      . How does communication heal? Pathways linking clinician-patient communication to health outcomes. Patient Educ Couns 2009;74:295–301. doi:10.1016/j.pec.2008.11.015
      OpenUrlCrossRefPubMedWeb of Science
    66. ↵
      1. Robinson A,
      2. Thomson R
      . Variability in patient preferences for participating in medical decision making: implication for the use of decision support tools. Quality and Safety in Health Care 2001;10:i34–8. doi:10.1136/qhc.0100034
      OpenUrlAbstract/FREE Full Text
    67. ↵
      1. Nease RF,
      2. Brooks WB
      . Patient desire for information and decision making in health care decisions: the autonomy preference index and the health opinion survey. J Gen Intern Med 1995;10:593–600. doi:10.1007/BF02602742
      OpenUrlCrossRefPubMedWeb of Science
    68. ↵
      1. Ende J,
      2. Kazis L,
      3. Ash A, et al
      . Measuring patients’ desire for autonomy: decision making and information-seeking preferences among medical patients. J Gen Intern Med 1989;4:23–30. doi:10.1007/BF02596485
      OpenUrlCrossRefPubMedWeb of Science
    69. ↵
      1. Murtagh FEM,
      2. Thorns A
      . Evaluation and ethical review of a tool to explore patient preferences for information and involvement in decision making. J Med Ethics 2006;32:311–5. doi:10.1136/jme.2005.012484
      OpenUrlAbstract/FREE Full Text
    70. ↵
      1. Rising KL,
      2. Papanagnou D,
      3. McCarthy D, et al
      . Emergency Medicine Resident Perceptions About the Need for Increased Training in Communicating Diagnostic Uncertainty. Cureus 2018;10:e2088. doi:10.7759/cureus.2088
    71. ↵
      1. Politi MC,
      2. Clark MA,
      3. Ombao H, et al
      . Communicating uncertainty can lead to less decision satisfaction: a necessary cost of involving patients in shared decision making? Health Expect 2011;14:84–91. doi:10.1111/j.1369-7625.2010.00626.x
      OpenUrlCrossRefPubMedWeb of Science
    72. ↵
      1. Giacalone A,
      2. Blandino M,
      3. Talamini R, et al
      . What elderly cancer patients want to know? Differences among elderly and young patients. Psychooncology 2007;16:365–70. doi:10.1002/pon.1065
      OpenUrlCrossRefPubMed
    73. ↵
      1. Jenkins V,
      2. Fallowfield L,
      3. Saul J
      . Information needs of patients with cancer: results from a large study in UK cancer centres. Br J Cancer 2001;84:48–51. doi:10.1054/bjoc.2000.1573
      OpenUrlCrossRefPubMedWeb of Science
    74. ↵
      1. Russell BJ,
      2. Ward AM
      . Deciding what information is necessary: do patients with advanced cancer want to know all the details? Cancer Manag Res 2011;3:191–9. doi:10.2147/CMR.S12998
      OpenUrlPubMed
    75. ↵
      1. Schattner A
      . What Do Patients Really Want to Know? Oxford University Press, 2002:135–6.
    76. ↵
      1. Leydon GM,
      2. Boulton M,
      3. Moynihan C, et al
      . Cancer patients’ information needs and information seeking behaviour: in depth interview study. BMJ 2000;320:909–13. doi:10.1136/bmj.320.7239.909
      OpenUrlAbstract/FREE Full Text
    77. ↵
      1. Innes S,
      2. Payne S
      . Advanced cancer patients’ prognostic information preferences: a review. Palliat Med 2009;23:29–39. doi:10.1177/0269216308098799
      OpenUrlCrossRefPubMedWeb of Science
    78. ↵
      1. De Vleminck A,
      2. Pardon K,
      3. Roelands M, et al
      . Information preferences of the general population when faced with life-limiting illness. Eur J Public Health 2015;25:532–8. doi:10.1093/eurpub/cku158
      OpenUrlCrossRefPubMed
    79. ↵
      1. Beresford N,
      2. Seymour L,
      3. Vincent C, et al
      . Risks of elective cardiac surgery: what do patients want to know? Heart 2001;86:626–31. doi:10.1136/heart.86.6.626
      OpenUrlAbstract/FREE Full Text
    80. ↵
      1. Degerliyurt K,
      2. Gunsolley JC,
      3. Laskin DM
      . Informed consent: what do patients really want to know? J Oral Maxillofac Surg 2010;68:1849–52. doi:10.1016/j.joms.2010.04.004
      OpenUrlCrossRefPubMed
    81. ↵
      1. McNair AGK,
      2. MacKichan F,
      3. Donovan JL, et al
      . What surgeons tell patients and what patients want to know before major cancer surgery: a qualitative study. BMC Cancer 2016;16:258. doi:10.1186/s12885-016-2292-3
    82. ↵
      1. Say R,
      2. Murtagh M,
      3. Thomson R
      . Patients’ preference for involvement in medical decision making: a narrative review. Patient Educ Couns 2006;60:102–14. doi:10.1016/j.pec.2005.02.003
      OpenUrlCrossRefPubMedWeb of Science
    83. ↵
      1. Chamot E,
      2. Charvet AI,
      3. Perneger TV
      . Variability in women’s desire for information about mammography screening: implications for informed consent. Eur J Cancer Prev 2005;14:413–8. doi:10.1097/00008469-200508000-00015
      OpenUrlCrossRefPubMedWeb of Science
    84. ↵
      1. Ziegler DK,
      2. Mosier MC,
      3. Buenaver M, et al
      . How much information about adverse effects of medication do patients want from physicians? Arch Intern Med 2001;161:706–13. doi:10.1001/archinte.161.5.706
      OpenUrlCrossRefPubMedWeb of Science
    85. ↵
      1. Elkin EB,
      2. Kim SHM,
      3. Casper ES, et al
      . Desire for information and involvement in treatment decisions: elderly cancer patients’ preferences and their physicians’ perceptions. J Clin Oncol 2007;25:5275–80. doi:10.1200/JCO.2007.11.1922
      OpenUrlAbstract/FREE Full Text
    86. ↵
      1. Butow PN,
      2. Maclean M,
      3. Dunn SM, et al
      . The dynamics of change: cancer patients’ preferences for information, involvement and support. Ann Oncol 1997;8:857–63. doi:10.1023/a:1008284006045
      OpenUrlCrossRefPubMedWeb of Science
    87. ↵
      1. Lagarde SM,
      2. Franssen SJ,
      3. van Werven JR, et al
      . Patient preferences for the disclosure of prognosis after esophagectomy for cancer with curative intent. Ann Surg Oncol 2008;15:3289–98. doi:10.1245/s10434-008-0068-y
      OpenUrlCrossRefPubMedWeb of Science
    88. ↵
      1. Davis MA,
      2. Hoffman JR,
      3. Hsu J
      . Impact of Patient Acuity on Preference for Information and Autonomy in Decision Making. Acad Emerg Med 1999;6:781–5. doi:10.1111/j.1553-2712.1999.tb01206.x
      OpenUrlCrossRefPubMedWeb of Science
    89. ↵
      1. Gordon HS,
      2. Street RL,
      3. Sharf BF, et al
      . Racial differences in doctors’ information-giving and patients’ participation. Cancer 2006;107:1313–20. doi:10.1002/cncr.22122
      OpenUrlCrossRefPubMedWeb of Science
    90. ↵
      1. Street RL,
      2. Gordon H,
      3. Haidet P
      . Physicians’ communication and perceptions of patients: is it how they look, how they talk, or is it just the doctor? Soc Sci Med 2007;65:586–98. doi:10.1016/j.socscimed.2007.03.036
      OpenUrlCrossRefPubMed
    91. ↵
      1. Siminoff LA,
      2. Graham GC,
      3. Gordon NH
      . Cancer communication patterns and the influence of patient characteristics: disparities in information-giving and affective behaviors. Patient Educ Couns 2006;62:355–60. doi:10.1016/j.pec.2006.06.011
      OpenUrlCrossRefPubMedWeb of Science
    92. ↵
      1. Cooper LA,
      2. Roter DL,
      3. Carson KA, et al
      . The associations of clinicians’ implicit attitudes about race with medical visit communication and patient ratings of interpersonal care. Am J Public Health 2012;102:979–87. doi:10.2105/AJPH.2011.300558
      OpenUrlCrossRefPubMedWeb of Science
    93. ↵
      1. Golin CE,
      2. DiMatteo MR,
      3. Leake B, et al
      . A diabetes-specific measure of patient desire to participate in medical decision making. Diabetes Educ 2001;27:875–86. doi:10.1177/014572170102700613
      OpenUrlCrossRefPubMed
    94. ↵
      1. Duggan C,
      2. Bates I,
      3. Sturman S, et al
      . Validation of a “desire for information” scale. Int J Pharm Pract 2011;10:31–7. doi:10.1111/j.2042-7174.2002.tb00585.x
      OpenUrlCrossRef
    95. ↵
      1. Epstein RM,
      2. Peters E
      . Beyond information: exploring patients’ preferences. JAMA 2009;302:195–7. doi:10.1001/jama.2009.984
      OpenUrlCrossRefPubMedWeb of Science
    96. ↵
      1. Barry CA,
      2. Bradley CP,
      3. Britten N, et al
      . Patients’ unvoiced agendas in general practice consultations: qualitative study. BMJ 2000;320:1246–50. doi:10.1136/bmj.320.7244.1246
      OpenUrlAbstract/FREE Full Text
    97. ↵
      1. Schouten BC,
      2. Meeuwesen L
      . Cultural differences in medical communication: a review of the literature. Patient Educ Couns 2006;64:21–34. doi:10.1016/j.pec.2005.11.014
      OpenUrlCrossRefPubMedWeb of Science
    98. ↵
      1. Ferguson WJ,
      2. Candib LM
      . Culture, language, and the doctor-patient relationship. Fam Med 2002;34:353–61.
      OpenUrlPubMedWeb of Science
    99. ↵
      1. van Vliet LM,
      2. van der Wall E,
      3. Albada A, et al
      . The validity of using analogue patients in practitioner-patient communication research: systematic review and meta-analysis. J Gen Intern Med 2012;27:1528–43. doi:10.1007/s11606-012-2111-8
      OpenUrlCrossRefPubMed

    Footnotes

    • X @caitriona_cox, @drzoefritz

    • Contributors CLC and ZBMCF conceived of the study and wrote the protocol. ZBMCF, CLC and TH contributed towards the development of the video vignettes and undertook pilot-testing. The initial quantitative analysis was performed by MP, with qualitative analysis by CLC and TH. The first draft was produced by CLC with input from TH and MP and supervision from ZBMCF. The final draft is the result of close collaboration and input from all authors. ZBMCF is the guarantor of the work.

    • Funding This study was funded by Wellcome Trust (208213/Z/17/Z).

    • Competing interests None declared.

    • Provenance and peer review Not commissioned; externally peer-reviewed.

    • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.

    • ↵Note that in an AB|BA design, a carry-over effect cannot be distinguished mathematically from an interaction effect (between treatment (A or B) and period (first or second)), or a sequence effect (AB or BA).