Abstract
Research in palliative care and hospice populations is important for improving quality of care, quality of life, and provider understanding of individuals at the end of life. However, this research involves many potential challenges. This review seeks to inform and assist researchers targeting to design studies targeting hospice and palliative care patients by presenting a thorough review of the published literature. This review covers English-language articles published from 1990 through 2009 listed in the PsycInfo, Medline, or CINAHL research databases under relevant keywords. Articles on pediatric hospice were not included. Issues discussed include study design, informed consent, and recruitment for participants. Synthesized recommendations for researchers in these populations are presented.
Introduction
In 2007, an estimated 1.4 million people received hospice services, and 38.8% of all deaths that occurred in the United States were under hospice care. 1 With a growing number of individuals under hospice care, concerns about research with individuals at the end of life have become more prominent. As officials at the National Institutes of Health note, “While there is a growing body of research covering a wide range of issues, the research is, in many ways, still in its infancy in terms of rigorous testing and evaluation of models of care, in terms of patients and family outcomes, and in terms of resource utilization.” 2 However, research with patients at the end of life is replete with potential ethical and logistical pitfalls, the likes of which have frequently been discussed in the literature. 3 –10 The implicit question embedded in this discussion is, “How can researchers best proceed in this difficult area?” Unless and until this question is answered, the pace and quality of this research is likely to suffer. The goal of this review is to examine the literature regarding research design with hospice and palliative care populations which has arisen over the last 20 years and to synthesize this information into a functional set of guidelines to assist researchers in avoiding the possible pitfalls of conducting research in this area. This review will cover the areas of the ethicality of either including or excluding hospice and palliative care populations and issues of study design, informed consent, and recruitment.
Methods
This literature review was conducted with the search engines PsycInfo, Medline, and CINAHL. Different search terms were used to match the variable subject headings used by each engine. In PsycInfo, the terms “hospice,” “palliative care,” “experimental ethics,” “informed consent,” “methodology,” and “experimental subjects” were used. In Medline, the corresponding terms “hospices,” “palliative care,” “ethics, research,” “informed consent,” “research design,” and “research subjects” were included. In order to limit the amount of irrelevant information returned from the Medline search, “ethics, research” was used as a mandatory search term. In CINAHL, the corresponding terms “hospices,” “palliative care,” “ethics,” “consent,” “research methodology,” and “research subjects” were included. To limit the amount of information returned from all searches and to provide up-to-date information, results were limited to sources published between 1990 and 2009. Although it was not included as a search limit, articles focusing on pediatric hospice were not included. Fifty-two articles were returned, of which forty-six were relevant to the topics covered in this article and were reviewed.
Results
Ethicality of Including or Excluding This Population as Participants
Historically, there has been debate over whether it is ethical to include or exclude hospice and palliative care patients as research participants. Some researchers have argued that hospice and palliative care patients are too ill and too vulnerable to allow for valid and generalizable research, that there is not enough to learn about this population to justify its use as research participants, and that this population is too heterogeneous to provide meaningful results. 11 Researchers have also expressed concern that it is unethical to ask hospice and palliative care patients to participate in research because participation may limit the already restricted time and energy these individuals have and because these participants will not have the opportunity to benefit from the results of the research. 12
While these concerns are valid, current literature is overwhelmingly in favor of the ethicality of including this population. Researchers have pointed out several reasons why it may be unethical to exclude this population, including respect for patients’ autonomy. 3,11 –14 Not offering hospice and palliative care patients the opportunity to participate in research negates their right to make their own decisions about research participation and to have their voices heard. While this population will not likely benefit from the knowledge generated by the research, individuals may experience subjective benefits from their participation which would be unethical to deny. 13 These potential benefits will be discussed in greater detail in subsequent sections.
Terry et al 15 argue that reservations about including hospice and palliative care patients as research participants reflect a societal taboo against speaking about death and dying rather than valid ethical concerns. In regard to a National Institutes of Health consensus conference report which expressed concerns about the “decency or propriety” of asking these patients to participate in research, they rebutted, “These feelings of shame are, we suggest, those of the researchers. This is the language of taboo and we do not know of any other area of medical research that has, in recent years at least, provoked it. This is unfortunate because one reason to value the opportunity to participate in research commonly expressed by our patients was that it would confirm that they were still, and were regarded as, real people: not taboo.” 15 While it is important to be aware of the possible vulnerability of this population, it is equally important to respect their rights of autonomy and their humanity.
Hospice Patients’ Interest in Participating in Research
Researchers have found indications that patients at the end of life are indeed interested in participating in research through both formal research and their experiences conducting research with other aims in this population. 11,16,17 In one study, 46% of a sample of hospice patients reported being interested in research involving surveys or interviews and 45% reported being interested in research involving therapeutic interventions. 18
There are many reasons why hospice and palliative care patients are interested in participating in research. Many participants express willingness to participate in research in order to help future patients or the care providers. 6,8,13,15,16,19,20 As one study participant explained, “It would be a way to give something back now before I die, I would have done something good for the future.” 15 One study found a wide variety of motivations for participating in research reported by hospice patients, including helping the doctor or nurse (37%), feeling good about helping others (33%), maybe feeling better (28%), improving symptoms (24%), contributing to science (22%), having a sense of purpose (20%), adding meaning to life (19%), the possibility of being followed more closely by the patient’s doctor or nurse (19%), getting better care (14%), or having the opportunity to be social (11%). 18 The opportunity to be social may be especially important for patients who have suffered disability arising from their illness and who may feel isolated. 13 The opportunity to provide feedback regarding services received may also be an important motivator in patients receiving hospice or palliative care services. 13 Interestingly, Phipps et al 21 found that patients who are in more pain may be more likely to participate in research. They hypothesized this may be because patients hope that participation will mean their pain issues are better addressed. This finding has an important implication for recruitment within these populations, as researchers need to be prepared to refer patients with treatable pain problems that are not the focus of research to appropriate providers. 21
There are also many barriers to research participation in this population. Hospice patients were more likely than ambulatory senior citizens to see being too sick, having too little energy, and concern for creating caregiver burden as being barriers to research participation. 18 Phipps et al 21 recorded the reasons given by end-of-life patients for refusing to participate in research. Other priorities, such as spending time with family, were cited by 42% of refusing patients. Research-related concerns, such as having had a bad experience with research participation in the past, were cited by 36% of refusing patients, and health-related concerns, such as physical weakness or emotional distress, were cited by 28%. General disinterest was cited by 61%, making it the most commonly used reason for participation refusal. When consenting and refusing patients were compared, they found that consenters thought they had more to gain from participation than refusers. Additionally, consenters reported that aches and pains were more problematic for them as compared to refusing patients. Thus, while many hospice and palliative care patients demonstrate interest in participating in research, there are many motivations and barriers which come in to play when making the decision to participate.
Study Design
Research with hospice and palliative care populations is a complex and often delicate process. In order for the research to proceed smoothly, an appropriate and thoughtful study design is necessary. Areas of study design that are particularly relevant to hospice and palliative care research include allowing adequate preparation time, identifying the target population, choosing an appropriate research design, determining the sample size, and including diversity for a representative sample.
Preparation time
Because of the complexities involved, research with hospice patients can be expected to be time consuming well before data collection begins. It has been reported that almost 1 year was required to plan and obtain approval for a study on heart failure in older adults. 22 Researchers must take into account the time required to identify and carry out specific training or educational needs and plan site visits to clarify access, researcher role, and ethical conduct when planning a timeline for such projects. 22
Identifying the target population
Hospice and palliative care patients are a heterogeneous group, and the identification of a specific target population is an important, and sometimes problematic, aspect of study design with this population. 23 While hospice services have more concrete guidelines regarding life expectancy and exclude patients who plan to continue with curative treatments, patients in palliative care services may span a wide range of life expectancies and concurrent treatments. One third of patients receiving palliative care services are not under the care of palliative care specialists, meaning that using this as an inclusion criteria could exclude a large percentage of palliative care patients who may differ in some way from those under the care of specialists. 12 For these reasons and others, it may be impossible to identify every member of a palliative care population in order to select an unbiased sample, especially as lengthy screening procedures are contraindicated for this population and are often not economically feasible. 8
Calculating an adequate sample size
Due to the high rates of attrition that occur in hospice and palliative care research, research must begin with an adequate sample size to allow for attrition. 24 Additionally, researchers should consider the proportion of hospice and palliative care patients that will be excluded from the study due to inclusion/exclusion criteria. In one study, 64% of hospice patients were ruled out by the criteria that stated participants must be over the age of 19, English speaking, judged able to give informed consent by either the investigator or a hospice nurse, enrolled in hospice services for over a week, and not considered too sick (eg, actively dying or in a crisis situation) to participate by the hospice nurse or physician. 18 Outside of exclusion criteria, many potential participants may decline to participate for a variety of reasons. For example, one study had a return rate of 36% of surveys which were distributed to hospice patients eligible to participate in their study. 18 Researchers must plan for these factors when calculating their target sample size and planning their recruitment efforts.
Increasing diversity in research samples
Palliative care and hospice research often includes a paucity of minority participants. For example, only 1 of 14 participants in a focus group study in South Africa was of a minority ethnic background. 25 Researchers have expressed concern that more efforts needed to be made to include diversity in hospice and palliative care research. 16 There are many potential reasons for the lack of diversity in this field of research. As Kendall et al 16 state, “Given the recruitment difficulties within the majority population, it is not surprising that many studies fail to engage with people from these small populations.” Language barriers and the fact that a lower proportion of minority individuals than majority individuals seek palliative care services likely also play a role in this issue. Certain research practices may increase the likelihood that diverse individuals will choose to participate. As a focus group of researchers suggested, offering something (ie, information sessions, art-based activities, or social events) in return for participation, gaining approval from community leaders, and ensuring that the study design, research materials, and dissemination methods are culturally appropriate may serve to increase the amount of minority patients who participate in hospice and palliative care research. 16
Research Design
Randomized-controlled trials
While randomized-controlled trials (RCTs) represent the gold standard of research and are needed in palliative care and hospice studies, they present several challenges which may make them difficult to implement. 10,26 For instance, many palliative care interventions are well established and believed to be efficacious despite the lack of supporting empirical research, which can result in the withholding of a standard treatment from the control group to be considered unethical. 10 As Grande and Todd explain, “For a trial to be ethically justifiable there must be real uncertainty as to whether the new treatment is superior to no treatment, or existing treatments.” 27 This concept is known as “equipoise.” However, even when equipoise exists, a new treatment may be perceived as more desirable by professionals, patients, and/or families. In this situation, wait list controls are a possible solution; however, the limited life expectancy of palliative care and hospice patients limits the usefulness of this strategy. 27 Alternatively, researchers and care staff may establish participation in the RCT as an entry point to the intervention, thus framing trial randomization as the offer of an opportunity to access a currently unavailable treatment rather than as a means of limiting the availability of the treatment. 27 Cluster randomization, in which potential participants are grouped by a common characteristic (eg, region, hospice program, etc) and are assigned to various experimental conditions, may be a more acceptable alternative because which groups will be assigned to a treatment or control condition before individual participants are identified. 7,8,27 Educating professionals involved in research, as well as potential participations, about the concept of equipose may also decrease such challenges.
A great deal of discussion in recent literature has focused on placebo-controlled trials in this population. The general consensus is that all research participants in clinical trials should have access to the best available standard of care. 24,28 Placebo use may be justified if the participants who receive the placebo also receive the current standard of care, if the symptom being studied has no effective treatment, or if the participants who receive the placebo have access to breakthrough treatment should it becomes necessary. 24 Placebo-controlled trials require careful consideration of the study design to ensure the placebo group does not receive substandard care, and as additional time and effort may be required to educate medical staff and potential participants on this research methodology. 4 Terry et al 15 found that most palliative care patients were reluctant to participate in placebo-controlled trials. However, their reluctance was related to the assumption that researchers knew that the patients receiving the placebo would have worse outcomes than the patients receiving the treatment. It appears that trials comparing a new treatment to an established, active treatment are more likely to be acceptable to potential participants.
Qualitative research
Because of the difficulties inherent in RCTs, qualitative approaches are popular among hospice and palliative care populations. Qualitative research frequently utilizes interviews, participant observation, questionnaires, focus groups, case studies, and documentary analysis, 19 and qualitative methods are often preferred by end-of-life researchers because of their ability to allow participants to discuss issues that are currently important to them. 16 Interview methods offer a particularly adaptive, flexible methodology for gaining information about hospice and palliative care patients and their experiences while also supporting the needs and well-being of the participants. 29 Researchers must still be sensitive to possible ethical concerns. Coercion and perceived benefit to participants are particular areas of potential risk; researchers must ensure that the perceived benefit of participating in research does not cross the line into coercion. 30
A phenomenological paradigm (ie, a qualitative approach which searches to understand how people make sense of the everyday world) may be especially applicable to palliative care research. 31 In phenomenological research, the researcher attempts to gather data on the unique experiences of individuals and their interpretations of these experiences. Data collection is considered complete when no new material is generated by data collection, and cases which are atypical are emphasized rather than ignored. Seymour and Clark 31 explain, “Phenomenology is not a method … Rather it should be seen as a philosophical paradigm within which methods are chosen and questions framed.”
Case studies are another qualitative approach frequently used in hospice and palliative care settings. Case studies can be an especially diverse approach. As Walshe et al 32 explain, “Case studies can use either qualitative or quantitative methods, can be prospective or retrospective, can have an inductive or deductive approach to theory, can focus on one case or many, can describe, explain or evaluate” and are useful for studying “complex social situations or interventions, where multiple variables exist.” This methodology may be most useful when complex situations need to be addressed, context is central to the study, multiple perspectives need to be recognized, the study design needs to be flexible, researchers want the research to be directly congruent with a clinical practice approach, there is no strong theory to which to appeal, or other research methodologies could be difficult to conduct. 32 However, case studies must be used rigorously and only with appropriate reasoning for their use. 32
Focus groups allow for a group of individuals to provide information collaboratively and may also be useful with hospice and palliative care patients. However, some additional factors must be taken into consideration when using this approach. For example, facilitators must balance individual expression of thoughts and opinions with structured discussion of the topics of study, allowing for participants to breaks from the discussion and/or to share their personal stories if desired. 22
Mixed methods
Mixed methods, which combine qualitative and quantitative methodologies, can be used to collect and synthesize both types of data. 33 While little mixed methods research has been done in hospice and palliative care settings, it offers the potential to combine the benefits of both procedures and offer an in-depth set of data for analysis.
Survey research
Surveys are frequently used in palliative care and hospice research. 18,34 While surveys are most commonly completed by the patients, a proxy may be used to obtain information regarding patients who are unable to complete the surveys themselves. While this may be necessary in some cases, it may complicate the data collection process by raising the questions of whether to approach the patient or the proxy first regarding participation, when to approach the proxy (eg, it would be inappropriate to contact the proxy when the patient is rapidly declining), how to schedule data collection in a way that fits a caregiver’s schedule, and how to determine who the proxy should be. 35 However, it does offer an opportunity for retrospective survey research (ie, research regarding the patient that occurs after the patient’s death) as well as for collecting data on care at the time of death and data from patients who were too ill to participate in direct data collection. 12,35 When conducting retrospective survey research using proxies, it is essential to consider the timing of data collection, as collecting data too soon, when the proxy is not yet ready to discuss the deceased patient, or too late, when the proxy has forgotten important details, may both end with unsatisfactory results. 35
Collaborative approaches
Collaborative approaches, which involve patients working as co-researchers with the investigators, have been suggested and utilized for research with hospice and palliative care populations. 25,36 As these methods are very sensitive to the needs of the population and allow participants to influence the research in ways that result in study outcomes that are important to them, they may be especially applicable in this population. 25 As the result of a study in which patients and caregivers volunteered as co-researchers by leading focus groups of hospice and palliative care patients, Wright et al 25 suggested several recommendations for similar studies which include taking care to apply the same ethical guidelines that apply to participations to co-researchers, offering emotional support for all co-researchers if needed, and working to develop a collaborative dynamic in which the experienced researchers and co-researchers work together to ensure the quality of the research.
Williams et al 18 used a collaborative approach known as community-based participatory research (CBPR) to conduct research in a nursing facility for people with end-stage AIDS. In CBPR, community members with strong ties to the research population, rather than members of the research population itself, are included as co-researchers. In this particular case, staff at the facility were included throughout the research process in order to create a participation opportunity that was feasible and relevant to the community. This method can be particularly useful in populations who are vulnerable and marginalized as it invites the opportunity for those close to the population to advocate for both the researchers and the participants in a manner that creates a beneficial research experience for all parties. 18
Grounded theory
While it was not represented in the articles returned, grounded theory is a popular research methodology in medical sciences which focuses on building hypotheses from the data obtained rather than predefining hypotheses. Interested readers are referred to Stiel et al for further information on the use of grounded theory in palliative care populations. 37
Recruitment
Once the study design is solidified and the informed consent procedures are determined, study recruitment can begin. Due to the heterogeneity of hospice and palliative care populations, inclusion and exclusion criteria must be developed to recruit patients who are applicable for the study and who are likely to complete participation. Additionally, researchers must decide how they will access this population while minimizing inappropriate gatekeeping and how they will introduce their study to potential participants. Lastly, researchers should be aware of the factors affecting the interest of potential participants in research participation.
Inclusion/exclusion criteria
The development of inclusion and exclusion criteria is a necessary part of research design and recruitment. Appropriate criteria increase the likelihood that recruitment will target individuals who are relevant for, and likely to complete, study participation. Additionally, appropriate inclusion and exclusion criteria can lower the probability that participants from this sensitive population will be harmed by their participation. In a longitudinal study of the palliative care needs of older adults with heart failure, Barnes et al 38 assumed that individuals who were cognitively impaired, psychotic, or severely mentally ill were more likely than others to be unable to successfully complete study materials and to be alarmed by the issues raised in a palliative care study. Along similar lines, Williams et al 18 excluded potential participants who had been enrolled in hospice services for less than a week or who were considered by the hospice nurse of physician as too sick to participate in their study exploring hospice patients’ hypothetical interest in research participation.
Using inclusion/exclusion criteria that are too stringent can limit study recruitment. Mitchell and Abernethy 39 compared the methodologies of 2 studies of palliative care patients. The researchers of one study approached only patients who had been approved by the patients’ palliative care staff, general physician, and caregiver and who had a life expectancy of over a month. Of the 1137 potential participants screened, 52% were eligible for inclusion, and 51% of their sample died or withdrew before 1 month. In contrast, the researchers conducting the contrasting study used maximal inclusion criteria and minimal exclusion criteria to facilitate eligibility. Of the 1949 potential participants screened, 79% were eligible for inclusion. There is a fine balance to be reached in order to protect potential participants while simultaneously facilitating a large enough sample size to account for the high attrition and withdrawal rate that is likely in this population.
Life expectancy is an important factor to take into account for research with hospice and palliative care populations, especially in longitudinal studies. As Jordhøy et al 8 point out, “The main challenge, however, is to define eligibility criteria that can ensure patients’ entry at a time when survival will be long enough for the supposed effect both to occur and to be assessed.” Unfortunately, life expectancy is commonly overestimated. 8,39 In determining inclusion and exclusion criteria, it is important to allow for a reasonable length of time before follow-up data are collected 8,11 and to consider using prognostic factors (eg, performance status) and providing an allowance for overestimation of life expectancy. 8 In shorter term studies, including patients who have a short life expectancy may increase the likelihood of gatekeeping by ethics boards, medical professionals, and/or families. Hudson et al 14 recommend excluding patients that are very unwell or very close to death in studies which do not require patients with these characteristics in order to lessen the likelihood of gatekeeping problems.
Additionally, mental status has been identified as an important eligibility consideration for studies and clinical trials with palliative care patients due to its implications in informed consent and the ability for patients to participate in the tested interventions. 6,23 Dobratz 6 suggested using a baseline cognitive assessment tool to assess for cognitive impairments in all potential participants, regardless of diagnosis. However, excluding potential participants on the basis of mental status may exclude a large portion of the palliative care and hospice population and result in a sampling bias. Thus, it is important to consider how the use of mental status as an inclusion/exclusion criterion will affect the generalizability of the results. 23
Because hospice and palliative care patients constitute a diverse range of age, functional ability, cognitive status, and life expectancy, different exclusion criteria may be needed for different segments of this population. In order to increase flexibility and sensitivity to unforeseen problems in study recruitment, Hopkinson et al 17 left their inclusion and exclusion criteria open to adjustment during data collection as well as protocol development. However, this method can be detrimental in studies that use recruiters outside the research team. Bakitas et al 23 recommend that eligibility criteria be “clear, objective, and easily understood by recruiters and referring clinicians.” Additional thought must be put in to the wording of eligibility criteria due to the sensitive nature of end-of-life research. In regard to posters or flyers used to recruit participants, Bakitas et al 23 warn, “[Translating eligibility criteria into lay language] calls for creativity in recruiting seriously ill patients, some of whom will be unaware of their ‘eligibility’ as their condition may not have been presented to them by their physician as ‘serious’ or ‘life-limiting.’ Even when clinicians inform patients of their advanced illness status, patients may be in denial regarding the seriousness of their illness and unlikely to identify with an advertisement that is looking for ‘seriously ill, dying, or terminally ill’ patients.” Thus, it is important to consider who is being targeted by recruitment materials when considering how inclusion and exclusion criteria should be conveyed.
Access
Conducting research with hospice and palliative care patients may involve passing several levels of gatekeepers, including ethics boards, medical staff, hospice staff, family, and caregivers. In order for research to be successful, it is essential that researchers be able to obtain the support and understanding of all levels involved in gaining access to this population. Inappropriate gatekeeping has 2 main consequences: restricting patients’ autonomy and reducing research quality. 14 Gatekeepers can reduce the representativeness of the sample and the generalizability of the data by introducing sampling bias. Additionally, gatekeepers can misemploy the time and efforts of both researchers and participants. As Hudson et al 14 explain, “When patients or families participate in a study weakened by gatekeeping that prevents sound conclusions, their time and energies are misused. Given the limited life expectancy, it is imperative their contributions are worthwhile.” Thus, it is important to understand how and why gatekeeping can occur in hospice and palliative care research, as well as what can be done to prevent it.
Because hospice and palliative care patients are a vulnerable population, there has been contention about their appropriateness to participate in research. 14,30 As discussed previously, there has been historical debate about whether hospice patients should be excluded as research participants due to their end-of-life status and vulner-ability or included to protect their rights to autonomy. The paternalism of ethics boards was a common complaint in a focus group of end-of-life researchers. While the researchers agreed that potential studies should be scrutinized and their ethicality ensured, there was shared concern that ethics boards “[acted] as gatekeepers for perceived ‘vulnerable’ participants, rather than seeing them as individuals capable of making their own decisions.” 16
Medical and hospice staff are often used as referral sources in hospice and palliative care research in order to recruit individuals who are most likely to be able to participate successfully and without harm. 4,6,11,15 –18,21,25,26,36,38 While medical and hospice staff are invaluable for this purpose, problems can arise when these professionals block patients from participation for reasons other than the stated inclusion/exclusion criteria. As Barnes et al 38 argue, “Although clearly well intentioned, this responsibility for patient protection may cause sample bias, as only those deemed ‘well enough’ may be put forward as potential participants.” Other research have described encountering this problem when working with hospice and palliative care populations. 12,16 Barnes et al 38 recorded that 7 of the primary care practices from which patients were recruited removed 31 potential participants (2% of the participants produced by preliminary search) for reasons other than the exclusion criteria, evidencing the problem that gatekeeping can cause in unbiased study recruitment. This may become even more problematic in studies that involve a control or placebo group. As one researcher acknowledged, “Even the primary investigators were slow to recommend the study to their own patients because of the placebo control design.” 26
Buss and Arnold 4 reported the reservations of the nurses involved in recruiting hospice patients for a study measuring the safety and effectiveness of an antinausea drug. They found concern about the value of the study, the risk and burden to dying patients, and the ethics of a placebo trial were common worries. Because medical staff often play a key role in participant selection and recruitment, educating involved professionals on research-related issues, especially the ethical challenges of conducting research with hospice and palliative care patients, is an important way to decrease inappropriate exclusion of potential participants. 11,23 Assuring that recruitment does not unduly increase the responsibilities of medical and hospice staff can increase participant recruitment. 38 For example, in one study, referring clinicians were provided with flyers and pocket cards explaining the relevant inclusion/exclusion criteria, as well as scheduling meetings with each group to provide professionals with an opportunity to ask questions and raise concerns about the study. 23 The researchers found very few instances of gatekeeping when the clinicians had a clear understanding of the eligibility criteria.
While researchers do not necessarily need family or caregiver permission to invite palliative care or hospice patients to participate in research, they often act as gatekeepers in providing access to these patients. 17 For example, a caregiver could answer the researcher’s call at the time of recruitment and thus become the researcher’s initial contact. In one study, 13 potential participants were excluded because the caregiver declined on the patient’s behalf. 21 While caregivers may act in what they feel is the best interest of the patient, they often have a different perspective on the patients’ level of illness and ability to participate than the patients themselves. It has been found that caregivers of hospice patients perceived level of illness as a barrier to research participation more frequently than the patients themselves (49% vs 39%). 18 Also, caregivers were more concerned about emotional distress or pain as consequences of research participation (24% vs 13%). 18 Thus, while caregivers may feel appropriately protective of the individuals under their care, they may not make the same choices as patients for research participation.
Hudson et al 14 suggest several methods to decrease the likelihood of gatekeeping in hospice and palliative care research. For example, offering information sessions and brief face-to-face updates to recruiting medical professionals may increase their understanding of the study and of recruitment, as well as offering professionals the chance to bring up concerns about the study and to problem solve with the researchers. Also, excluding patients who are very ill or have a very limited life expectancy may increase study recruitment as gatekeeping is more likely to occur with patients who are perceived as very vulnerable. Lastly, excluding medical professionals who are involved in the direct care of the patient may decrease the possibility of medical professionals inappropriately protecting patients by not offering the opportunity for study participation or by discouraging patients from participating. It should be noted that this suggestion may be in opposition to patient preferences, as will be discussed below.
Preliminary contact
Medical professionals and researchers are often the first point of contact for potential participants in this population. Researchers investigating patients’ preference for being approached about end-of-life research have found that patients desire to be approached by the medical professionals most involved with their care rather than the researchers conducting the study. 15,18,36 This may be especially important when working with minority or stigmatized populations who may be more mistrustful of the researchers and able to communicate more openly with the medical professionals they already have a connection with. 36 While having medical professionals explain and introduce the study to potential participants increases the burden of participation on the professionals, it may also decrease gatekeeping and sampling bias by giving medical professionals more control over the timing of study introduction. In a study in which medical clinicians were responsible for referring patients to participate, researchers found that the referring clinicians rarely neglected to mention the study to eligible patients and would often adjust the timing of recruitment to account for factors such as distress about a terminal diagnosis. 23 However, there is concern that including medical professionals directly in the recruitment process may introduce more opportunity for gatekeeping and inconsistent recruitment approaches. 14
Informed Consent
It is essential that potential participants are given adequate informed consent, particularly given the vulnerable status of this population. Researchers must ensure that participation is voluntary and that potential participants are competent to give consent. Careful consideration should be given to the wording and method of delivery of consent.
Vulnerability and voluntariness
Hospice and palliative care populations are generally considered to be vulnerable research participants. 5,40 One reason for this is that patients at the end of life may choose to participate out of desperation 40,41 or loneliness 22 rather than an informed choice. 40,41 For these reasons, it is extremely important to ensure that research participation is voluntary. Patients’ relationships with the investigators and the involved institution may influence the decision to participate and must be understood in order to ensure that patients are not, and do not feel, coerced into participating. 5,42 Researchers must be aware that patients may feel pressure to participate in research projects in order to please their doctor and must actively help patients realize both that it is alright for them to decline participation and that refusal will not affect their relationship with their doctor or other staff. 41,43 Longer-term research carries additional challenges. Patients who are competent to consent at the time of informed consent may deteriorate to the point where they are no longer competent to consent or understand changes in their condition that warrant their withdrawal. 24 One potential solution to minimize this risk is that a friend or family member of the patient be present at the time of consent to ensure that there is no coercion. 43 Similarly, a neutral third party may be present to advocate for the patient or proxy consenter, minimize coercion, and promote the participant and/or proxy consenter’s understanding. 42
While it is important that researchers be aware of issues of coercion and voluntariness in informed consent, several researchers have presented evidence that patients are willing to refuse research participation. For instance, Rees and Hardy 44 reported a refusal rate of 32% in their trial of antimuscarinic drugs for treating the noisy breathing that occurs when dying patients are unable to clear secretions formed in their airways, often known as “death rattle”; 34 out of 107 palliative care patients approached declined to participate. Terry et al 15 reported that the palliative care patients who participated in their focus groups argued that “there is a ‘freedom’ in being close to death so that they felt they could say precisely what they wished and had nothing at all to lose by voicing their own opinion.” As one patient explained, “When you go to the emergency they ask you things, they don’t think you have stopped thinking. Why would you not know what to say just because you are dying in the hospice [?].” In light of concerns that patients may participate out of desperation to extend their lives, Terry et al 15 reported that the majority of participants stated they would only participate in research if it did not have the possibility of extending their life.
As several researchers have pointed out, the potential for coercion into research with hospice and palliative care patients is not unique to this population and many strategies, such as emphasizing the voluntary nature of participation and ensuring that researchers are sensitive to subtle forms of pressure, may be used to overcome these challenges. 41,45 Fine argues that researchers should “not equate vulnerability with involuntariness” and that “dying should not be equated with coercion.” 45 Instead, both patients and research should be evaluated in light of their context.
Competence
Once vulnerability and voluntariness are addressed, researchers must establish that patients are competent to consent to participate in research. In terms of informed consent, competence concerns “the ability to make a decision.” 42 While it has no universally agreed upon definition, it is commonly noted to consist of the capacity to understand information, to appreciate one’s situation and its consequences, to consider information pragmatically in light of one’s values, and to make a decision. 42 Because this population is diverse with regard to the ability to make competent and informed decisions, patients must be individually evaluated for competence to give informed consent. 5,6,24,41,45 However, researchers should not assume palliative care and hospice patients are incapable of informed consent or that this is a rare phenomena. Gysels et al 13 noted that most of the palliative care patients they interviewed were capable of making the decision to participate in an interview, as well as advocating for the conditions under which they wished to participate. Also, competence can fluctuate vastly over time, making it an ambiguous area to assess. 6,43 Researchers must be aware that known cognitive impairment does not rule out decision-making capacity. However, the presence of cognitive impairment may make the process of informed consent more difficult or impossible. 23,41
When evaluating competence to consent, consideration must be given to the potential risks and benefits offered by the study. 5,12,45 Fine 45 suggests that when minimal risks are posed by the study, a formal capacity assessment should not be required. When the study poses greater than minimal risks, Fine suggests, formal capacity assessment should be considered when potential benefits are offered and should be required when there are no potential benefits. An assessment, such as the MacArthur Competency Assessment Tool for Clinical Research (MacCAT-CR), may be used for assessing competence to consent. 45 In such cases, investigators should determine what will be cutoff point for competence before administering this assessment to potential participants. 45 The Mini Mental Status Examination (MMSE) has also been commonly used in this field of research 23,36 as well as the Mini Mental Status Questionnaire (MMSQ). 3 Bruera 3 has noted that patients should not be excluded from research due to incompletion of the MMSQ. Nonetheless, researchers should be concerned that such participants may not be capable of completely understanding the study. One downfall of the MMSQ is that it lacks specificity and may result in type I error. 46 Many researchers have chosen to use informal assessment of competence by a researcher or medical professional rather than a formal assessment. 6,18 Despite the prevalence of this option, it should be noted that one researcher who used these methods commented that use of a formal assessment of cognitive function would have been helpful in her study and recommended it for future use. 6
Mental status is valuable as inclusion/exclusion criteria for ensuring that competent consent is obtained, but it may result in issues with generalizability. Studies which only include patients who possess a normal mental status may not be generalizable to the palliative care and hospice population as a whole, in which patients are often sedated, delirious, or experiencing other declines in mental functioning. 23 In light of this possible complication, Bakitas et al 23 recommend that researchers consider whether a normal mental status is necessary for the patient to participate in and possibly benefit from the intervention or if it is solely necessary for the research procedures (eg, informed consent).
Proxy consent
If patients are not competent to give consent, a legally authorized proxy may be able to consent on their behalf. 23,24,47 However, “dual consent,” in which both the participant and the proxy provide consent, is recommended to ensure that the participants receive as much information as possible even if they cannot provide consent independently. 24 Before using proxy consent in a study, researchers must first understand their state laws regarding proxy consent for research, as some states may restrict or prohibit proxy consent. 24 Also, researchers should proceed with caution as the process of surrogate decision making still requires further exploration and the families of palliative care and hospice patients may still comprise a vulnerable population as they may feel pressured to do everything they can to increase their loved ones’ well-being or to extend their lives. 42
Process consent
Process consent is one alternative to the traditional, one-time method of obtaining informed consent. In process consent, the researcher regularly asks the participants whether they still wish to participate. 12 Process consent is recommended when a study involves several interviews or observations over a period of time. 16 It is designed to “acknowledge the dynamic and emerging nature of the research design, include the initial input and suggestions of the participants in the study, and negotiate further input regarding changes over time.” 48 Hudson et al 14 caution against gathering informed consent at every data collection point due to the danger of overloading the participant. Rather, they recommend that research staff remain constantly prepared to initiate discussion regarding consent, especially when they sense hesitance in the participant.
Advance consent
In situations where the potential participants will be unable to provide consent at the time they become eligible to participate, advance consent allows researchers to obtain consent to collect data if they should develop target symptoms. 5,24,44 For example, in a trial of antimuscarinic medications intended to treat the noisy breathing that occurs at the end of life, researchers asked patients whether they would be willing to participate should they develop noisy breathing at the end of life, at which time they would not be expected to be able to provide informed consent. 44 On each subsequent hospital admission, the participant was asked to re-sign the informed consent. If they were no longer competent to consent, the patient’s relative or caregiver was asked whether there was any reason the patient might have changed his or her mind about the decision to participate. If no indication was stated, the previously signed consent was considered to remain valid. While advance consent is necessary for studies in which the participant is expected to lose capacity to consent or in which the capacity to consent is intermittent, it should be used only when necessary and should be obtained as close to the time of expected study enrollment as is possible. 24 Advance consent may be especially practical when the study poses greater than minimal risk and/or no potential benefits. 24
Cluster consent
Another alternative method for obtaining informed consent is cluster consent, which involves obtaining consent from an entire group of patients with a “cluster guardian” responsible for providing consent as a proxy for the group and a “cluster gatekeeper” responsible for advocating and taking responsibility for individual participants. 7 Researchers comparing cluster consent with randomized consent methods found that cluster consent resulted in a larger number of participants recruited and less burden for medical staff, as consent was given at a unit, rather than individual, level. 7 However, this method is not without disadvantages. Researchers warn that the randomization of groups as opposed to individuals reduces statistical power.
Wording and method of delivery
The wording and delivery of informed consent can affect patient understanding and ability to provide consent. In order to lessen the risk of physical fatigue, mental fatigue, or emotional distress in the patient or family, the consent form should be concise and simply worded and the patient or proxy should be given ample time to read the consent form fully and discuss it with other family members or medical staff if desired. 3,42
Researchers should also be aware that terms such as “palliative care,” “terminally ill,” and “end of life” may be detrimental to recruitment as not all patients have been given a clear, terminal prognosis by their doctors and may not identify as terminally ill or dying. 29 While some researchers have expressed concern that avoiding such terminology may prevent potential participants from having all the information needed to provide informed consent, 38 others argue that this information is unnecessary and should not be considered deceptive. 5 Plant 29 suggests researchers devote time to establishing understanding the potential participants’ view of their condition before attempting to obtain consent. Additionally, wording that suggests a doctor–patient relationship, such as “doctor,” “patient,” and “drug” as opposed to “investigator,” “subject,” and “study agent,” may suggest that patients are consenting to treatment rather than research. 42 Researchers should similarly avoid the use of ambiguous expressions of probability such as “likely,” “probable,” “expected,” or “moderate.” 42
Terry et al 15 reported the palliative care patients in their sample strongly preferred oral information to written information. As one participant pointed out, “Do you see anyone writing here? Reading things is so hard.” However, potential participants should be given some form of back-up information regarding consent, such as leaflets, videos, or information sheets, because adjustment to a terminal diagnosis may result in poor information retention. 43 While explaining consent orally provides the option for participants to engage in a dialogue about the research process, and thus gain better understanding, the researcher’s nonverbal cues, such as tone of voice and behavior, may bias the information being provided. 42 More research is needed to better understand how patients wish to be approached regarding research participation and informed consent. 11
Exceptions
In some situations, it may be impossible to obtain consent in the population or to conduct the research in a population that is capable of providing consent. In these cases, researchers may seek approval from their ethics board to forgo the informed consent process. Although there is precedence for this, there is controversy concerning whether or not this method is ethically sound. 47 This approach may be acceptable if patients are unable to provide their own consent, seeking proxy consent would result in delay that would be detrimental to the treatment’s efficacy, participation involves minimal risks that are comparable to the risks present in standard treatment, and the research cannot be done with a population that can be expected to give consent, such as healthy volunteers. 47
Discussion
In conducting research with hospice and palliative care patients, researchers face many challenges in the areas of study design, informed consent, recruitment, risks and benefits, ethical issues, and methodological issues. The following recommendations include many previously suggested in the literature and others based on the information previously reviewed. In order to minimize these challenges and increase the quality of future research, researchers are recommended to:
allow plentiful time for planning and study design 22 ;
plan for high attrition rates. 8,11 Attrition rates of 30% to 50% have been found in longitudinal studies 3,39 and 15% in 1-time studies, 6 with rate increasing over the length of the study, 11 have been reported in the literature;
seek to increase diversity in samples by using culturally appropriate materials and involving the cultural community in recruitment 16 ;
consider implementing a “silent opt out,” in which patients can decline to participate through inaction, to minimize coercion;
use dual consent when proxy consent is required 24 ;
deliver informed consent orally, with simple, straightforward written materials for participants to keep for reference 43 ;
use oral, versus written, information while providing a concise written summary for future reference 15,43 ;
understand the specific target population’s perception of their life expectancy or terminal status and word recruitment materials accordingly to minimize distress and increase participant recruitment 29 ;
include life expectancy as an inclusion/exclusion criteria to ensure that participants will be able to complete study participation and decrease gatekeeping 8,11,14,39 ;
Educate other involved professionals on research-related issues to minimize gatekeeping and maximize study recruitment 23 ;
understand participants’ preference to be approached by the medical professionals most involved in their care 15,18,36 ;
train data collectors to be alert to signs of adverse effects. 6,19 Researchers involved in a focus group warned of the potential harm that can be caused by data collectors who do not pay attention to the emotional state of the participant or who do not end data collection with hospice and palliative care patients if a participant becomes distressed 16 ;
use retrospective or routine assessment data whenever possible to minimize the burden to patients 49 ; and
conduct pilot tests if there is concern about the burden or feasibility of the planned study. 38
In conclusion, the aim of this review was to outline a set of recommendations for researchers working with hospice and palliative care populations by synthesizing the literature regarding research design with hospice and palliative care populations which has arisen over the last 20 years. This literature includes areas of the ethicality of either including or excluding hospice and palliative care populations and issues of study design, informed consent, and recruitment. By utilizing the knowledge base present in the current literature, researchers can avoid common pitfalls in this area of research and decrease the potential for problems to arise. Appropriate, thoughtful study design is essential in order to balance the need for continued research and expansion of knowledge in this area with the necessary care and precautions warranted by this vulnerable and valuable population.
Footnotes
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.