In this systematic review of qualitative studies we aimed to uncover how the volunteer role in end-of-life care is understood by those who are closely involved with volunteering, for example, by being a volunteer or receiving volunteer care. Our findings suggest that a volunteer role exists which is largely distinct from that of paid staff, and is broadly social in nature. The findings will now be discussed including some areas of contradiction.
Distinct role versus the professional-like aspects of the role
Although the volunteer role seemed to be distinct from that of paid staff, the synthesis also found that volunteers in some studies appeared to take on a quasi-professional role. It may be that volunteers need to feel that their role is important in order to justify the time and effort they put into it and, therefore, they are keen for it to be seen as a distinct role. However, they may also feel that patients and families (and maybe the researchers) will take them seriously only if they appear to be professional-like. The concept of professionalism involves lay people putting their trust in professionals who in return maintain confidentiality
. Therefore, ‘professionalising’ the volunteer role in this way may be an important part of how patients and volunteers ‘negotiate’ their interaction. Volunteers’ taking on a professional-like role was more common in studies where volunteers were respondents, contrasting with studies where family members were respondents where volunteers were more likely to be cast in a surrogate family role. There was a similar contradiction between the description of the role as ‘complementary not substitute’ and the finding that volunteers ‘do what is needed’ (seen in the characteristic ‘flexibility’). This could imply that volunteers perform tasks which staff do not have time for, which suggests that the volunteer role may not really be a distinct role, but a way of stretching available funding. It is not clear from the included studies where participants talked about ‘doing what is needed’ what the tasks involved were, and whether they were part of paid staff’s roles or not. Further primary research, such as an ethnographic study following volunteers in relevant settings, could usefully explore all these anomalies.
Types of role
The two types of volunteer role which emerged from the synthesis, independent roles and surrogate family roles, show how the volunteer role may adapt to the care setting in which it is played out. The included studies were undertaken in different settings, including patients’ homes and inpatient care facilities. Surrogate family roles arose most frequently in studies of volunteers in homecare or mixed (i.e. both homecare and inpatient) settings. This may be because in inpatient facilities volunteers appear to patients and families to be more like staff in a setting where paid staff were also working. In the home, where patients and families are in a familiar environment, volunteers may feel most comfortable taking on the role of family member or friend. This may be particularly the case where there has been contact over a longer period of time or where only one volunteer has been involved with a family as can be the case when volunteers provide care in patients’ homes. On the other hand, patients and families using inpatient facilities are likely to encounter relatively large numbers of staff and volunteers. Again, this contrast between roles could be usefully explored by further primary research, such as a qualitative study comparing relatives’ experiences of volunteers in different settings.
The dominant characteristic of the volunteer role emerging from the synthesis was its social nature, in particular that respondents characterised the role in social terms rather than in terms of the tasks they undertake. The perceived social aspect of volunteering in a hospice setting has been found to be a motivating factor for volunteers
[39, 40]. The social cluster of themes was found in studies interviewing volunteers and in those interviewing family members. It was also found in studies in both home- and hospice-based care, indicating that this is a fundamental aspect of how the volunteer role is understood. Forming relationships with patients and their families are an important and sometimes neglected aspect in palliative care
, for example, palliative care patients receiving volunteer visits have been shown to live for longer than those not receiving visits
The synthesis also identified ambiguity in how the volunteer role is understood. On one hand the role was attractive to volunteers because of its flexibility and informality, qualities which were viewed positively by volunteers (although they are not defined in the studies, so it is unclear to what volunteers were specifically referring). On the other hand, volunteers found it stressful when the tasks to be undertaken were not clearly established, and when this was the case, the role was described as ambiguous. Since a largely quantitative study reports low levels of role ambiguity and role conflict among 97 hospice volunteers in Australia
 this aspect of our findings would also be worthy of further primary research, such as a qualitative study focusing on volunteers’ perceptions of this aspect of the role. Given the increased formalisation of volunteer roles resulting, for example, from legislation around health and safety
, these aspects may effect volunteers’ motivation and satisfaction with the role
The synthesis highlighted the relationship between volunteers and paid staff as important in shaping the volunteer role. This included the way staff ‘control’ the role, for example, in some studies through restricting access to patient information. It is clearly expected that staff direct the volunteer role, for example, designating the tasks volunteers undertake, since they have overall responsibility for patient care. However, a quotation from a volunteer "in one study I expect to be on the periphery" suggests that this volunteer sees their role as subordinate to that of staff (although the quotation begs the question, periphery of what?). The synthesis also identified issues of staff insecurity, as well as examples where members of staff were supportive of volunteers. All of these findings suggest that power relationships may shape the volunteer role
, p.154, and further primary research exploring this would be worthwhile, for example, a qualitative study focused on the relationship between volunteers and staff.
In a previous survey, we found that some people volunteered their professional skills, most commonly hairdressers and beauty therapists
. In the studies included in the present review, authors did not differentiate between volunteers offering professional skills and others, implying that none of the volunteers were offering professional skills, although this is not explicitly stated. A study specifically of volunteers offering professional skills and comparing their experiences with those of other volunteers might be of value to volunteer coordinators to help manage both groups more effectively.
Implications of the findings
Understanding how a role is conceived by those involved (in the present review mostly volunteers, but also some family members and volunteer managers/coordinators) will help planners and managers of palliative care services to make the most of volunteers’ contribution. For example, our findings could help volunteer coordinators present the role as realistically as possible to prospective volunteers. In particular, understanding that the role is not simply a set of tasks to be undertaken and that the role is strongly social in nature is important. Also, an appreciation of how the role is understood will be helpful in training paid staff to work alongside volunteers as effectively as possible. A specific issue which could be addressed includes the contrast between ambiguity in the role, which was stressful for volunteers, and flexibility, which volunteers liked (although this may be being eroded by increasing formalisation of the role due to the introduction of management policies, such as job descriptions
). Also, it would be useful to help volunteers understand and deal with any ambiguity they may experience in their role, as well as reducing any ambiguity that may exist by training staff to understand what volunteers can contribute to the patient and family experience of care.
The quality assessment process
Three studies contributed very little to the analysis
[3, 5, 33]. One of these scored poorly on both quality assessments
, and one other
, scored poorly on one set of criteria
 and lower than most of the other studies on the other
. Retaining them in the analysis did not appear to compromise the synthesis since the weaker studies merely contributed to fewer themes rather than changing the themes. The third study
, although scoring well on both quality assessments, differed in other ways from the other studies. It was the only study not undertaken in a developed country, and volunteers in this study undertook different tasks compared with those in the other studies (for example, providing clinical care). Its focus was also more on what volunteers do rather than on what it was like either to be a volunteer or to receive volunteer care. Furthermore, although this study’s authors collected qualitative data, their analyses which used content analysis, were less inductive than most of the techniques used in the other studies. It should also be noted that the CASP criteria rely on quality of reporting which, in turn, is affected by publisher word limits, so not meeting these criteria may have been because of inadequate reporting rather than because these issues were not considered by study authors.
Turning to the second quality assessment tool used
[22, 26, 27], which focuses in addition on the usefulness of a study’s findings, three studies scored two points out of a maximum of four for ‘overall usefulness’
[5, 32, 33] which included an assessment of the extent to which the study privileged the role of volunteers. In one of these the focus was on volunteer stress and burnout
. In another, which is a comparison of paid and unpaid workers, the findings included little material suitable for synthesis
. In the third study the material presented focuses more on descriptive material about families’ experiences of the whole care process rather than specifically on experience of volunteer care
. This demonstrates the utility of this quality assessment tool and also reflects the experiences of other researchers undertaking qualitative synthesis
The inclusion of one study
 also raises the question of whether our inclusion criteria should have excluded studies from developing countries, particularly given that the majority of available studies were from the developed world. Given that the synthesis method tends to privilege themes which recur within the available data, any study which stands out from the others is likely not to have much influence on the final themes reported. Although there may be a universal volunteer role common across cultures and healthcare systems, we consider that removing this study would not have benefited our research. An important reason why this study contributed so little was that the data collected by the authors was not analysed using the sophisticated methods employed in other studies rather than simply because it was from a developing country.
In addition, we excluded papers where the primary focus was on topics such as volunteer motivations, stress and burnout, and training, unless findings focused on care given by volunteers in patient/family-facing roles (Table
1). We appreciate that this may have led to our omitting relevant data but we needed to set realistic inclusion criteria because of the time-consuming nature of thematic synthesis. We consider that data saturation was reached since similar themes recurred in the included studies, so additional studies would not necessarily have added relevant new themes
We also included studies with a range of respondents, volunteers, family members and volunteer co-ordinators. This allowed us to gain as wide an understanding as possible of how the volunteer role is conceived, rather than just one respondent group’s view, and whether some universal understanding of the role exists. It also acted as a triangulation method. The included studies were also from a variety of countries, just over half from North America and a third from Europe, as well as the one from Africa.
Another strength of our review was that we were able to explore the difference that care setting made on our findings, in particular the finding that our quasi-professional theme came mostly from studies based in inpatient care facilities and the surrogate family member theme from studies in home-based care.
The included studies employed a range of analysis methods which raises the issue of whether it is appropriate to combine data arising from different epistemological approaches. Little research has been undertaken on this issue
 which largely remains theoretical. We could discern little difference between studies based on analysis method, except that those studies which did not state their analysis method contributed very little to the synthesis
We analysed the questions asked by the researchers in our included studies when collecting their primary data, so far as the information in the published papers would allow. These show a wide range of topics covered, some directly relevant to our research question (for example, ‘Participants had the opportunity to share their….understanding of their role…’
, p.17, and less directly relevant (for example, ‘Opening question, "What made you become a volunteer here?". Then the volunteers were asked to describe personal experiences from their voluntary work.’
). This demonstrates how a systematic review of qualitative findings can go beyond those findings, and even the original intentions, of individual studies to answer a separate research question.
Thematic synthesis method
An important issue concerns the method of synthesis chosen, thematic synthesis, and how this suited our research question. Our question could be seen as being constructivist in nature, being about how the role of volunteer is understood. This could be construed as similar to how a role is constructed by relevant stakeholders, which is an idea based in an idealist epistemology (where reality is socially constructed
). However, our method draws on realist epistemology, where reality is considered knowable, albeit imperfectly
[17, 18]. However, methods for qualitative synthesis are underdeveloped and which method to apply in particular situations may become clearer
In using thematic synthesis, we followed the method set out by Harden and Thomas
. However, this has previously been used mainly with relatively broad research questions that look at participant views on loosely defined topics (for example, children’s views on obesity
 and barriers to, and facilitators of, healthy eating in children
). We, on the other hand, had a research question which was simultaneously less likely to be literally answerable from existing studies (from our searches of the wider literature on volunteers in palliative care settings we were not expecting to find studies which had focused specifically on how the volunteer role was understood) and was relatively tightly defined. Therefore, we kept our research question in mind throughout the synthesis process. This meant, for example, that when coding the studies for themes, we did not code literally every sentence of the findings or results sections as Thomas and Harden did
. Instead we coded only those portions of the text which were directly relevant to our research question, omitting sections which were clearly off-topic, such as some text on volunteer motivation
In this way, the second stage of the thematic synthesis method as outlined by Thomas and Harden
 was fulfilled, that is, developing descriptive themes. Thomas and Harden then describe using the descriptive themes to answer review questions by inferring answers from the themes, thus generating analytical themes. However, because of our constant mindfulness of our review question, we considered that we had effectively answered our question by the end of the second stage. We had largely focused on our review question when coding, for example, by omitting study findings which were clearly not relevant. Also, when comparing the codes generated in the initial coding and arranging them in a hierarchy we kept our research question in mind, asking ourselves, ‘What does this code tell us about the volunteer role?’. We believe that we have ‘gone beyond’ the original studies to produce a synthesis which does far more than merely describe the findings of the included studies.