Perceptions and Utilization of Palliative Care Services in Acute Care Hospitals

Abstract

Objective

To understand perceptions of palliative care in acute care hospitals and identify barriers to earlier use of palliative care in the illness trajectory.

Methods

We conducted semistructured interviews with 120 providers involved in decision making or discharge planning and “shadowed” health care providers on intensive care unit rounds in 11 Pennsylvania hospitals, and then used qualitative methods to analyze field notes and transcripts.

Results

Most participants characterized palliative care as end-of-life or hospice care that is initiated after the decision to limit treatment is made. Few recognized the role of palliative care in managing symptoms and addressing the psychosocial needs of patients with chronic illnesses other than cancer. Participants viewed earlier and broader palliative care consultations less in terms of clinical benefits than in terms of cost savings accrued from shorter terminal hospitalizations. In general, participants thought nurses were most likely to facilitate palliative care consults, surgeons were most likely to resist them, and intensive care specialists were most likely to view palliative care as within their own scope of practice. Suggestions for increasing and broadening palliative care integration and utilization included providing workforce development, education, and training; improving financial reimbursement and sustainability for palliative care; and fostering a hospital culture that turns to high -intensity care only if it meets the individual needs and goals of patients with chronic illnesses.

Conclusions

Initiating palliative care consultations earlier during hospitalization will require an emphasis on patient benefits and assurances that palliative care will not threaten provider autonomy.

INTRODUCTION

Palliative care has been associated with better outcomes, including enhanced patient care,^1-3 facilitation of understanding of the diagnosis and prognosis,⁴ increased patient satisfaction,³ improved symptom control,^5-7 decreased health care utilization,³ and improved financial performance of institutions.^8-11 Particularly in response to these last two outcomes – decreased utilization and improved financial performance – palliative care consult services have become more common in academic and community hospitals.^1,2,13 For these programs to grow, referring physicians need to understand what palliative care is, what services it provides, and to whom. The specialty of care grew out of the hospice movement in the 1970s and as such has been associated with care for “dying” patients, particularly those with cancer. Furthermore, hospital-based palliative care has been supported by administrators largely based on the expectation of cost-savings accrued by decreasing length of stay or reducing per-diem resource use among terminal hospitalizations. Thus, for many, palliative care is synonymous with dying.¹⁴

Palliative care physicians want to change this perception and encourage a broader view of palliative care as treatment focused upon symptom relief and the promotion of quality of life among patients with life-limiting chronic illnesses before they are perceived as “dying.” This transition “upstream” in the course of life-limiting illness is motivated by two issues. First, the uncertainty about when dying begins in non-cancer patients has resulted in very late palliative care referral and sub-optimal treatment. Second, many studies report high symptom burden, caregiver stress, and communication difficulties during both terminal and non-terminal hospitalizations among patients with life-limiting illnesses.^15-17

While there are some studies of how patients view palliative care services,^18-20 few studies have explored how physicians and other health care providers in acute care hospitals perceive and utilize these services. A better understanding of provider perceptions could inform efforts to broaden the use of palliative care by its proponents. To address these issues, we performed a qualitative analysis of the contents of our field notes and a sample of interviews conducted during 2-day site visits to 11 acute care hospitals in Pennsylvania as part of a study to identify organizational programs, policies, and practices that influence the intensity of treatment given to elderly patients with life-limiting illnesses.

METHODS

Our study was part of a larger project that investigated hospital-level variations in end-of-life treatment intensity. The University of Pittsburgh Institutional Review Board (IRB) approved both the larger project and the current study and exempted them from the requirement for written informed consent. One participating hospital required additional review and approval by its own IRB, and obtained this approval.

Settings and Participants

We limited the sampling frame to Pennsylvania hospitals to minimize differences in treatment patterns attributable to state-related legal, regulatory, and financing influences. Within Pennsylvania, we used purposive sampling to identify a diverse group of hospitals with respect to location and rural-urban continuum code category, bed size, teaching status, and observed treatment intensity among hospital decedents (rates of intensive care unit [ICU] admission, mechanical ventilation, hemodialysis, and gastrostomy tube placement among terminal admissions aged 65 or older in 2000). Of 20 hospitals sampled, 13 agreed to participate, and we completed site visits to 11 hospitals in 2004. Two-day site visits included observation of morning ICU rounds, when feasible, and semistructured interviews with the Chief Nursing Officer, the Director of Case Management, the physician Director of the Emergency Department (ED), the physician Director of the ICU, the ICU Nurse Manager, the Chief of Surgery, two “high-volume” physicians, a bedside ICU nurse, an ICU social worker, an oncology social worker, the Director of Pastoral Care, the Director of Palliative Care (if present), and the Chair of the Ethics Committee.

Data Collection

During ICU rounds, an investigator took handwritten notes about the clinical practices and discussions that took place concerning life-sustaining treatments and palliative care delivered to critically ill patients with life-limiting chronic illnesses.

From each participant who agreed to be interviewed, the investigator collected information about sociodemographics, medical specialty, years in practice, years at the current hospital, and perceptions of standards of care and determinants of care for chronically ill patients at the end of life. To ensure uniformity in topics covered and allow for valid comparison of responses, the investigator used a standardized interview protocol and specific sequencing of questions. Each interview lasted about 60 minutes and was audiotaped.

At the beginning of each interview, the investigator asked the participant to present his or her most recent case involving the treatment of a patient who was aged 65 or older and died in the hospital. Treatment decisions included whether to admit the patient to an ICU and whether to initiate processes such as intubation, dialysis, enteral or parenteral nutrition, transfusion, or cardiovascular support. The investigator followed up with questions about why certain decisions were made and whether they were affected by factors related to the patient, patient's family members, community, health care providers, and organizational policies, procedures, or resources.

At the end of the site visits, the handwritten field notes were typed and the audiotaped interviews were transcribed by a professional transcriptionist.

Development of Coding Criteria

Two study investigators read the field notes and a sample of the interview transcripts to identify common themes and concepts that emerged concerning palliative care.^21-23 After discussing their initial findings, they developed a set of criteria to code text segments in terms of content pertaining to the participants’ perceptions of palliative care services, the circumstances under which these services were used, and the barriers to and facilitators of their use.

To test the criteria, 2 coders independently applied the criteria to text segments that included discussions of palliative care in 6 randomly chosen sets of field notes. For purposes of coding, they defined discussions of palliative care as any talk concerning symptom management, quality of life, the use of pain medications, or the use of other methods to alleviate suffering and provide comfort. In cases in which the coders disagreed on the coding of particular text segments, they discussed their differences and refined the criteria. They continued to refine the coding scheme until they reached 100% agreement.

Coding and Analysis of Field Notes

When we applied the finalized coding scheme to the field notes, we coded for a variety of factors, including: how palliative care is defined by participating individuals and hospitals; whether a hospital has a palliative care team; who is currently on the team; how long the team has functioned; who spearheads the palliative care efforts; who initiates a palliative care consult; how a person initiates a consult (via formal or informal channels); who approves and implements the use of palliative care services; and why these services are or are not instituted (goals, facilitators, and barriers).

We systematically analyzed patterns in the recurrence, distribution, and relationships of emergent palliative care concepts and themes in data regarding individual hospitals and the group of hospitals.

RESULTS

Characteristics of Hospitals

Of the 11 hospitals (Table 1), 4 (36.4%) were located in large cities, 3 in medium-sized cities, 2 in small towns, and 2 in rural areas. While 5 (45.5%) were major teaching hospitals, 4 were minor teaching hospitals and 2 had no residency programs. The mean number of hospital beds was 417.1 (range, 104–761 beds). We observed ICU rounds in 7 of the 11 hospitals. In 2 hospitals, there was no form of rounding.

Table 1.

Characteristics of the 11 Hospitals in the Study

Hospital	Location	Teaching Role	Number of Beds	Percentage of Patients
				In ICU	With Mechanical Ventilation	With Tracheostomy	With Hemodialysis	With Gastrostomy	With Pulmonary Artery Catheterization
A	Large city	Major	491	77.9	40.0	8.2	5.2	0.0	0.0
B	Large city	Major	694	86.3	35.4	11.2	6.0	1.0	2.0
C	Large city	Major	701	93.4	43.4	9.7	5.0	0.6	8.6
D	Small town	Major	548	81.9	31.7	1.9	4.4	1.0	0.0
E	Small town	Major	330	90.7	43.1	3.4	3.4	0.5	0.0
F	Large city	Minor	237	66.3	28.7	2.2	0.0	1.1	3.4
G	Medium city	Minor	761	58.8	24.2	4.5	7.6	1.1	0.6
H	Medium city	Minor	232	58.8	23.0	1.5	3.9	0.5	1.5
I	Medium city	Minor	353	84.5	23.5	1.9	9.9	0.9	2.5
J	Rural area	None	104	68.4	6.6	1.3	2.6	0.0	0.0
K	Rural area	None	137	81.3	17.3	0.0	10.7	1.3	0.0

Two of the 11 hospitals had no current formal form of palliative care. Of the remaining 9 hospitals, 1 had a palliative care curriculum for family practice residents but no palliative care service, 3 had a palliative care consult service, 2 had only oncological-palliative care (1 with a 5-bed palliative care unit on the oncology floor, and 1 with some palliative care beds in the oncology unit), 2 had newly instituted dedicated beds and a consult program, and 1 had a palliative care consult service that was recently reinstituted. No differences in discussion of themes emerged based on maturity of the program.

Characteristics of Individual Participants

We interviewed 139 individuals (Table 2). Most (59.0%) were female, and the vast majority (95.0%) were white. While 42.4% had a nursing degree, 38.8% had a medical degree, 6.5% had a social work degree, and the remainder had other professional degrees. The mean number of years in practice was 19.4 (range, 1–45 years). About half of the individuals (52.5%) were involved in direct patient care, and the mean number of patients who died under their care during the past month was 6.3 (range, 0-40 patients). Of the 139 individuals, 120 had a complete (60-minute) interview. The others arrived late and had a shorter interview.

Table 2.

Sociodemographic Characteristics of the 139 Participants in the Study^*

Characteristic	Number (%)
Sex
Male	56 (40.3)
Female	82 (59.0)
Unknown	1 (0.7)
Race
White	132 (95.0)
African American	2 (1.4)
Asian or Pacific Islander	3 (2.2)
Other	1 (0.7)
Unknown	1 (0.7)
Attends religious services
More than weekly	13 (9.4)
About once weekly	47 (33.8)
About once monthly	37 (26.6)
About once yearly	23 (16.5)
Less than once yearly	7 (5.0)
Never	11 (7.9)
Unknown	1 (0.7)
Professional degree
Doctor of medicine (MD) or osteopathic medicine (DO)	54 (38.8)
Registered nurse (RN)	59 (42.4)
Licensed clinical social worker (LCSW)	9 (6.5)
Other	16 (11.5)
Unknown	1 (0.7)
Years in practice (range, 1–45)
1–9	23 (16.5)
10–19	49 (35.3)
20–29	42 (30.2)
30–39	20 (14.4)
40–49	3 (2.2)
Unknown	2 (1.4)
Years at current hospital (range, 1–46)
1–9	58 (41.7)
10–19	46 (33.1)
20–29	24 (17.3)
30–39	8 (5.8)
40–49	2 (1.4)
Unknown	1 (0.7)
Primary job function
Clinical care	63 (45.3)
General social work	10 (7.2)
Case management	8 (5.8)
Management/supervision	29 (20.9)
Other administrative	12 (8.6)
Other nonadministrative	15 (10.8)
Unknown	2 (1.4)
Primary place of work
Emergency department	8 (5.8 )
Intensive care unit (ICU)	33 (23.7)
Oncology unit	2 (1.4)
Throughout the hospital	31 (22.3)
Palliative care or hospice	3 (2.2)
Surgery or operating room	5 (3.6)
Administrative offices	11 (7.9)
Other	12 (8.6)
Unknown	34 (24.5)
Provides direct patient care
Yes	73 (52.5)
No	35 (25.2)
Unknown	31 (22.3)
Has experienced the death of a loved on in a hospital
Yes	93 (66.9)
No	31 (22.3)
Unknown	15 (10.8)
Has experienced the death of a loved on in an ICU
Yes	57 (41.0)
No	67 (48.2)
Unknown	15 (10.8)

^*Because of rounding, percentages may not all total 100.

Themes Related to Palliative Care

Rather than ask participants directly about palliative care, we asked them about treatment decisions in general. Their discussions about palliative care at the hospitals varied widely and included talk about palliative care training, curriculum, programs, services, teams, and specific experiences. However, from our analysis of notes and transcripts, we discerned 4 major themes: perceptions of palliative care, current utilization of palliative care, barriers to utilization of palliative care, and suggestions for increasing palliative care consults.

Perceptions of Palliative Care

Most participants perceived palliative care to be a type of care that focuses on terminal pain and symptom management and on facilitating decisions to stop life-sustaining treatments. Some viewed palliative care as reserved for patients with terminal cancer, and some viewed it as care for “actively dying” patients (i.e., during the last days or hours of life). For example, one physician stated: “[The patient's] on a non-rebreather mask in the ER, he looks terrible and he looks like he's gonna’ die. I said, ‘Well, get him a bed in palliative’.” Aside from nurses, not many participants identified palliative care's role earlier in the disease process.

Conflation of the terms “palliative care,” “hospice care,” “end-of-life care,” and “terminal cancer care” tended to be at the root of confusion related to who should receive palliative care, who should recommend and provide it, and when it should be started. For example, as one physician and ED director stated: “[The patients] are coming [to the ED] for medical treatment, not for palliative treatment.... The [‘Hospice House’] is palliative treatment. They are...going there [to ‘Hospice House’] because, you know, perhaps they have intractable pain and they are near the end.” This view of palliative care was based on disease progression (i.e., when it is “time”) rather than on the need for services.

In other cases, the participants worried that other individuals (physicians, patients, and patients’ family members) viewed the institution or consideration of palliative care as a signal that health care workers had abandoned all hope for a specific patient. For health care providers, the options are sometimes viewed as “doing everything or doing nothing,” and palliative care is essentially viewed as “pulling the plug” or “doing nothing.” A number of participants viewed palliative care as incompatible with the hospital and ED goals of saving lives, and some surgeons perceived instituting palliative care as being “soft” or “giving up.”

Although physicians rarely discussed palliative care as a supportive service that could be useful 6–12 months before death, nurses tended to believe that it should be available to all patients, not just as an option after disease-oriented care fails or becomes too burdensome and not just when patients reach the end of life. Further, nurses emphasized the role of palliative care in facilitating discussions and decision making about goals of care and quality of life; issues that they felt other treating physicians frequently neglected.

Many physicians felt that symptom management for pain and shortness of breath was part of their job and core competency. Thus, they felt that palliative care consults in non-terminally ill patients were “lateral consults.” Others recognized palliative care as a specialty whose practitioners can be called upon for various roles with actively dying patients, such as intervening when a patient's expressed desires are being ignored, dealing with “difficult” patients and family members who have unrealistic expectations about the patient's disease outcome, and taking the time to listen to patients and “sort out” what they want. As one palliative care provider described it, “What we're doing is what everyone else is NOT doing.” In this sense, many discussed palliative care consults as central to conflict resolution concerning care goals and decision making during the last days of life.

Current Utilization of Palliative Care

As noted above, discussion about why medical providers choose to use palliative care often focused on conflict around care goals. Among the circumstances under which participants felt that palliative care consults should be arranged were cases in which aggressive care was futile, end-of-life preferences were not being met, there were disputes over the course of care, or there were concerns about high-intensity treatment, particularly in elderly patients. The focus of palliative care was getting the family or other health care providers to do “less.” The sense was that the palliative care team would be more successful in getting terminally ill patients and families to go along with a comfort limitation or withdrawal of life-sustaining treatments. Cases in which there was earlier consultation for comfort measures in a patient who might soon die included those in which patients were stridently committed to comfort care, were competent, and supported by their family.

Consults are physician dependent. As one nurse indicated, the willingness to use palliative care “greatly depends on the medical doctor” and “everything is very doctor-driven.” Although anyone can informally suggest a palliative care consult, the patient's physician usually has to formally order it. In one hospital, physicians and nurses launched an effort to make it allowable for registered nurses or social workers to initiate a palliative care consult, but this effort failed.

Nevertheless, the palliative care providers whom we interviewed said they received a lot of “anonymous” tips from registered nurses and even radiation therapists who believe that a patient's care may be futile or inappropriate for expected outcomes. This suggests that palliative care is gaining recognition, but primarily in negotiation of “more appropriate” treatment plans in actively dying patients.

Further, the nurses try to push physicians to transfer patients out of the ICU if they are either stable or better served by receiving “palliative care” (i.e., comfort measures only). Some experienced nurses discussed efforts to educate newer nurses about how to get a palliative care consult. For example, one surgical ICU nurse stated:

[A palliative care consult is] either nurse or family directed.... Often times it's the family coming to the nurse and saying, “What do you think? Should we...keep going?” And then the nurse kind of initiates a conversation, but based on a family member's comment. And then we will take that to the physician, but present it as “the family asked me about [this],” not as “we need to talk about [this].”

Similarly, social workers try to alert physicians to psychosocial issues they identify, and generally the physicians are receptive to having these issues addressed.

Some hospitals reported having high rates of inter- and intra-hospital transfers to their palliative care services from other hospital units when death is imminent. Sometimes the palliative care services admitted patients directly from the community. For example, they admitted home hospice patients whose families become afraid at the end of the patient's life. In one case, a woman had difficulties dealing with home hospice for her husband, a patient with esophageal cancer, but the hospice service said it would not follow the patient into a nursing home, so the patient was admitted to the hospital and died there.

Nursing homes in one city were fearful of being sued or accused in the newspaper of “doing nothing” for their patients who experienced changes in mental status and became difficult to manage, so they tended to transfer these patients to the ED. Many were subsequently admitted with terminal delirium to the geriatric psychiatric service, where a palliative care consultation was obtained.

Additionally of note, when participants spoke about critically ill patients’ needs, they mentioned relying on a variety of non-palliative care programs. Those discussed included ethics committees, pastoral care services, hospitalist programs, and long-stay review boards. For example, at two hospitals, medical doctors and registered nurses relied heavily on the chaplain. He or she served as the bridge between the family and the medical doctors.

Barriers to Utilization of Palliative Care

Participants were more likely to mention barriers to using palliative care near the end of life than they were to mention barriers to using it earlier in the disease trajectory, not surprisingly given their aforementioned perceptions of palliative care. Participants identifying underuse of palliative care earlier in the disease trajectory were palliative care providers or advocates, such as nurses, social workers, and administrators. They perceived that the biggest barriers to early use were the narrow view about palliative care's range of services discussed above and physicians ' concerns about loss of autonomy. Specifically, physicians worried that palliative care providers would change the goals of care before they [the treating physician] were ready.

With respect to barriers to using palliative care near the end of life, participants frequently mentioned the following factors as barriers to their using or others using palliative care services: the pervasiveness of a culture currently geared toward high-intensity treatment; the concerns of hospitals and health care providers about reputations and lawsuits; the beliefs and perceptions of patients; the beliefs, age, and medical specialty of individual physicians; and economic constraints that affect hospital practices and staffing.

Numerous participants described the hospital culture as one that embraces only high-intensity treatment (e.g., an active life-sustaining treatment plan) and resists palliative care and comfort measures. For example, one oncology and palliative care nurse stated:

Even though [patients] are in palliative care [a designated inpatient unit],...we give blood, we've had people still on dialysis, we have people that are not just on 1 antibiotic but 4 and 5.... You have the patients at the other end where it truly is comfort measures only,...but many have...aggressive medical therapy and they are still a “palliative” DNR.

In another hospital, the culture was described as “very aggressive unless there is a written advance directive and it's very clear.” In this type of hospital culture, patients and families were described as expecting “the best,” which was felt to be synonymous with aggressive, technology-intensive treatment.

The physicians’ concerns about which patients to refer to palliative care services and when to refer them are often coupled with the physicians’ individual resistance to transitioning from cure to palliation. Although non-physicians perceived physicians generally to resist this transition, some specialists were singled out. As an example, one participant described an oncologist as wanting to give “another round of chemo” even though it seemed futile to other members of the care team, and, in another hospital, a social worker said of oncologists, “If they can't cure them,” they “kill them” in the process of trying. Surgeons were singled out as the most resistant, often irascibly so. The nurse ICU manager at one hospital remarked that surgeons “never say 'die',,,the vascular surgeons will resist until the patient's totally mottled and the toes are falling off. They just absolutely will not ask for a palliative care consult.” According to participants at one hospital, cardiologists rarely consulted palliative care providers and only did so after consulting neurologists and other subspecialists. In contrast some participants viewed hospitalists as more likely than other physicians to initiate life-sustaining therapy but also more likely to withdraw the therapy if it wasn't meeting treatment goals.

Some physician resistance stemmed from lack of training in comfort measures and the conflation of personal moral beliefs and treatment goals and outcomes. For example, some physicians felt it was “inhumane” or “unethical” not to continue feedings, even if the patient had an advance directive indicating no enteral or parenteral feedings. As one participant stated, “Even though they [the physicians] were receptive to palliative care, they still insisted on IV fluids.” One participant felt that physicians who were “pro-life” were more likely to push for more intensive treatment.

In general, participants believed that young physicians were more likely to support palliative care and more likely to consider the opinions of nurses about the needs of patients, in part due to medical school and residency training with explicit end-of-life curricula. However, older physicians were sometimes felt to be better at finessing difficult conversations like those related to code status and transition to comfort measures, but also older physicians were more resistant to high doses of pain medication and to palliative care consultants’ threats to their autonomy.

These physician-based perception barriers were largely cited as culprits in failed administration and provider initiatives to change relevant hospital standards of care. Additionally, though, proponents and providers of palliative care cited resource limitations as barriers to increasing utilization of palliative care. For example, one hospital had only enough money to fund several part-time physicians and nurse practitioners in palliative care. Further economic concerns were voiced in relation to insurance reimbursement issues. In one hospital, a case manager explained that when patients were at the end of life and were transitioned to palliative care, the insurer would lower the rate paid to the hospital to a skilled nursing rate, creating tension between hospital finances and providers’ perceptions of the patients’ best interest.

Suggestions for Increasing Palliative Care Consults

In most hospitals we visited, participants felt that palliative care services were underutilized (particularly at the end of life) and that change would take time. Hospital staff suggestions included the following: arranging negotiations involving hospital administrators, chaplains, ethics committee members, and units with continued problems; providing workforce development, education, and training about palliative care; increasing the number of staff involved in palliative care; improving financial reimbursement and sustainability for palliative care; raising community awareness; and changing the normative hospital culture.

Recommendations to increase end-of-life consults, particularly for withdrawal of life support, centered on the need to attend to frustrations with futile treatment, communicate with resistant doctors, and change the hospital culture geared toward high-intensity care. These motivations were commonly centered on potential cost savings accrued from shorter terminal hospitalizations rather than on clinical benefits. Although less commonly discussed, recommendations to move palliative care “upstream” in the illness trajectory or broaden its use included such strategies as incorporating palliative care providers into care teams for high risk patients (e.g., heart failure patients being considered for ventricular assist devices) and mandating palliative care consults for all medical intensive care unit admissions. To increase consults for actively dying patients and to move consults “upstream,” several physicians mentioned the value of role modeling and collegiality – if palliative care providers had strong clinical skills, were “good citizens” of the medical staff, were true to their word, followed through with ideas and programs, were highly reliable and were not perceived as usurping consulting physicians’ autonomy, they might increase utilization of their services over time.

Finally, to overcome the barrier of the prevailing perception of palliative care as the “death service,” several interviewees suggesting changing service branding. For example, one palliative care consult team had billed themselves as the “pain and symptom management consult service.” Several interviewed palliative care providers believed that chronic disease management, a holistic approach, was a better description of what they did and possibly even a better name for palliative care.

DISCUSSION

When we visited 11 hospitals and had full-length interviews with 120 health care providers, we found that there were several common misperceptions that must be changed before palliative care services can be integrated and utilized optimally in the acute care hospital setting. These include beliefs that palliative care is applicable only in the last days of life, is antithetical to intensive care, is reserved for patients with cancer, is a signal that the physicians have given up on the patient, and would be a “lateral consult” since symptom management is within most physicians’ scope of practice.

Although we found that most participants knew that palliative care focuses on pain relief and decision-making guided by goals of care for patients near the end of life, few identified its role in managing other symptoms of life-limiting illnesses and addressing the needs of patients who were not actively dying. When palliative care consults were felt to be initiated “too late” it was primarily because the respondent perceived limitation of life-sustaining treatment should have occurred earlier in the hospital course. Very few identified the initiation of palliative care consultation earlier in the illness trajectory as an unmet need.

Opportunities to address a variety of care needs, including symptom control, psychosocial concerns, spiritual needs, physician-patient-family communication, and the needs of health care providers, may be missed if palliative care services are considered only when death is thought to be imminent or disease-oriented care has failed. In some cases, it is difficult to predict how long patients will live, particularly if they have diseases other than cancer. For instance, patients with congestive heart failure typically die of sudden arrhythmia and do not have a predictable terminal phase of progressive decline.²⁴ Thus, it would be optimal if discussions of palliative care were instituted earlier in the disease trajectory. Moreover, palliative care should be regarded as coexisting with disease-oriented therapy throughout a patient's care.^25,26 As the disease progresses, the patient may decide to forgo some types of therapies, at which time the relief of symptoms and attention to psychosocial needs may become paramount.

In discussing the benefits of earlier consultation, participants were more likely to mention potential cost savings than clinical benefits. Indeed, hospital-based palliative care is promoted by the Center to Advance Palliative Care as a means for hospitals to reduce costs through shorter length of stay or lower costs per day once the patient has been referred to palliative care. Generally such cost savings are realized by moving patients to lower acuity settings (e.g., out of the ICU) or by decreasing resource intensity (e.g., discontinuing treatments) and have as their hallmark, “doing less,” particularly for actively dying patients. This professional marketing and the types of consultation it creates serve to reinforce clinician misperceptions of palliative care as a service reserved for those in whom all other (life-sustaining) treatments have failed. Ultimately, clinicians often view palliative care as “death services” because most hospital-underwritten palliative care programs were created for, and thus cater to, this narrow demand.

We found that most nonphysicians felt that treatment decisions, including palliative care consultation or hospice referral, depended almost entirely on the idiosyncrasies of treating physicians, rather than the clinical or social circumstances of the patient. In general, experienced nurses were perceived to be most amenable to palliative care service utilization, educating new nurses and informally facilitating initiation of consults via families, residents, or sympathetic nonprimary attending physicians. Among physicians, surgeons were perceived as the most resistant to palliative care services, fearing interference with their decisions to continue life-sustaining treatment. And intensivists often believed that they had the same skills as palliative care consultants.

Although all physicians should have basic palliative skills, not all will be able to treat more complicated patients. The idea that all physicians are inherently capable of delivering such services perhaps stems from the perception that palliative care refers primarily or even exclusively to symptom management.²⁷ This particular aspect of palliative care is something that physicians start learning the first day of internship, but it is seldom taught in a systematic manner and seldom covers special problems of symptom management that emerge in specific hospital settings.^28-30 For example, in the ICU setting,^31,32 problems of particular concern include performing pain-provoking procedures,³³ greater difficulty in assessing pain and discomfort,³³ and encountering higher levels of stress in patients’ families.^{34, 35}

A number of patients will require specialist palliative care from an experienced interdisciplinary team whose core activity is dealing with more complex problems, and access to such advice or care should be widely available. In our study, hospital staff suggestions for increasing integration and utilization of palliative care included providing education and training. Education and training must address issues concerning how to communicate effectively with patients and their family members,³⁶ how to involve patients in decision making early in the course of their illness,³⁷ and how to ensure that the advance care planning goals of patients are met.³⁸ An in-house education program for hospital staff about available palliative care services and how staff can work with consultative teams to provide such services to patients with life-limiting illnesses may help to achieve such ends. Further, by acting as an advisor to the patient's primary medical provider and staff, the palliative care consultant or team can promote seamless integration among institutions, settings, and services by using the full array of inter-institutional and community resources (e.g., hospital, home care, hospice, long-term care, adult day services).

Some of our findings are consistent with results of earlier studies. Although palliative care may be gaining an increasing foothold in medical practice, there continue to be problems in the way that it is perceived by health care professionals and patients.^8,29,39 Views that it is synonymous with hospice care have been shown to fuel the perception that palliative care has little place in acute inpatient medicine.^30,40,41 Although nurses in acute care settings have been shown to be more accepting of a multidisciplinary team approach,³¹ many studies note the need for further staff education and training, particularly regarding the dimensions of palliative care and the identification of patients who could benefit from this care.^27,32,42,43 Of particular importance are doctors’ views on palliative care, which are likely to be particularly influential in shaping an individual patient's care course.^29,44,45 For example, the belief that the decision to limit life-sustaining treatment is primarily physician-driven was confirmed in a recent study of 15 university-associated ICUs.⁴⁶

Our study involved a limited number of hospitals and health care workers and was conducted in only one state. These factors limit the generalizability of our results concerning perceptions and utilization of palliative care in acute care hospitals. However, despite the limitation of our sample, by interviewing nonphysicians in addition to physicians, we obtained more nuanced information about clinical decision making. Indeed, physicians often had little insight into their own behavior but nurses and case managers, who have the opportunity to observe multiple physicians taking care of similar patients, were able to detect and describe variations in physician behavior and attitudes. It is possible that our findings of the conflation of palliative care and terminal care are a result of our interview protocol, which asked respondents to discuss treatment decisions among terminal hospitalizations. However, we did not prompt mention of palliative care and instead found many respondents revealed their preconceptions right away upon hearing that our study was about end-of-life care, when they often responded, “oh, you mean palliative [or hospice] care?”

Replication of our methods in other settings with a larger and more diverse sample of health care providers is warranted, because insights into individual-level and hospital-level factors associated with variations in the utilization of palliative care services will help eliminate the barriers to receiving these services.