|Home | About | Journals | Submit | Contact Us | Français|
With improved effectiveness of early detection and treatment, many patients with cancer are now living with advanced disease and associated symptoms. As cancer becomes a chronic illness, adequate attention to patients' symptoms and psychosocial needs in the community setting requires positioning of palliative care alongside cancer care. This article describes the current palliative care needs of a population of community-dwelling patients with advanced cancer who are not yet ready for transition to hospice.
This secondary analysis used quality-monitoring data collected in three community-based palliative care organizations. Analyses focused on people with cancer-related diagnoses who were receiving palliative care during 2008 to 2011.
The analytic data set included 4,980 people, 10% of whom had cancer. Median age was 71 years. Forty-eight percent had been hospitalized at least once in the 6 months before palliative care referral. Forty-nine percent had a Palliative Performance Score (PPS) of 40% to 60%; 40% had PPS ≤ 30%. Although 81% had an estimated prognosis of ≤ 6 months, 58% were expected to live weeks to months. Thirty-three percent had no identified healthcare surrogate; 59% had no do-not-resuscitate order despite declining functional status and limited prognosis. Ninety-five percent reported ≥ 1 symptom, and 67% reported ≥ 3 symptoms; a substantial proportion did not receive treatment for symptoms.
Patients referred to community-based palliative care experience multiple often-severe symptoms that have been insufficiently addressed. They tend to have declining performance status. Earlier palliative care intervention could improve outcomes but will require delivery models that better coordinate inpatient/outpatient oncology and community-based palliative care.
Cancer is evolving into a chronic illness; in 2007, almost 12 million Americans with a diagnosis of cancer at some point in their history were alive.1 This number, which represents almost 4% of the US population, has dramatically increased from the 3 million cancer survivors alive in 1971.2
A remarkable and expanding portfolio of antineoplastic and supportive care treatments, combined with early detection methods that identify cancers at a treatable stage, accounts for the fact that many individuals are currently living with advanced cancer.3 But despite this encouraging progress in survival, many cancer survivors experience the accumulating symptom burden, deterioration of function, and other concerns inherent to slowly progressive, life-limiting disease. Most of these individuals are living and receiving care “in the community,” that is, in their personal homes or other private homes, nursing homes, and retirement communities. Thus they confront the psychological, social, emotional, and spiritual issues related to advanced cancer, as well as burdensome cancer-related symptoms, in a context often far removed from a cancer center.
Traditionally, hospice (and palliative care, frequently equated with hospice) has assisted patients in the last days or weeks of life, when curative, disease-focused treatments have ceased and end-of-life care is needed.4 Under current reimbursement schemes, most people living with advanced cancer in the United States are not eligible for hospice, the dominant nonhospital palliative care mechanism, because they are receiving ongoing antineoplastic treatment and/or have a prognosis that does not preclude the potential for meaningful survival. However, because of growing appreciation that palliative care can improve the seriously ill patient's experience—especially when delivered much earlier than at the end of life, as targeted by hospice—the delivery model of palliative care and hospice is evolving (Figure 1). Early palliative care has demonstrated utility in improving survival,5 symptom burden,6 quality of life,7 costs,8 and caregiver burden.5–9 In the emerging delivery model for cancer care, palliative care is provided across the disease trajectory, alongside the care of an oncologist.
Now a recognized medical subspecialty, palliative medicine is rapidly evolving; an important aspect of its development is exploration of approaches to meet the needs of people with advanced cancer. In the United States, the growth of inpatient consultative palliative care was a critical first step in the evolution of palliative care.10 Today, the vast majority of patient encounters with nonhospice palliative care in the United States occur in acute care settings, ranging from emergency departments to hospital intensive care units. The focus of these inpatient consultative palliative care services includes symptom control, advance care planning, and care transitions.
Inpatient consultative palliative care is, by definition, a hospital-based model. It lacks a well-developed structure to support longitudinal, postdischarge, follow-up care by palliative care experts of patients with complex needs. Inpatient palliative care often, legitimately, focuses on symptom management related to the acute issues that led to hospitalization, at the expense of important outpatient care needs such as medication reconciliation, general symptom management, and discussion of goals, as well as important future needs such as outpatient planning, transition management (eg, from hospital to rehabilitation to home), and proactive/preventive symptom management to prevent emergency room visits and hospital admissions. Inpatient providers are challenged to anticipate home needs and caregiver concerns when they see a patient only in the hospital environment.
In response to the limitations of inpatient palliative care, new community-based palliative care models (eg, “bridging programs”) are starting to populate the palliative care landscape.11 Providing longitudinal, community-based, palliative care, these new models seek to meet the needs of the large and growing population of patients living with cancer who are well enough to be in home-like settings yet who experience the very real burdens of advanced and progressive illness. Community-based palliative care works in concert with standard cancer care to assess and address multidimensional symptoms. Together, this model also serves as a system for coordination of care, streamlining the transition from palliative to end-of-life care when disease-focused treatments are no longer reasonable or effective.
Given the projected increases in the numbers of cancer survivors, that is, people living with advanced cancer as a chronic disease, both oncology and palliative care are intensifying their focus on community-based palliative care. Questions currently under debate include, When should patients with cancer be referred for community-based palliative care? What type of workforce will best be able to care for these patients? What are appropriate reimbursement mechanisms? In practical terms, how can oncology and palliative care work together in the community setting?
To establish a foundation from which to consider these challenges, this article describes the status and palliative care needs of a population of community-dwelling patients with advanced cancer who were referred to outpatient palliative care consultative services.
The Carolinas Palliative Care Database Consortium12 is a community/academic partnership between Duke University and three community-based palliative care organizations in North Carolina (Four Seasons, Flat Rock; Forsythe Medical Center, Winston-Salem; Hospice of Wake County and Horizons Palliative Care, Raleigh; and Charlotte Region Hospice and Palliative Care, Charlotte). Since 2008, Consortium members have been systematically collecting quality improvement data for the purpose of understanding patients' needs and informing practice change. Sites represent diverse practice venues (urban, suburban, rural), patient demographics (socioeconomic status, racial diversity), medical cultures (prevalence of medical specialists, proximity to academic medical center), and palliative care delivery models (eg, community based, bridging, inpatient consultative, hospice affiliated).
Quality improvement data were collected by palliative care clinicians at the point of care using the Quality Data Collection Tool (QDACT) Version 1.0, a palliative care–specific needs assessment tool developed by the Consortium.12,13 Demographic data, including patient age, sex, and race, were obtained from local administrative data sets, supplemented by clinician data entry. To supply clinical data, clinicians entered the primary medical condition; diagnoses were confirmed by comparing with International Classification of Diseases (ninth edition; ICD-9) codes submitted for the encounter. The McCorkle Symptom Distress Scale,14 considered the original gold standard for symptom measurement in patients with cancer, was the foundation for symptom queries; questions used four-point Likert scales ranging from “not a problem” (0) to “severe problem” (4). In alignment with prior validation studies,15,16 any answer choice of “moderate” or “severe” was considered clinically significant; in most analyses “moderate” and “severe” were combined. The Palliative Performance Scale17 was used to document performance status and to aid in prognostication; the Palliative Performance Scale is derived from the Karnofsky Performance Status Scale, and categories of the two instruments have similar meanings (Appendix Table A1, online only). Pharmacologic and nonpharmacologic symptom interventions for pain, dyspnea, depression, and constipation were documented by the palliative care provider. Quality of life was qualitatively categorized as “poor,” “fair,” or “good,” using a global quality-of-life scale derived from the McGill Quality of Life Questionnaire.18
Data from patient encounters between June 1, 2008 and March 1, 2011 were extracted in de-identified aggregate from the Carolinas Consortium Palliative Care Database. To ensure that results reflected community-based nonacademic palliative care, any records from Duke University Health System were excluded. Analyses were limited to people with cancer-related diagnoses (ICD-9 140-239.9) and to the initial consultation so as to include patient care needs before changes initiated by the palliative care provider. Duplicate records, and those without complete demographics to satisfactorily identify duplicates, were excluded. Basic descriptive statistics were used for analyses, which were conducted using Stata/IC (version 11, StataCorp, College Station, TX).
The analytic data set included 4,980 people receiving palliative care in three community settings in North Carolina; 513 (10%) of these individuals had cancer (Table 1). Median age was 71 years, and most (67%) were over age 65. Half (48%) of patients had been hospitalized at least once in the 6 months before palliative care referral. Half (49%) of patients had a Palliative Performance Score of 40% to 60% indicating a need for some assistance with most activities of daily living; 40% of patients had a Palliative Performance Score ≤ 30%, reflecting severe debilitation and a need for considerable assistance. Most (81%) had a provider estimated prognosis of 6 months or less, but the majority (58%) were expected to live only “weeks to months.” Typically, a patient with cancer with a solid tumor and Palliative Performance Status of 50% has a 3-month prognosis; thus, the majority of included patients were being seen not at the end of life but rather during a 3-month window of failing function, with impending death in the foreseeable future. Consistent with this scenario, most (88%) patients reported “poor” or “fair” quality of life.
Regarding goals of care and planning, one-third (33%) of patients did not have an identified healthcare surrogate at the time of consultation. The Medical Orders for Scope of Treatment (MOST) form, the legally adopted, preferred, portable physician order directing care at the end of life in North Carolina, was not completed for 97% of patients at the time of referral to palliative care. Despite poor functional status and limited prognosis, 59% of patients were “full code” (ie, had no do not resuscitate order).
With only 10 possible symptoms that could be recorded using the QDACT v1.0 interface, nearly all (95%) patients with cancer had at least one reported symptom, and two thirds (67%) reported three or more symptoms (Figure 2). Review of interventions prescribed before referral to palliative care demonstrates that a substantial proportion of patients were not receiving treatments aimed at addressing their symptom concerns; for example, 75% of patients reporting moderate/severe constipation were not being treated for this symptom.
The evolution of cancer into a chronic disease, and the correspondingly accumulating burden of advanced cancer on patients in the community, demand new approaches to meet patient needs. Oncology and palliative care are logical partners in this endeavor. How can these two disciplines most effectively intersect?
This snapshot of current populations and practices in community-based palliative care suggests that people are frequently referred to such care when their physical performance status begins to decline. Referrals encompassed a large range of anticipated survival and frequently extended far past the typical hospice period of 6 months or less, showing both patient and physician demand for palliative care services before hospice. Severe symptoms were common (67% in this sample), and most people had at least one clinically significant, uncontrolled symptom; for example, depression and constipation, two symptoms for which treatments are available and familiar, were not addressed in more than 75% of patients before referral to palliative care.
A clear need for better discussion of advance care planning and goals of care was found. Partnership between palliative care and oncology in the community setting might help to foster these conversations. In this sample, most patients' records did not document discussions regarding resuscitation and goals of care. This lack raises suspicion that other goals not traditionally documented in the medical record (eg, transition planning if condition changes) may likewise not have been addressed. Deteriorating health and functional status, apart from symptoms alone, is a common contributor to distress in patients with advanced illness.19 Worries about the future, flagging hope, and social needs have been identified as the most burdensome issues facing patients with cancer.20 Therefore, improved advance care planning that effectively addresses these concerns, especially when patients' functional status begins to deteriorate, will require early multidisciplinary involvement.
This article provides one of the first depictions of the use of palliative care by patients with advanced cancer living in the community. Consistent with expectations, these patients are being referred to community-based palliative care when they have limited prognoses, generally in the range of months to less than a year. They present with a large portfolio of unmet needs, indicating a need for more upstream involvement of palliative care concurrent with standard cancer care. Support for earlier palliative care intervention is increasing; ASCO recently published its vision of full integration of palliative care into comprehensive cancer care by 2020.21 Addressing the needs of patients with cancer needs across all settings is paramount in order to meet this goal, but it is not yet clear how primary palliative care (ie, care provided by oncologists) and secondary palliative care (ie, care provided by palliative care specialists) can best be coordinated.
Two major barriers to the integration of palliative care into cancer care are late referrals and inconsistent availability of clinic-based palliative care. For example, in this cohort, only 10% of the people being seen in the included palliative care programs had cancer as a primary diagnosis; by contrast, 43% of patients receiving hospice care have cancer. The low proportion of patients with cancer in community-based palliative care may suggest that oncologists have a difficult time deciding to refer.22 Late referral has been described as a national problem23 that delays the proven patient and family benefits of palliative care.24,25 In our data set, the vast majority of patients were hospice eligible, according to both performance status and estimated life expectancy, on consultation. Although somewhat expected, as up to one-third of patient with cancer die within 6 months of diagnosis23 (ie, many patients with cancer are hospice eligible from the outset), this pattern again highlights the need to bring palliative care more upstream, to benefit the other two thirds.21 In this study, we did not document the concurrent cancer therapies being provided, but clinicians reported that the majority of patients were also receiving cancer treatments that included radiation and systemic antineoplastic therapies; hence, these patients were not yet appropriate for hospice, though most were well qualified for palliative care.
For patients with advanced cancer, when is the appropriate time to refer for palliative care consultation? There is no simple answer. Triggers for consultation are needed, and barriers that impede integration of palliative care into cancer care must be identified and addressed.
Consultation triggers or thresholds have been increasingly studied. Outside of oncology, repeat hospitalizations are often used as an indicator for referral26; in our population of patients with cancer, however, hospitalization-based triggers may not be sufficient as fewer than one quarter of the included patients had two or more prior hospitalizations. A recent consensus report from the Center to Advance Palliative Care suggests the following as consultation triggers for hospital-based palliative care: frequent hospital admissions, admissions for difficult-to-control physical or psychological symptoms, complex care requirements, decline in function, or failure to thrive. These global indicators should prompt screening for unmet palliative care needs,26 but they are inpatient approaches; outpatient flags that indicate palliative care consultation is needed may be more difficult to define.
In the outpatient setting, triggers for referral to community-based palliative care could be simple, such as diagnosis of metastatic disease in lung cancer, consistent with the eligibility criteria used in the study by Temel et al.5 Prognosis-based referral is a practical approach, but prognosis may not reflect an actual need for palliative care services.27 Performance status may be more closely related to need, and thus decline in performance status may offer a better trigger for referral to community-based palliative care. In our analysis, most patients were referred when performance status had started to wane; 88% had documented degradation in Palliative Performance Status. As demonstrated in Figure 1B, this pattern is consistent with referral to palliative care services at a time of declining function. Use of performance status-based triggers for palliative care will require continuous performance status monitoring, conducted in a standardized way as a usual part of oncology care. Accumulating symptom burden or decrease in quality of life may also be used as triggers for palliative care referral, but systematic documentation of these patient-reported concerns may be more difficult in many community oncology settings. Continuous electronic patient-reported outcomes monitoring is becoming more prevalent, and in the future may support resolution of this problem.28–30
The lack of palliative care availability in the outpatient setting—where cancer care largely takes place—may be a leading barrier to integration of palliative care earlier in the disease trajectory. Hui et al recently reported a survey showing that only 20% of non–NCI-designated cancer centers, which are often in community settings, have an outpatient palliative care program. The value of outpatient palliative care services for patients with cancer with complex needs has been demonstrated,31,32 but these studies have been primarily restricted to academic cancer centers with access to large pools of personnel and other instrumental resources. As new studies report the demonstration of care models outside of large cancer centers11 and integrated within cancer pain clinics,33 a workforce delivery problem for secondary palliative care is emerging.
Workforce planning is a critical concern for oncology and palliative care as they consider, together, how best to care for the increasing population of patients living with advanced cancer as a chronic disease (Figure 1C). Uncertainties regarding who, specifically, will provide community-based palliative care to patients with cancer cause considerable distress in both disciplines. Symptom control is a clear need for this population; most patients with advanced cancer experience seven or more uncontrolled symptoms34 as well as low quality of life at the time of referral.35 Providers of community-based palliative care must be prepared to address a substantial concurrent constellation of symptoms using evidence-based pharmacologic and nonpharmacologic strategies. Multidisciplinary teams that are prepared to address early goal setting and advance care planning, as well as a range of physical and psychological symptoms, will be needed. Many other practical issues must be addressed in order to provide adequate care for the growing number of people experiencing cancer as a chronic illness in the community. For example, new reimbursement mechanisms that allow for effective and efficient care must be developed; the current Medicare benefit funds hospice services but not palliative care. In addition, communication channels that coordinate palliative care and oncology, especially when palliative care providers are seeing patients at home, will be critical to the provision of best care.
There are many limitations to this type of analysis. First, data were collected primarily for the purpose of quality improvement and were used secondarily for this exploratory, descriptive study. The data we used thus do not contain the breadth of variables often collected and reported for hypothesis testing in research. For example, symptom prevalence was limited to 10 key symptoms that are either referenced in palliative care quality measures or included in well-validated screening tools. Second, the analytic data set included only people who were seen for subspecialty community-based palliative care consultation; it does not reflect the status or needs of people who were not referred to palliative care, whether because they did not have any perceived unmet needs, because community-based palliative care was not available to them, or because their cancer care providers were unaware of clinical services. Hence our analysis provides an incomplete picture of palliative care needs in the community. Third, standardized data collection procedures are still in development in palliative care. Because some patients will have persistent care needs after the first palliative care visit, our estimates of the needs and unaddressed symptom burden of patients with cancer are likely underreported. Nevertheless, collecting data within palliative care, both in academic and research environments, as well as in historically underrepresented community and rural settings, is vital to both the improvement of patient-centered care and the development of service delivery models needed to implement this care.36 We and others37 have demonstrated the feasibility and value of standardized data collection; a recently formed research collaborative in the United States is building infrastructure38 to facilitate this task.
Cancer is becoming a chronic disease. Among community-dwelling patients with advanced cancer, the burden of uncontrolled symptoms and the prevalence of inadequate advance care planning are high. Understanding the palliative care needs of community-based patients in the advanced stages of cancer represents a first step toward the development of delivery models that integrate care across settings (inpatient, outpatient), and that coordinate community oncologists and palliative care providers to meet identified needs of this population. Areas for further work include development of consultation triggers that result in earlier integration of palliative care into oncology care, and novel approaches for ensuring the availability of community-based palliative care at the time when it is needed.
Although all authors completed the disclosure declaration, the following author(s) indicated a financial or other interest that is relevant to the subject matter under consideration in this article. Certain relationships marked with a “U” are those for which no compensation was received; those relationships marked with a “C” were compensated. For a detailed description of the disclosure categories, or for more information about ASCO's conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors.
Employment or Leadership Position: None Consultant or Advisory Role: Janet Bull, Salix Pharm-Scientific Advisory Board (C), Archimedes-Scientific Advisory Board (C); Amy P. Abernethy, Helsinn (U), Amgen (U), Novartis (U), Pending: BMS (U) Stock Ownership: None Honoraria: Janet Bull, Salix Pharm-Speakers Bureau, Pfizer-Speakers Bureau Research Funding: Amy P. Abernethy, Pfizer, Lilly, Novartis, Pending: Endo Holdings, BMS, Helsinn, Amgen, Kanglaite, Alexion, Biovex, DARA, MiCo Expert Testimony: None Other Remuneration: None
Conception and design: Arif H. Kamal, Amy P. Abernethy
Financial support: Donald H. Taylor Jr, Amy P. Abernethy
Administrative support: William Downey
Provision of study materials or patients: Arif H. Kamal, Amy P. Abernethy
Collection and assembly of data: Arif H. Kamal, Dio Kavalieratos
Data analysis and interpretation: Arif H. Kamal, Janet Bull, Dio Kavalieratos, Donald H. Taylor Jr
Manuscript writing: Arif H. Kamal, Janet Bull, Dio Kavalieratos, Amy P. Abernethy
Final approval of manuscript: All authors