Palliative care has been progressively adopted by American cancer centers; however, referrals to palliative care continue to occur late in the trajectory of illness. It was hypothesized that the perceived association between the name palliative care and hospice was a barrier to early patients' referral. The objectives of this study were to determine the perception of the impact of the name palliative care compared with supportive care on patient referral and to determine whether there was an association between demographic factors and the perceptions of the 2 names by medical oncologists and their midlevel providers (advance practice nurses and physician assistants) at a comprehensive cancer center.
A survey was conducted among a random sample of 100 medical oncologists and 100 midlevel providers from The University of Texas M. D. Anderson Cancer Center. Information was collected on demographics, previous experience in palliative care, and attitudes and beliefs toward the impact of the name palliative care compared with supportive care on patient referral.
A total of 140 of 200 (70%) participants responded (74 midlevel providers and 66 medical oncologists). Median age was 43 years (range, 34.5-50 years), and there were 83 (60%) women. Midlevel providers and medical oncologists generally agreed in their responses to most of the items. More participants preferred the name supportive care (80, 57%) compared with palliative care (27, 19% P < .0001). Medical oncologists and midlevel providers stated increased likelihood to refer patients on active primary (79 vs 45%, P < .0001) and advanced cancer (89 vs 69%, P < .0001) treatments to a service named supportive care. The name palliative care compared with supportive care was perceived more frequently by medical oncologists and midlevel providers as a barrier to referral (23 vs 6% P < .0001), decreasing hope (44 vs 11% P < .0001), and causing distress (33 vs 3% P < .0001) in patients and families. There were no significant associations among the perception of the 2 names and age (P = .82), sex (P = .35), or prior training in palliative care (P > .99).
Palliative care focuses on the assessment and management of physical and psychosocial distress of patients with advanced cancer and on family support and advance care planning.1-3 In the United States, palliative care developed during the late 1980s and is currently a recognized subspecialty by the American Board of Medical Specialties (ABMS).4 Clinical palliative medicine programs have developed rapidly in cancer centers, academic hospitals, and community hospitals in the United States during the past 10 years.5 At the University of Texas M. D. Anderson Cancer Center (UT MDACC), the palliative care program was established in 1999. Since its establishment, the percentage of patients who access palliative care before death has steadily increased.6 However, access has not been homogeneous for all patients with advanced cancer. Data from UT MDACC7 and from other centers8, 9 suggest that patients with hematological malignancies have an overall lower rate of palliative care access and a higher rate of both inpatient hospital mortality10, 11 and intensive care unit mortality.7 In addition, referrals to palliative care programs occur late in the trajectory of the illness of patients with cancer.12, 13 Even when the number of palliative care referrals increased over time, there is no clear trend toward earlier referrals.14
The World Health Organization indicates that “palliative care is applicable early in the course of illness, in conjunction with other therapies that are intended to prolong life, such as chemotherapy or radiation therapy, and includes those investigations needed to better understand and manage distressing clinical complications”.15 Early referrals to palliative care are of great importance in allowing clinical teams to provide early relief of physical and psychosocial distress, establish a relationship with patients and families, and discuss issues of discharge planning that might result in preventing both inpatient and intensive care unit deaths.16 It is unclear why referrals to palliative care occur late in the trajectory of illness. Focus groups conducted by our team among referring specialists as well as 2 recent studies17, 18 suggest that the term palliative care might act as a deterrent to early referrals by medical oncologists. In a convenience sample of 20 of each of the groups advanced cancer patients, their family members, physicians, and nurses in palliative care and acute care cancer settings of 5 regional cancer institutes in Japan, Miyashita et al17 found that the negative image of palliative care was the most important barrier to early patient referral to an inpatient palliative care unit as perceived by the patients, families, and staff. When comparing the terms palliative and supportive care, Morstad et al18 found that the term palliative care evoked more negative emotions and was less favored by staff and parents of children with cancer. These 2 available studies have been conducted among Japanese general physicians17 and American pediatricians.18 There have been no studies conducted among American medical oncologists and midlevel providers.
The National Cancer Institute (NCI) defines supportive care as care given to improve the quality of life of patients who have a serious or life-threatening disease. The goal of supportive care is to prevent or treat as early as possible the symptoms of the disease; side effects caused by treatment of the disease; and psychological, social, and spiritual problems related to the disease or its treatment.19 The background of physicians and nurses running a clinical supportive care program is similar to that of the palliative care programs.
We hypothesized that the name palliative care compared with the name supportive care has a negative impact on the timeliness of palliative care referrals. The primary objective of this preliminary study was to determine the perception of the impact of the name palliative care compared with supportive care on patient referral as indicated by referrals at specific times in the cancer illness trajectory. The secondary objectives of this study were to determine whether there was an association between demographic factors and the perceptions of the 2 names and to characterize the perceived meanings and implications of the 2 names by medical oncologists and midlevel providers in a comprehensive cancer center.
MATERIALS AND METHODS
We conducted an anonymous Web-based survey among a random sample of 100 full-time clinical faculty and 100 midlevel providers including advanced practice nurses and physician assistants in the Division of Cancer Medicine, UT MDACC. The random sample was generated by our biostatistician using SAS data step and Ranuni function with seed 1 from base lists of 178 oncologists and 150 midlevel providers (advance practice nurses and physician assistants) from the Division of Cancer Medicine. The clinical faculty in Cancer Medicine has been chosen because the vast majority of palliative care referrals for cancer patients in our center are made by practicing medical oncologists.20 In addition, the midlevel providers frequently participate in the decision to refer the patients to the palliative care service. Subjects were excluded if they do not spend at least 1 day a week on direct hands-on patient care. An e-mail was sent to all potential respondents that explained the nature of the study and directed them to the Web page of the survey on which a consent for participation was included. For those who did not respond within 10 days, an e-mail reminder was sent to all participants followed by a telephone call to those who did not respond within 2 weeks of the second e-mail. The study was approved by our institutional review board.
The instrument used for this study was an ad hoc survey questionnaire designed by the investigators after extensive discussion and a comprehensive literature review of the challenges and opportunities facing palliative care services. The survey underwent content validation by 2 experts who were not directly involved with the study.
The questionnaire included questions about demographics of the respondents (age, sex, number of years in practice, ethnicity, and service area within medical oncology), previous experience and training in palliative care, frequency of caring for dying patients, frequency of use of the palliative care service, preference of use of the names palliative versus supportive care when referring patients, general perceptions of the 2 names, and attitudes and beliefs toward the impact of the name palliative care compared with supportive care on the timeliness of patient referral as determined by use of the services for patients at different stages of illness. We determined previous training in palliative care based on the following categories: 1) no training; 2) attendance to courses, continuing medical education lectures, or conferences; 3) formal rotation during residency or fellowship; 4) 6 months or more of formal training in palliative care; 5) appointment at a hospice or a palliative care service. Participants whose responses fell into categories 2 through 5 were considered to have had previous training in palliative care. We divided service areas into liquid and solid tumor. Participants from leukemia, lymphoma, myeloma, and stem cell transplantation departments were considered a liquid tumor service area. All other participants were considered to fall into the solid tumor service area.
The primary objective of the study was to determine the perception of the impact of the name palliative care compared with supportive care on patient referral as indicated by referrals at specific times in the cancer illness trajectory. We divided the stages of illness into patients who fall into these categories: newly diagnosed with cancer, receiving active primary treatment for cancer, no evidence of active cancer (remission), receiving treatment for advanced cancer, and finally no longer receiving treatment for advanced cancer or in transition to end-of-life. We did not define “active primary treatment for cancer” as treatment with curative intent in our survey. However, at M. D. Anderson Cancer Center, this is the term that is generally used when a patient is treated with curative intent. Because patients with advanced cancer, who are receiving antineoplastic treatment, are generally considered to have a life-limiting illness and the treatment intent is usually palliative in nature, we identified this stage as the most appropriate for the introduction of palliative care services. The primary objective of the study was assessed by asking the participants to rate their likelihood of referral to a service with either the name palliative care or supportive care for a symptomatic patient with advanced cancer who is receiving active treatment, on a scale that ranges from strongly disagree, disagree, neutral/undecided, agree, to strongly agree. In addition, referral at the other stages of cancer was assessed using the same rating scale. Participants were also asked questions related to their perception of the meaning of the 2 names and rated the answers on the same scale. For specific questions, participants were asked to rate their responses on a numerical Likert scale that ranges from 0 to 10, where 0 indicates the best and 10 indicates the worst. All information was collected and analyzed anonymously. The survey instrument is available upon request.
Descriptive statistics were used to summarize the data. Preference for name (pro palliative care, pro supportive care, no preference/other) was compared with demographic variables using chi-square tests or Kruskal-Wallis tests, depending on the distribution of the demographic variable. Responses about perceptions of palliative care and supportive care overall and by type of provider were compared using chi-square tests. Wilcoxon rank-sum tests were used to determine differences between groups as related to severity of distress and usefulness of existing palliative care services. Responses about perceptions of palliative care and supportive care overall and by type of provider were compared using McNemar tests. Analyses were made using SAS version 9.1 (Cary, NC).
A total of 140 (70%) participants responded (66 medical oncologists and 74 midlevel providers). Table 1 summarizes the demographics of participants. Median age was 43 years, and there were 83 (59%) women. The majority of medical oncologists and midlevel providers were providing care to patients with solid tumors (66%), and a little more than half (55%) of them had any type of training in palliative care.
Table 1. Characteristics of the Participants
Midlevel Providers, n=74
Medical Oncologists, n=66
All Participants, N=140
SD indicates standard deviation.
*Agree and strongly agree.
Mean age (SD)
19 (29 )
Mean years in oncology, (SD)
Service area (%)
Training in palliative care (%)
No. of patients who died of cancer in the past year under participant's care (%)
Medical oncologists and midlevel providers stated more likelihood to refer patients to a service with the name supportive care than palliative care for symptom control in patients newly diagnosed with cancer (93, 66% vs 48, 34%; P < .0001), undergoing active primary treatment for cancer (111, 80% vs 61, 44%; P < .0001), in remission or with no evidence of disease (92, 66% vs 48, 36%; P < .0001), and undergoing treatment for advanced cancer (125, 89% vs 96, 69%; P < .0001), respectively. However, there was no difference in likelihood of referring patients who are no longer candidates for therapy or in transition to end-of-life to a service with the name supportive care compared with palliative care (126, 91% vs 127, 92%; P = .8). In addition, there was a clear trend toward referring patients who are later in the trajectory of illness to a service with the name palliative care (P < .0001). These results are shown in Table 2.
Table 2. Likelihood of Referral for Patients at Different Stages of Cancer
I would use this service for a patient with symptoms and who is
PC indicates palliative care; SC, supportive care.
Agree and strongly agree.
Newly diagnosed with cancer
Undergoing active primary treatment for cancer
Without evidence of active cancer (remission)
Receiving active treatment for advanced cancer
No longer receiving treatment - advanced cancer - or is in transition to end of life
When asked about their preference for using the 2 service names when discussing referrals with patients and families, 27 of 140 (19%) preferred using the term palliative versus 80 of 140 (57%) for supportive care (P < .0001).
Table 3 summarizes participants' perceptions of the names palliative versus supportive care as pertains to the following: being a barrier to patient referral, being synonymous with hospice and end-of-life care, ability to decrease hope in patients and families, and their association with treatment of chemotherapy-related side effects (P < .0001).
Table 3. General Perceptions of the Names Palliative and Supportive Care
Midlevel Providers n=74 (%)
Medical Oncologists n=66 (%)
All Participants N=140 (%)
PC indicates palliative care; SC, supportive care.
*Agree and strongly agree.
Service name is a barrier for me to refer patients
Service name is synonymous with hospice and end of life
Service name can decrease hope in patient and families
Service name is associated with treatment of chemotherapy side effects
A total of 20 (14%) participants indicated that using the name palliative care when discussing referrals with patients and families causes them distress versus 6 (4%) when using the name supportive care (P = .0042). In addition, 46 (33%) participants indicated that using the term palliative care causes distress to patients/families versus 5 (4%) for supportive care (P < .0001). The severity of distress evoked by using the 2 names is shown in Figure 1.
Figure 2 shows participants' rating of the usefulness of the existing Palliative Care Service at M. D. Anderson Cancer Center in the assessment and management of symptoms and psychosocial distress in patients and families.
Finally, there was no association between preference for name (categorized as pro palliative care, pro supportive care, or no preference/other) and demographic factors including age (P = .82), sex (P = .35), primary oncologic specialty (P = .06), previous training in palliative care (P > .99), or the number of years practicing oncology (P = .90). However, there was a statistically significant association between preference of the name supportive care and white race versus other races (P = .04).
Our preliminary study revealed that medical oncologists and midlevel providers stated more likelihood to refer patients at earlier stages of the cancer illness trajectory to a service named supportive care compared with palliative care. They were as likely to refer symptomatic patients at the end-of-life to a service named supportive care compared with palliative care (Table 2). These results suggest that, although the name supportive care might have a positive impact on early patient referrals for symptom issues, it does not deter practitioners from referring patients at the terminal stages for the appropriate palliative care measures and advanced care planning.
Participants in our study expressed more likelihood to refer to palliative-care patients who are no longer receiving treatment or in transition to end-of-life care than patients with advanced cancer who are receiving therapies that are generally considered palliative in nature. This finding agrees with the results of a recent study of healthcare practitioners in Pennsylvania using semistructured interviews.21 Content analysis showed that most interviewees characterized palliative care as end-of-life or hospice care that is offered after the decision is made to sustain curative or life-prolonging therapy. For instance, some participants viewed palliative care as reserved for patients with terminal cancer, and some viewed it as care for patients who are in the last hours or days of their life.
The European Society of Medical Oncology (ESMO) has recognized the need to integrate supportive and palliative services to the care of cancer patients.22 Supportive care is defined by ESMO as care that aims to optimize the comfort, function, and social support of the patients and their family at all stages of the illness, including those with curable cancer. Palliative care is defined as care that aims to optimize the comfort, function and social support of the patients and their family when the illness is incurable.22 Our study showed that medical oncologists strongly prefer a service with the name supportive care more than palliative care for patients who are being treated with curative intent. In the context of patients being beyond all anticancer treatment options and in patients approaching the end-of-life, there was no preference to either term. This finding suggests that clinicians perceive a difference between the terms that is related to clinical context. Patients with curable cancer have different needs and require a different emphasis of care. On the other hand, for patients with incurable illness and dying patients, the context of the impending end-of-life makes these issues so much more important and central to the entire clinical endeavor. This distinction between palliative and supportive care based on need and clinical context requires a different approach to care. In addition, whereas the service packages might be different for patients with curable cancer than for those with advanced disease, it should be considered that the delivery of these services is done by the same team to emphasize continuity of care and patient's quality of life issues, and to overcome healthcare delivery system challenges.
The name palliative care compared with supportive care was perceived more frequently as a barrier to patients' referral, synonymous with hospice and end-of-life care and reducing hope in patients and families when compared with supportive care. These results are similar to a study by Morstad et al, which showed that the name supportive care compared with palliative care consistently evoked more positive feelings in pediatric healthcare professionals.18 Our findings suggest that using the name supportive care to define some outpatient palliative care programs might potentially facilitate earlier referral of patients with advanced cancer to such services. However, there are possible limitations to the name supportive care that are worth mentioning. The name supportive care was perceived as less distressing to both patients and clinicians than the name palliative care. This finding might lead clinicians to avoid discussions regarding end-of-life and the palliative nature of the disease and treatment. In addition, the majority of participants (61%) associated the name supportive care with treatment of chemotherapy-related side effects. It is conceivable that this association might deter clinicians from utilizing a service with the name supportive care for other symptom issues. More research is needed to better define the impact of the service name on early patient referral.
Although participants perceived the palliative care services as useful in many aspects and expressed general willingness to refer patients to the palliative care service, there was a clear trend toward increased use of the palliative care service when patients are approaching the end-of-life. These findings suggest that the perception that palliative care service is offered to patients who are no longer candidates for disease-directed therapy is not necessarily related to referring clinicians' satisfaction with these services, but rather to a fundamental misunderstanding of their role and definition.
Interestingly, participants in our study perceived the name palliative care as a cause of distress to not only patients and families when used in a conversation discussing referral to the service, but also to themselves. This finding combined with the finding that palliative care was frequently viewed by our study participants as terminal care that is synonymous with hospice, allude to clinicians' own discomfort with end-of-life discussions and possibly concerns of taking away hope from patients and families. Our study also shows that medical oncologists and midlevel providers prefer to use the term supportive care more than palliative care when discussing referrals with patients and families regardless of their age, years in practice, oncologic specialty, or previous training in palliative care.
Our study demonstrates a significant difference in likelihood of referral of patients with advanced cancer receiving active treatment favoring a service with the name supportive care. However, we did not ask this question with regard to the specific subset of patients with incurable cancer who are receiving chemotherapy with purely palliative intent. Because there could be a difference in these 2 clinical scenarios based on the type and stage of advanced cancer, future research should aim at delineating this further.
Our findings highlight the clamant need for education of healthcare practitioners on the definition of the term palliative care, the role of palliative care services, and the advantages of earlier referrals to these services. Previous research has suggested that such education interventions are likely to influence clinicians' perceptions.23, 24 Therefore, it is imperative that palliative care providers become more proactive in promoting the integration of palliative care services early in the course of life-limiting illnesses and in changing the perception that palliative care is offered only to dying patients who are no longer candidates for life-prolonging therapy.25, 26 In addition, palliative care services should emphasize their positive role in improving quality of life and symptom control so that referrals to these services are not viewed as distressing or countering hope in patients and families.
Although our study did not aim to identify advanced cancer patients' perceptions of the names palliative and supportive care, this is certainly an area that future research will need to address. Previous findings suggested that, although these patients were aware of palliative care (84%) and hospice (81%) services, only 10% and 6%, respectively, considered these services as an alternative to phase 1 clinical trials.27 Another study by Koedoot et al showed that patients with metastatic cancer persistently chose palliative chemotherapy over best supportive care.28 These findings suggest that these patients may decline both palliative and supportive care services when offered as an alternative rather than complementary aspect of their disease-specific treatments.
Our study has some limitations that are worth mentioning. First, this study was conducted in a tertiary referral NCI-designated comprehensive cancer center; therefore, these results might not be generalized to American medical oncologist and midlevel providers practicing in community settings. Similarly, these findings might not be pertinent in countries where palliative care is widely adopted and integrated in the medical system, such as the United Kingdom. Second, as with most survey studies, these responses cannot always be taken as accurate descriptions of what the respondents actually do or really believe about the 2 service names and therefore do not conclude a cause-effect relationship. On the other hand, whereas our study findings do not necessarily reflect how medical oncologists and midlevel providers will react to the 2 names in their daily practice, they highly suggest that the name palliative care might have an impact on referral patterns at different stages of the cancer illness trajectory. Consequently, education of referring clinicians on the definition of palliative care might increase their level of comfort and lead to a positive impact on their utilization of these services. Defining some outpatient services as supportive care and improved education about the scope of these services might result in early and less distressing referrals. In addition, the impact of naming services supportive and palliative care is an alternative that should be studied.