Article Text

Download PDFPDF

Vulnerability in palliative care research: findings from a qualitative study of black Caribbean and white British patients with advanced cancer
  1. J Koffman1,
  2. M Morgan2,
  3. P Edmonds1,
  4. P Speck1,
  5. I J Higginson1
  1. 1
    Department of Palliative Care, Policy and Rehabilitation, King’s College London, London, UK
  2. 2
    Department of Public Health Sciences, King’s College London, London, UK
  1. Dr Jonathan Koffman, King’s College London, Department of Palliative Care, Policy and Rehabilitation, Weston Education Centre, Cutcombe Road, London SE5 9RJ, UK; jonathan.koffman{at}


Introduction: Vulnerability is a poorly understood concept in research ethics, often aligned to autonomy and consent. A recent addition to the literature represents a taxonomy of vulnerability developed by Kipnis, but this refers to the conduct of clinical trials rather than qualitative research, which may raise different issues.

Aim: To examine issues of vulnerability in cancer and palliative care research obtained through qualitative interviews.

Method: Secondary analysis of qualitative data from 26 black Caribbean and 19 white British patients with advanced cancer.

Results: Four domains of vulnerability derived from Kipnis’s taxonomy were identified and included: (i) communicative vulnerability, represented by participants impaired in their ability to communicate because of distressing symptoms; (ii) institutional vulnerability, which referred to participants who existed under the authority of others—for example, in hospital; (iii) deferential vulnerability, which included participants who were subject to the informal authority or the independent interests of others; (iv) medical vulnerability, which referred to participants with distressing medical conditions; and (v) social vulnerability, which included participants considered to belong to an undervalued social group. Participants from both ethnic groups populated all these domains, but those who were black Caribbean were more present among the socially vulnerable.

Conclusions: Current classifications of vulnerability require reinterpretation when applied to qualitative research at the end of life. We recommend that researchers and research ethics committees reconceptualise vulnerability using the domains identified in this study and consider the research context and interviewers’ skills.

Statistics from

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.

Palliative care research raises important ethical concerns.14 Given the perceived vulnerability of those included in research, it is often believed that the methods of data collection may be upsetting, a situation amplified by patient or family distress associated with their disease or circumstances.5 Questions have therefore been raised as to whether it is appropriate to engage these individuals in research: because, first, they may not wish to make use of their remaining time engaged in research; second, research could be viewed as exploitative, since they are a captive audience;6 third, as with other research, it is unlikely that they stand to benefit directly from their participation.7 8 Although some research bodies9 have emphasised that special care should be taken where research participants are considered vulnerable, there is confusion about what this term implies and how it should be applied to research in general and palliative care in particular.

Definitions of “vulnerability” vary, usually referring to individuals with limited cognitive abilities or diminished autonomy.1012 Vulnerable populations, however, may possess autonomy but lack capacity to communicate opinions regarding participation in research. Moreover, this definition does not adequately engage with the context (social as well as medical) of research participants, which may create situations of vulnerability. Kipnis presents a more helpful definition of vulnerability as being an intrinsic or situational condition that puts some individuals at greater risk of being used in research in ethically inappropriate ways.13 Within these two domains, Kipnis developed six caution areas intended to identify individuals, or groups of people, considered to be vulnerable, and advice on how researchers should manage them in research studies. The emphasis of this taxonomy is on conducting clinical trials. While each category of vulnerability acts as a signal for researchers to initiate study safeguards, the taxonomy also serves as a checklist of criteria to potentially prevent the inclusion of these groups. This supports the principle of fairness, highlighted by the Belmont report,14 to exercise care among those who, historically, were selected inappropriately for research. However, there are also concerns about the unfair distribution of benefits of research. Protections that surround vulnerable populations may have unintended consequences. First, the principle of justice, often regarded as being synonymous with fairness, is compromised, as they are denied the opportunity of contributing to the research process.15 Second, being excluded from research potentially deprives them of the benefits and outcomes of being studied.16 Therefore, medical advances may not be relevant to their needs or situation. These polarised positions should represent a concern for researchers and members of research ethics committees, since it is not always apparent what to do when confronted with persons considered to be vulnerable.

In this paper, we use secondary analysis of data collected as part of a recent qualitative study that examined the experience of advanced cancer among black Caribbean and white British patients.17 18 We attempt to examine a number of inter-related domains of vulnerability that characterised our study population (cancer patients often cared for in an institution) within the context of qualitative research. We also examine to what extent being a member of a black and minority ethnic (BME) group is a factor that might constitute a domain of vulnerability. Although there has been some interest in research ethics in qualitative research,19 we were unable to find where this concern has been explored in relation to vulnerable populations. The findings from this analysis may therefore offer lessons for people planning research among these groups and may also guide research ethics committees’ decision-making about research studies that engage with vulnerable populations using qualitative methods.


Study setting and participants

The study took place in three south London boroughs characterised by social deprivation and a high concentration of ethnic minorities. We attempted to recruit a diverse group of adult black Caribbean and white British participants from palliative care and oncology teams, paying attention to age, cancer site and location of care. Ethics approval was obtained before data collection.


After providing informed consent, participants engaged in informal interviews, beginning with a general discussion about how they felt about cancer and its causation, and then progressed to explore symptoms that bothered or troubled them, particularly as their illness had progressed. Questions also explored how participants lived with their illness. All interviews were tape-recorded and lasted on average 40 minutes (range 20–60 minutes).


Secondary analysis of the interview transcripts was conducted using the framework method to organise data and identify emerging themes. This matrix-based approach was used to develop a thematic framework through independent repeated reading of the interview transcripts.20 To address issues of analytical rigour and trustworthiness, a selection of interview transcripts were also reviewed by MM and IJH and discussed with JK. Where coding differed, areas were reconsidered until consensus was achieved. Participants’ ethnic identities are preceded by “BC” (black Caribbean) and “WB” (white British). We also reflected on the researcher’s field notes to supplement this analysis.


Thirty black Caribbean and 22 white British cancer patients were approached to participate in the study. Of these, four black Caribbean and three white British patients declined to participate. Reasons for not participating included wanting not to be disturbed or not to discuss their cancer. Table 1 presents the main characteristics of the participating patients. Five aspects of vulnerability emerged from the qualitative analysis and included five types of vulnerability: communicative, institutional, deferential, medical and social. We have presented text from the interview transcripts in order to illustrate issues. All participants’ names have been changed to protect their identity.

Table 1 Characteristics of cancer patients interviewed

Communicative vulnerability

Communicative vulnerability was seen in patients who were impaired in their ability to communicate because of distressing symptoms. We observed instances of communicative vulnerability associated with overt emotions, most frequently from cancer pain. This is illustrated by Elston, a 65-year-old black Caribbean man with prostate cancer, who reported intractable pain that was refractory to medication. While the interview process provided an important insight into understanding this symptom, it introduced a conflict. There were times during the interview when he was unable to speak, his emotions and tears impairing his communication:

Elston’s wife: He lost all movement in his legs.

Elston: Before I was having this back pain I was suffering nights and days. Oh.

JK to Elston: Are you in pain now, Mr C?

Elston: Yes I am. (Long pause while participant cries.) (BC21)

The participant then requested some tissues, regained his composure and proceeded again with the interview.

Institutional vulnerability

Institutional vulnerability refers to those who existed under the authority of others at the time of the study. In this study, this related to 16 of 45 interviews conducted among participants cared for in hospital. This setting raised a number of issues. First, compared with interviews that took place at home, participants were often in clothing different from the type of clothing they wore before their admission—for example, nightwear, or in some instances hospital gowns that provided very basic cover. The manner in which they were dressed, a consequence of the interview setting, may have created a perceived differential in the relationship between participants and the researcher. Other researchers21 refer to their appearance in research, but only in relation to the perceived formality by other professions, not among participants.

Second, the researcher’s field notes raised concerns about the physical position of participants interviewed. When they were well enough, interviews were conducted while participants sat in a chair by their bed. Other interviews, however, were conducted with bed-bound participants, who were frequently lying down. In these situations, the researcher sometimes had to sit on the side of their bed rather than in a chair, which created a height differential and a difference in perceived power.

Third, we noted the absence of a private venue to conduct some of the interviews, particularly among participants located in the larger hospital wards shared with patients and healthcare professionals. These participants were separated from others by a linen curtain that did not prevent all external noise interference or the possibility of being overheard. Given the frailty of some participants, moving them to a more private location, if available, was difficult.

Fourth, we observed that the concept of time had been suspended for a number of participants, particularly for those cared for in hospital. While the interviews may have represented a way of breaking up the monotony of their day, they could also be viewed as capitalising on participants’ loneliness, despite being surrounded by people who always appeared to be busy going about their everyday business. A small number of participants intimated they wanted to pursue a longer relationship with the study. For example, Franklyn, a 72-year-old black Caribbean man with prostate cancer, asked the researcher:

When are you coming back? (BC07)

Deferential vulnerability

Deferential vulnerability refers to those who were subject to the informal authority, or the independent interests of others. These informal interests can be based on ethnicity, social class, gender or the perceived inequalities between the researcher and participants.

There were other encounters that assumed a different kind of deferential vulnerability; in this case, the researcher was viewed as a healthcare professional and therefore not a party with whom information and views could always be shared. This is illustrated by Grace, a 35-year-old black Caribbean woman with gastrointestinal cancer, who stated that the interview had some of the characteristics of a recent encounter with a liaison psychiatrist at a local hospice:

JK to Grace: Would you like to tell me anything more, or do you have any questions?

Grace: No, I haven’t got any questions really, cos I don’t know, I mean what can you say? Er I mean I saw a psychiatrist. I understand the method behind, you know, the, the … how they interview you, but I don’t like it, it’s not for me. I don’t feel a 100% comfortable, because you’re in there for half an hour, an hour or whatever, and it’s just you talking all the time. It’s almost like with you but not … but worse (laughter) … You know, you’re just sitting there. (BC14)

In all interviews, the researcher attempted to put participants at their ease by allowing them to talk freely. However, this investment in gaining participants’ trust led to increased rapport and a third type of deferential vulnerability, the disclosure of intimate details in participants’ lives. The researcher observed this on several occasions when the direction of the qualitative interview, not bound by the constraints of other data collection methods, appeared to give some participants implicit permission to contribute information that did not initially appear to be contextually relevant to the study. This is illustrated by Jeanie, an older white British woman, who spoke in detail of the absence of practical and emotional support from her husband during her illness. To help the researcher comprehend her situation, she chose to share her history of physical and psychological marital abuse. This also included an unsuccessful suicide bid:

I mean he was always a heavy drinker. I got so fed up … I’ve tried to gas myself. And my daughter had gone off to school, but she’d forgot something. And she came back and she found that I’d turned the gas on and put a towel on the floor against the door. (WB30)

Last, in some interviews it became apparent that participants felt self-conscious or questioned their ability to offer valid accounts of their cancer. This is illustrated by Becky, a 34-year-old white British woman with an adenocarcinoma. While she provided a very eloquent account of her intractable pain, likening it to a devil living within her, there were times during the interview when she became uncertain about whether she was a credible witness:

In my head and how I feel? Er, I’ve always been a bit of a quiet person. I haven’t spoken about this before to anyone. Do I sound sick? (Laughs) Do I sound mad in the head? (WB34)

Interestingly, not all encounters with participants raised these concerns and instead could be viewed as liberating. For example, after being introduced to a black Caribbean woman here called Merlene, the researcher discussed with her when it would be convenient to conduct the interview. It was at this point that Merlene challenged the perceived hierarchy of the interview encounter, insisting she would only offer her time to the study if the interviewer returned with a homemade lunch. This request was taken very seriously. The next day, a lunch was prepared; on receiving it, she fulfilled her part of the “bargain” and provided a detailed account of her cancer experience. Feminist researchers2123 have used similar, though more deliberately reciprocal, strategies to develop rapport and decrease the power imbalance between research participants and themselves, one justification being that reciprocity brings better internal data validity.

Medical vulnerability

Medical vulnerability is a concern for potential participants who have serious and distressing medical conditions—for example, advanced cancer. Before interviews commenced, we explained our study, along with the potential benefits and risks associated with participation. We also emphasised that participants were free to terminate the interview at any point. However, no participants took advantage of this offer, despite all of them being medically vulnerable, given their poor prognosis and multiple symptoms. Many examples from the interview transcripts illustrate how compromised participants were by their symptoms. For example, a black Caribbean woman named June spoke openly about her hope that prayer would in some way relieve her breathlessness, a symptom that impacted on all aspects of her life. In her utter desperation to be free of her unabated distress, June petitioned the researcher:

I want you to help me breathe! Pray for me! To help me—especially for the shortness of breath. I don’t mind the pain, but the shortness of breath … terrible! (BC01)

Social vulnerability

Social vulnerability can refer to participants who are able to consent but who may be considered to belong to an undervalued social group. Typically, those who are poor or who belong to a BME community are considered to populate this group.24 An ongoing debate, since the 1950s, on inter-ethnic interviewing concerns the differences between the researcher and participants’ ethnic background that can potentially effect the veracity and accuracy of the information that participants share.25 26 It has also been suggested that ethnic differences can be threatening to participants.27 Strategies to redress these differences include ethnic matching, which, it is argued, promotes greater rapport in relation to the research topic.26 However, there are many practical as well as emotional problems with this solution. It assumes that ethnic identity is fixed and unaffected by gender, age, education and social class, among other factors.27 28 Second, feelings of commonality and difference may fluctuate during the course of the interview. Despite these perceived obstacles, we elicited very detailed accounts of cancer and its progression from participants, irrespective of their ethnic group. In one instance, however, we were made aware of a cultural gap that created a perceived social differential. This resulted from a question to explore what helped participants to live with their cancer and its progression. In many instances, black Caribbean participants volunteered that their religious faith and belief in God helped them and they spoke in detail about this resource. However, Rosemary, a woman with a genitourinary malignancy, took issue with this question about the nature of her relationship with God. She stated that only those who shared her faith could understand the relationship with her deity and the meaning of prayer. She also attempted to establish where the researcher stood in relation to religion. However, once she determined that she and the researcher did not populate the same “spiritual domain”, the interview was unable to explore this issue in more detail:

JK to Rosemary: What do you say in your prayers?

Rosemary: (Laughter) It is a very, very big question you are asking me! I go to him and ask him to heal me or to cure me or so-and-so. But I can’t explain it to you because you wouldn’t understand! You have to be in the faith to understand what I say to my Lord … I can’t answer you that question! Because you just have to live by faith. (BC09)


A key issue when considering carrying out any research is due consideration of whether a participant is vulnerable. Remarkably, however, vulnerability remains one of the most contentious yet least examined concepts in research ethics.12 16 29 While a number of issues highlighted by the study may exist in all research studies, there may be greater potential for their presence in qualitative research, which provides greater scope to reveal participants’ realities. Few bodies have adequately considered this issue in relation to this method of enquiry.9 Moreover, it may also seem surprising that it has received little attention in the palliative care literature, with its focus on patient and family-centred care.1 16

Regulations and policy documents that refer to vulnerability often do so within the context of populations or patient groups where patient autonomy and the ability to consent to research are compromised or may be absent.14 30 Researchers are required to protect the interests of such patients, since they may be susceptible to exploitation and harm. As a result, additional protections, both legal and in the form of guidelines, have been established to protect them. Recent advances in this area, with the development of a taxonomy of vulnerability, have made an important contribution to this field of research ethics. In general terms, it has been suggested that research among those considered to be vulnerable should not in itself imply that research is exploitive, and blanket exclusion from research may violate ethical principles governing fair subject selection.31 32 However, the current focus of the taxonomy, which is limited to clinical trials rather than other research methodologies,13 hinders progress in this important field.

Implications for researchers and research ethics committees

The secondary analysis of qualitative data and field notes from this study identifies tensions in the current understanding of vulnerability that warrants further consideration. We observed the following. First, we identify that the concept of vulnerability departs from the previously held notions that categorises people as either being vulnerable or not. Instead, we highlight the need to consider the relative importance of participants’ individual characteristics. In this study, participants were identified among five domains of vulnerability: communicative, institutional, deferential, medical and social. These are not exclusive categories and importantly researchers and research ethics committees should be aware that it is possible to populate more than one group. Second, as with current debates about capacity33 outlined within the 2005 Mental Capacity Act,34 we show that vulnerability cannot be viewed as being a “once-and-for-all” label. We observed some participants who were vulnerable in certain circumstances and not in others. Third, we observed many issues that were common to both ethnic groups. Vulnerability is therefore not a research issue exclusive to those from BME groups. However, we acknowledge, particularly with reference to social vulnerability, that the perception and experience of vulnerability among researchers and participants may be amplified, because BME groups may be socially disadvantaged on other levels. Fourth, the timing and location of interviews, and very importantly the individual personality of participants, governed reactions or responses to the research. An individual participant’s need for protection is therefore not governed only by inclusion within a vulnerability group. Instead, involvement in research should be viewed more flexibly; skilled researchers must reactively take account of the subtleties of the research process and the context in which data is collected.

We also observed that there were some participants who challenged previously held misconceptions about their vulnerability, a finding that has also been identified in relation to conducting research among bereaved children, a group where protections are deeply enshrined.35 Indeed, for these individuals the process of recruitment and conducting the qualitative interviews acted as a means of leveling the research playing field. On this basis, assumptions about the nature of power in the qualitative interviews process should be questioned. Indeed, research participants are arguably very powerful, since it is they who hold the data.36

Despite these challenges to the existing taxonomy, we believe that research among groups considered to be vulnerable is important and indeed possible. We therefore recommend that researchers may need to do more to prepare some participants for interviews of this nature. They should ensure that potential participants are clear about the inclusion of topic areas that may prove to be sensitive or distressing. While we acknowledge that this is potentially a difficult area to predict on all fronts, since the interviews were not experienced consistently among participants, we believe researchers should consider prompting participants to thoroughly consider the implications of thinking about and responding to potentially distressing questions. For example, they could say “Is it all right if we talk About X?” or “If you think you will find that upsetting, we can move onto to another topic.”

Although we attempted to provide reassurance and support to participants, there was also some evidence from the interviews that this sense of perceived comfort was beguiling. In a few instances, sensitive issues were shared that were not immediately understood as being relevant to the study aims. Questions have been raised as to whether the nature of informed consent should be presented as an ongoing process.19 However, we believe it is important to acknowledge that some participants may not wish to be continually asked for consent as events unfold during the course of an interview. Further, this approach may be time-consuming in patients who may be compromised by multiple symptoms. Ultimately, informed consent indicates trust and permits participants to hand over control without compromising their autonomy.37


Research ethics committees place important emphasis on protecting vulnerable populations,38 but little guidance is provided on how the concept should be made operational across all research paradigms or on the research skills required to conduct similar studies. Currently, studies are often categorised as including those who are vulnerable or those who are not. This delineation is important, but it fails to sufficiently protect those not considered to explicitly inhabit these groups. While qualitative research has the potential to unearth intrinsic and situational vulnerabilities through its research process, we believe that this research supports the principle of justice, providing a voice to previously under-researched groups.


The invaluable contribution of patients and staff at King’s College NHS Foundation Trust and Guy’s and St Thomas’ NHS Foundation Trust is gratefully acknowledged.

IJH with MM, JK and PE won funding for this study. JK collected the data supervised by IJH and MM. JK with IJH, MM, PE and PS analysed and interpreted the data. JK drafted the manuscript. All authors contributed to important intellectual revisions and approved the version to be published. JK is the guarantor.



  • Funding: The study was funded by Guy’s & St Thomas’ Charity.

  • Competing interests: None declared.

  • Ethics approval: The Research Ethics Committee, King’s College Hospital NHS Trust (LREC Protocol Number 01–204) and Guy’s and St. Thomas’ Research Ethics Committee (EC 00/018) approved this study.

  • Provenance and Peer review: Not commissioned; externally peer reviewed.

Other content recommended for you