Availability and Integration of Palliative Care at US Cancer Centers

Context The current state of palliative care in cancer centers is not known.

Objectives To determine the availability and degree of integration of palliative care services and to compare between National Cancer Institute (NCI) and non-NCI cancer centers in the United States.

Design, Setting, and Participants A survey of 71 NCI-designated cancer centers and a random sample of 71 non-NCI cancer centers of both executives and palliative care clinical program leaders, where applicable, regarding their palliative care services between June and October 2009. Survey questions were generated after a comprehensive literature search, review of guidelines from the National Quality Forum, and discussions among 7 physicians with research interest in palliative oncology. Executives were also asked about their attitudes toward palliative care.

Main Outcome Measure Availability of palliative care services in the cancer center, defined as the presence of at least 1 palliative care physician.

Results A total of 142 and 120 surveys were sent to executives and program leaders, with response rates of 71% and 82%, respectively. National Cancer Institute cancer centers were significantly more likely to have a palliative care program (50/51 [98%] vs 39/50 [78%]; P = .002), at least 1 palliative care physician (46/50 [92%] vs 28/38 [74%]; P = .04), an inpatient palliative care consultation team (47/51 [92%] vs 28/50 [56%]; P < .001), and an outpatient palliative care clinic (30/51 [59%] vs 11/50 [22%]; P < .001). Few centers had dedicated palliative care beds (23/101 [23%]) or an institution-operated hospice (37/101 [37%]). The median (interquartile range) reported durations from referral to death were 7 (4-16), 7 (5-10), and 90 (30-120) days for inpatient consultation teams, inpatient units, and outpatient clinics, respectively. Research programs, palliative care fellowships, and mandatory rotations for oncology fellows were uncommon. Executives were supportive of stronger integration and increasing palliative care resources.

Conclusion Most cancer centers reported a palliative care program, although the scope of services and the degree of integration varied widely.

Although significant progress has been made in cancer treatment, approximately half of all patients with cancer eventually die of their disease, and one-third of cancer deaths happen within 6 months of diagnosis.¹ Palliative care outpatient clinics, inpatient consultation teams, palliative care units (PCUs), and hospices all play an integral role providing symptom control, psychosocial support, and transition of care for patients with cancer and their families along the cancer care continuum. Based on this understanding, multiple national and international organizations support early incorporation of palliative care into oncology practice.^2-5 Most recently, the American Society of Clinical Oncology has taken a step further to support full integration of palliative care as a routine part of comprehensive cancer care in the United States by 2020.⁶

The understanding that palliative care is most effective when incorporated early in oncology care led to the development of the comprehensive cancer care model, which integrates palliative care along with anticancer therapy from the time of diagnosis.^7,8 The current state of palliative care availability and degree of integration of palliative care services into oncology practice at various cancer centers in the United States are unknown. Many cancer centers claim that they provide palliative care, although the structures, processes, and outcomes for their programs remain unclear. A more thorough picture of the level of palliative care that currently exists in these cancer centers would identify deficiencies and barriers to comprehensive cancer care and allow development of specific strategies to improve care delivery.

In this survey study, our goal was to determine the availability and degree of integration of palliative care services in US cancer centers, as well as their executives' attitudes toward palliative care. The palliative care services of National Cancer Institute (NCI)–designated cancer centers were also compared with those of non-NCI cancer centers.

The institutional review board at The University of Texas M. D. Anderson Cancer Center approved this study and waived the requirement for review and informed consent. Survey questions were generated after a comprehensive literature search, review of guidelines from the National Quality Forum,^9,10 and discussions among 7 physicians (4 medical oncologists and 3 palliative care specialists) with research interest in palliative oncology. Survey questions were generated based on the Donabedian tripartite division of structure, processes, and outcomes to evaluate the administrative, clinical, education, and research aspects of palliative care in cancer centers (see eTable),¹¹ and subsequently reviewed for readability and face validity. The surveys were further revised by our institutional Patient-Reported Outcomes, Survey and Population Research Shared Resource. Three physicians (D.H., M.D.L.C., and D.S.Z.) tested the final version of the surveys and assessed the time required for completion.

Two surveys were used in our study. The first survey was directed to cancer center executives (see eAppendix 1) and consisted of 22 questions, which provided information regarding access to and attitudes toward palliative care. The second 74-question survey was sent only to cancer centers with palliative care programs (see eAppendix 2) and asked palliative care program leaders details about their programs, including personnel, inpatient PCUs, consultation teams, outpatient clinics, hospice, and educational and research activities.

Cancer centers in the United States were identified by using the Commission of Cancer database. The Commission of Cancer accredits hospitals as cancer centers based on 36 standards that encompass a range of clinical, research, and quality improvement aspects of cancer care.¹² We obtained a list of 1482 Commission of Cancer–accredited cancer centers, which represent approximately 30% of all US hospitals and provide services to approximately 70% of patients with new cancer diagnoses.¹³ Among these institutions, 71 were NCI-designated centers (40 NCI-designated comprehensive cancer centers, 15 NCI clinical cancer centers, and 16 NCI community cancer centers). The NCI designation is based on scientific excellence related to cancer prevention, diagnosis, and treatment and does not bear any specific reference to the quality of patient care or palliative care research. National Cancer Institute–designated cancer centers are widely recognized as leading institutions in the United States and represent a well-defined cohort in previous studies.¹⁴ We surveyed all 71 NCI-designated cancer centers and a sample of 71 of 1411 non-NCI cancer centers using a simple randomization scheme.

We determined the names and contact information of cancer center executives through the Commission of Cancer and NCI Web site, with independent verification through telephone contact. Information on palliative care program leaders was identified by calling the palliative care department directly. All respondents received an initial invitation with a mail survey, along with a US $10 gift certificate. They or their delegates were asked to complete the survey anonymously and to return it by mail, fax, or electronically through a secured Web site. A reminder letter was sent to nonrespondents at 2 weeks and 4 weeks, followed up by a telephone or e-mail reminder at 8 weeks. Data collection was performed between June and October 2009.

Response rates were determined following the metrics of the American Association for Public Opinion Research.¹⁵ The formula for calculation was response rate = (complete responses+ partial responses)/(complete responses + partial responses + nonresponse + explicit refusals + implicit refusals). In this study, a complete response was defined as 80% or more of questions answered, a partial response as less than 80% of questions answered and completion of the key question in the corresponding survey regarding the availability of palliative care (question IIB for the executive survey and question IB for the palliative care program leader survey [eAppendixes 1 and 2]), and nonresponse as not answering the key question. Three of 120 palliative care program leader surveys sent were later found to be ineligible due to absence of palliative care programs in those centers and were not included in the analysis.

Our primary outcome was the availability of palliative care services in the cancer center, defined as the presence of at least 1 physician with dedicated time to the provision of palliative care (with or without board certification). We compared the proportions of NCI and non-NCI cancer centers that met this criterion using a binomial test. Given that we surveyed 71 NCI cancer centers and 71 non-NCI cancer centers with an anticipated response rate of 65% (therefore, 46 evaluable NCI centers and 46 evaluable non-NCI cancer centers), we were able to declare as statistically significant a difference of 28% between groups, assuming a 2-sided significance level of .05 and 80% power.

We summarized the availability of various types of palliative care services at different cancer centers using standard descriptive statistics, including medians, interquartile ranges (IQRs), proportions, and frequencies, together with 95% confidence intervals (CIs) where appropriate. Differences in services provided between NCI-designated cancer centers and non-NCI cancer centers were computed using Fisher exact tests for categorical variables. The Mann-Whitney test was used for nonparametric continuous variables. P < .05 was considered statistically significant.

SPSS version 16.0 (SPSS Inc, Chicago, Illinois) software and STATA version 10.0 (StataCorp, College Station, Texas) were used for statistical analyses.

Among the 142 executive surveys sent, we had 89 complete responses (63%), 12 partial responses (8%), 40 implicit refusals (28%), and 1 explicit refusal (1%). The response rate was 71% (101/142), and did not differ by cancer center type (NCI vs non-NCI) or whether the center had a palliative care program based on independent telephone verification.

Among the 120 palliative care program leader surveys sent, the number of complete responses, partial responses, implicit refusals, explicit refusals, and ineligible cases were 86 (72%), 10 (8%), 21 (18%), 0, and 3 (2%), respectively. The response rate was 82% (96/117). Palliative care program leaders from NCI cancer centers were more likely than those from non-NCI cancer centers to respond (61/67 [91%] vs 35/50 [70%]; P = .007).

We did not detect any differences in Commission of Cancer center type between the non-NCI cancer centers that were surveyed and those that were not surveyed (P = .98).

The barriers to delivery of palliative care identified by cancer center executives are shown in the Figure. Poor reimbursement and limited institutional resources were the most commonly cited reasons. A small number of executives were concerned that the presence of palliative care could negatively affect their hospital mortality rate and national rating.

Cancer center executives rated their current pain and palliative care services favorably and reported a significant improvement compared with 5 years ago (Table 1). They strongly agreed that further integration of palliative care services into oncology practice will benefit patients, and that more funding should be directed toward palliative care research (Table 2). National Cancer Institute cancer center executives were significantly more likely than their non-NCI cancer center counterparts to agree with an increase in palliative care resources at their institutions over the next 5 years.

A majority of responding cancer center executives reported an active palliative care program (Table 3). Compared with non-NCI cancer centers, NCI-designated cancer centers were significantly more likely to provide palliative care services, have at least 1 palliative care physician, an inpatient consultation team, and an outpatient clinic. Fewer centers had dedicated acute palliative care beds or an institution-operated hospice.

Table 4 provides an overview of existing palliative care programs based on information provided by palliative care program leaders, including the range of services and personnel. Palliative care services have been in place at NCI cancer centers longer than at non-NCI cancer centers. A majority of programs reported a short patient follow-up duration of less than 1 month. Although most palliative care teams had physicians (80%; 95% CI, 71%-88%), physician assistants (71%; 95% CI, 62%-80%), social workers (55%; 95% CI, 45%-65%), and nurses (47%; 95% CI, 37%-58%), other health care professions were present in less than half of the cancer centers surveyed. Only one-third of palliative care program leaders identified their professional background as palliative care. Board certification was not a requirement for physicians or nurses in a majority of programs.

Table 5 provides further information about the structures and processes for the 4 main branches of palliative care services, including inpatient consultation teams, outpatient clinics, PCUs, and institution-operated hospices. Inpatient consultation teams were not only the most common service, but also served a larger proportion of patients than PCUs and outpatient clinics. The median duration from referral to death was 7 days (n = 49; IQR, 4-16 days) for inpatient consultation teams, 7 days (n = 8; IQR, 5-10 days) for PCUs, and 90 days (n = 11; IQR, 30-120 days) for outpatient clinics. A large minority of programs with an outpatient clinic reported that they see patients in oncology clinics.

The inpatient consultation teams, outpatient clinics, and PCUs were generally larger and served more patients at NCI cancer centers than at non-NCI cancer centers (Table 5). In contrast, NCI cancer centers were less likely to report an institution-operated hospice than non-NCI cancer centers and had a smaller median daily hospice census.

National Cancer Institute cancer centers were more likely than non-NCI cancer centers to offer a palliative care fellowship or research programs (Table 6). The fellowship programs were generally small, with few having 5 or more clinical fellows, and even fewer with research fellowships. Almost half of the responding programs provided training for physician assistants. In NCI cancer centers, palliative care was a mandatory rotation for oncology fellows in a minority of programs. Less than half of the respondents had research programs in place, even for NCI cancer centers.

Our survey provides up-to-date information on the availability and degree of integration of palliative care services in US cancer centers. Despite significant growth in the number of palliative care programs during the past decade,^16-19 there remains much heterogeneity in the infrastructure and delivery of care in US cancer centers. Less than half of the palliative care programs were equipped with an outpatient clinic, PCU, or hospice, and even fewer had fellowship and research programs in place. Compared with non-NCI cancer centers, NCI cancer centers were more likely to offer palliative care services, particularly inpatient consultation teams and outpatient clinics. Patients with cancer were referred to palliative care late in the disease trajectory. Further efforts to consolidate existing infrastructure and to integrate palliative care in cancer centers are warranted.

Although a great majority of cancer centers reported the presence of a palliative care program, many of these programs were limited to providing inpatient consultation services and only few centers had a PCU. Although palliative care for many inpatients can be managed effectively through the consultation service, those patients with severe distress requiring intensive palliative care are best managed in a PCU under the care of an interprofessional team.^20-22

Less than half of the cancer centers offered palliative care outpatient services, although NCI cancer centers reported having more clinics in place. Given that oncology care is provided predominantly on an outpatient basis, the relative lack of palliative care outpatient clinics is an important finding. Palliative care clinics have the potential not only to significantly increase the number of referrals, but also to promote early access to integrated care. A novel outpatient model of integration involves palliative care personnel seeing patients at oncology clinics, which can potentially increase referral and enhance coordination and communication between oncology and palliative care teams.²³ Twenty percent of centers reported that they have combined oncology-palliative care clinics. The exact nature and outcomes of these clinics need to be further delineated.

Almost all cancer center executives, who play a critical role in defining the future of cancer care, agreed that a stronger integration between palliative and oncology care is necessary. Early incorporation of principles of palliative medicine in cancer care has been shown to be associated with improved symptom control, quality of life, and transition to hospice care.^24-26 Conversely, delayed referral significantly limits the effectiveness of palliation.^27-29 Our survey revealed that palliative care patients were referred too late in the disease trajectory, a marker of limited access and integration.

How can we facilitate this important process of integration? First, health care professionals, patients, and their families need to be educated on the benefits of palliative care and the concept of simultaneous care.³⁰ This can be accomplished through longitudinal discussions, mass media, and integration of relevant materials into heath professionals' educational curricula, scientific meeting programs, and certification/recertification requirements.⁶ Second, increased availability of palliative care outpatient clinics can facilitate early referral. Third, oncologists should be encouraged to make early palliative care referrals and to participate regularly in family conferences and palliative care educational rounds, whereas palliative care specialists should start attending interdisciplinary cancer treatment rounds. Fourth, enhanced training of oncologists in core competencies related to palliative care can increase the overall quality of care and prompt more timely referrals to palliative care.⁶ Fifth, NCI should dedicate more resources toward research on integration models, with particular attention paid to the documentation of clinically important outcomes.³¹ The Institute of Medicine also recommended that “NCI should add the requirement of research in palliative care and symptom control for recognition as a Comprehensive Cancer Center.”⁴ Finally, as highlighted in the executive survey, financial constraints remain the major obstacle to effective delivery of palliative care. It is prudent to advocate for increased resources and health policy changes through political engagement at the state and national levels.

The future of palliative care depends on research and education. Our study revealed the troublesome finding that few cancer centers have research programs available, and less than half offer palliative care fellowship programs and mandatory palliative care rotations for medical oncology fellows. Given the self-perceived lack of palliative care training among oncologists,^32,33 standardization and consolidation of palliative care in oncology fellowships is essential. To meet the anticipated growth in palliative care, the infrastructures for both research and education need to be developed further.

Our study had several limitations. First, we sampled only a small number of cancer centers, with some missing values. Second, responses from our survey could be biased, because they were based on recalled information and approximated data. Third, we surveyed only cancer centers approved by the Commission of Cancer and did not include any nonaccredited hospitals. This may result in an overestimation of palliative care services in non-NCI cancer centers.¹² Fourth, there remains no gold standard for what constitutes a palliative care program, which could have affected how our respondents interpreted questions in the surveys. An operational definition of palliative care related to program infrastructure, staff qualifications, or both would be helpful. Finally, although we addressed the surveys to the cancer center executives and palliative care program leaders, their delegates might have filled out the surveys instead. Nevertheless, in delegating this task, we believe that the executives and program leaders identified the most appropriate representatives for completing the surveys.

Although most cancer centers now have a palliative care program, significant gaps and delays in the delivery of care remain. To fulfill American Society of Clinical Oncology's vision of full integration of palliative care as a routine part of comprehensive cancer care by 2020,⁶ we urgently need to consolidate infrastructure such as outpatient clinics and PCUs, increase training of palliative care professionals and oncologists, conduct research on novel integration models and quality improvement measures, educate patients and their families, and advocate for public health policy changes. Oncologists, palliative care specialists, executives, and political leaders need to work closely together to ensure access to high-quality palliative care for all patients with cancer in the United States.

Corresponding Author: Ahmed Elsayem, MD, Department of Palliative Care and Rehabilitation Medicine, Unit 1414, The University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Blvd, Houston, TX 77030 (aelsayem@mdanderson.org).

Author Contributions: Dr Hui had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Drs Hui and Elsayem contributed equally to the manuscript.

Study concept and design: Hui, Elsayem, De La Cruz, Berger, Zhukovsky, Fadul, Palmer, Bruera.

Acquisition of data: Hui, Elsayem, De La Cruz, Berger, Zhukovsky, Evans, Fadul.

Analysis and interpretation of data: Hui, Palla, Palmer, Bruera.

Drafting of the manuscript: Hui, Bruera.

Critical revision of the manuscript for important intellectual content: Hui, Elsayem, De La Cruz, Berger, Zhukovsky, Palla, Evans, Fadul, Palmer, Bruera.

Statistical analysis: Hui, Palla, Palmer.

Obtained funding: Bruera.

Administrative, technical, or material support: Hui, Elsayem, De La Cruz, Berger, Zhukovsky, Palla, Evans, Fadul, Bruera.

Study supervision: Elsayem, Bruera.

Funding/Support: This work was supported in part by grants RO1NR010162-01A1, RO1CA122292-01, and RO1CA124481-01 from the National Institutes of Health (Dr Bruera); grant CA16672 from the National Cancer Institute (Patient-Reported Outcomes, Survey and Population Research Shared Resources); and the Clinician Investigator Program, Royal College of Physicians and Surgeons of Canada (Dr Hui).

Role of the Sponsors: The funding organizations had no role in the design and conduct of the study, in the collection, management, analysis, and interpretation of the data, or in the preparation, review, or approval of the manuscript.

Additional Contributions: We thank the Patient-Reported Outcomes, Survey and Population Research core facility at The University of Texas M. D. Anderson Cancer Center for assistance with conduction of this survey, and Kathryn Hale, MS, MLIS (Department of Scientific Publications, The University of Texas M. D. Anderson Cancer Center) for review of the manuscript. Ms Hale did not receive any compensation for her contribution. We are also grateful to all respondents for completing the surveys.