Early palliative care and metastatic non-small cell lung cancer

Abstract

Patients with advanced cancer experience a significant burden of physical symptoms and psychological distress at the end of life, and many elect to receive aggressive cancer-directed therapy. The goal of palliative care is to relieve suffering and promote quality of life (QOL) for patients and families. Traditionally, both the public and medical community have conceptualized the need for patients to make a choice between pursuing curative therapy or receiving palliative care. However, practice guidelines from the World Health Organization and leadership from the oncology and palliative care communities advocate a different model of palliative care that is introduced from the point of diagnosis of life-threatening illness. Early palliative care has been shown to provide benefits in QOL, mood, and health care utilization. Additionally, preliminary research has suggested that in contrast to fears about palliative care hastening death, referral to palliative care earlier in the course of illness may have the potential to lengthen survival, particularly in patients with advanced nonsmall-cell lung cancer. This review summarizes the literature on potential survival benefits of palliative care and presents a model of how early integrated palliative care could potentially influence survival in patients with advanced cancer.

Introduction

The American Society of Clinical Oncology recently issued a provisional clinical opinion that patients with metastatic cancer should be offered concurrent palliative and standard oncologic care from the time of diagnosis. ¹ This recommendation is based on evidence from randomized controlled trials (RCTs) demonstrating that palliative care is associated with improvements in quality of life (QOL), mood, caregiver distress, and changes in patterns of care at the end of life (EOL). While the provisional statement did not include improved survival in the list of benefits, some research has shown that palliative care interventions may also have a positive impact on survival. In this review, we summarize the literature regarding possible survival benefits from palliative care, present a conceptual model of how early palliative care (EPC) might influence survival and then explore hypotheses derived from a model in patients with metastatic nonsmall-cell lung cancer (NSCLC).

Methodological considerations

Definition of terms

Palliative care is a specialized medical care for patients with life-threatening illness. In contrast to hospice, the delivery of palliative care does not require patients to have an estimated survival of 6 months or less. The goal of palliative care, according to the National Consensus Project on Palliative Care, “is to prevent and relieve suffering and to support the best possible quality of life for patients and their families, regardless of the stage of the disease or need for other therapies.” ² Traditionally, in cancer care, patients with advanced disease are treated by their oncologists until late in the course of illness, at which point an abrupt transition to palliative care occurs close to death. Early, integrated palliative care is a model in which palliative care is introduced near the time of diagnosis along with standard cancer care and the involvement of palliative care increases over time until the EOL. This newer model is supported by the World Health Organization which emphasizes that palliative care should ideally be provided early in the course of life-threatening illness and extend throughout the continuum of care. ³

Palliative care for patients with cancer has moved beyond inpatient consults at times of crisis near the EOL. Clinicians are establishing new models of care including ambulatory palliative care clinics embedded in oncology centers in which patients can develop relationships over time with palliative care providers. This longer term relationship allows for the management of physical and psychological symptoms throughout the course of disease, facilitation of understanding of prognosis and goals of care, assistance with treatment decision making, and coordination of care. Recent guidelines for optimal palliative care emphasize the importance of an interdisciplinary team approach, close coordination with the oncology or primary medical team, and the benefit of being able to engage in direct patient care, while acknowledging the need to adapt the intervention in a way that does not add burden to patients or families. ^{2 –4}

Limitations of the current research

Multiple meta-analyses have highlighted the challenge of synthesizing research on palliative care because of variable definitions of this service and choice of outcome measures. Many studies have lacked adequate power to detect clinically meaningful changes in QOL or symptoms, and the majority of trials to date were not designed to assess the potential relationship between EPC and survival a priori. Additionally, patient recruitment has been limited considering that many studies have relied on referrals rather than employing a systematic screening process. ^1,5,6

Relevant RCTs

Of the existing literature, we identified four trials of palliative care interventions for patients with cancer that were randomized, adequately powered to detect differences in QOL or patterns of care at the EOL and included data on survival.

Trials of concurrent palliative care in the samples of patients with illness that included advanced cancer

Two studies were conducted in the samples of patients with mixed diagnoses including advanced cancer. Although palliative care was not provided from the time of diagnosis, it was delivered alongside curative care.

First, Brumley and colleagues conducted an RCT at two managed care sites that investigated the impact of referral to an in-home palliative care intervention compared with usual care. Physicians were asked to identify patients with an estimated 12-month life expectancy, and palliative care could be provided in conjunction with curative care. Home visits were made by an interdisciplinary team consisting of a physician, nurse, and social worker. Outcome measures included satisfaction with care, health care utilization, costs, and location of death. From 2002 to 2004, 310 patients were randomly assigned to usual care versus in-home palliative care. Approximately, one half of patients had cancer, and the patients in the rest of the sample had congestive heart failure and chronic obstructive pulmonary disease (COPD). Patients in the intervention group had higher satisfaction with their care at 1 month and 90 days (but not at 60 days), fewer emergency department visits, and fewer hospital days. Although patients assigned to palliative care were significantly more likely to die at home, the study groups did not differ in hospice use or length of stay. ⁷ Finally, Brumley and colleagues observed a nonsignificant trend toward decreased survival in the intervention group (median days survived = 196 days vs. 242 days in the control group, p = 0.08). The investigators hypothesized that patients in the intervention group may have had more severe disease, although they attempted to control for illness severity in their analysis. Additionally, they speculated that the differences observed in EOL care could potentially be related to survival in this sample.

Next, Gade and colleagues ⁸ conducted a multicenter, RCT comparing the impact of an interdisciplinary, inpatient palliative care service including a physician, nurse, social worker, and chaplain, with usual care in patients hospitalized with a life-limiting illness. Patients were eligible to participate if their attending physician said they would not be surprised if they died within 1 year. Primary outcomes included symptom control, emotional and spiritual support, satisfaction, and health care costs. Secondary outcomes included survival, number of advanced directives completed by discharge, and use of hospice within the next 6 months. The investigators randomly assigned 517 patients (27% with cancer diagnoses) to usual care versus interdisciplinary palliative care. Physical symptoms, QOL, mood, and emotional support did not differ significantly between groups. Although the likelihood of readmission to the hospital was similar, patients receiving palliative care had fewer intensive care unit stays compared to those in the usual care group. Rates of referral to hospice did not differ, but patients assigned to the intervention had longer median hospice stays (median days = 24 vs. 12, p = 0.04). Also, patients receiving palliative care were more likely to complete an advanced directive prior to discharge and reported higher satisfaction with hospital care. However, survival time did not differ significantly, with a median postenrollment survival 30 days in the intervention group compared to 36 days in the usual care group (p = 0.08). ⁸

Trials of EPC in patients with advanced cancer

The following two RCTs were designed to assess the impact of EPC provided within 8–12 weeks of diagnosis and included only patients with advanced cancer.

The first RCT was a nurse-led palliative care intervention (Project ENABLE) conducted in 322 patients newly diagnosed with advanced cancer including patients with NSCLC. ⁹ From 2003 to 2007, enrolled patients were randomly assigned to usual care or a telephone-based intervention led by advanced practice nurses in addition to usual care. Specifically, nurses conducted weekly telephone sessions focused on education and problem solving regarding communication and social support, symptom management, and advanced care planning. Until the time of death or the end of the study period, the nurses followed up at least monthly, and patients and caregivers were invited to attend monthly group shared medical visits led by a palliative care physician and nurse practitioner. The aim of the intervention was to enable patients and caregivers to be more actively involved in their own care. The primary outcomes included QOL, symptom intensity, and health care resource use. Mood was a secondary outcome. The study demonstrated higher QOL and decreased depressed mood in the intervention group. However, there were no significant changes in symptom intensity or health care utilization. Post hoc analyses showed a nonsignificant trend toward increased survival in the intervention group (14 months vs. 8.5 months in the usual care group, p = 0.14.).

Temel and colleagues conducted a phase III RCT of 151 patients newly diagnosed with metastatic NSCLC at an academic medical center. Ambulatory patients were recruited from a thoracic oncology clinic and assigned to standard oncology care or EPC within 8 weeks of diagnosis. The primary outcome of this trial was QOL at 12 weeks as measured by the Functional Assessment of Cancer-Lung Trials Outcome Index. Secondary outcomes included mood (specifically anxiety and depression), understanding of illness, and aggressiveness of care at the EOL. This trial demonstrated statistically and clinically meaningful improvements in QOL and depression at 12 weeks. Additionally, patients assigned to the EPC group had a more accurate understanding of prognosis, a higher rate of documentation of resuscitation preferences and less aggressive care at the EOL, including decreased intravenous (IV) chemotherapy within 2 weeks of death, and earlier referrals to hospice. Although survival was not a prespecified outcome, patients in the EPC group had a median survival of 11.6 months compared to 8.9 months in patients who received standard oncology care (p < 0.02). ^10,11

Together, these four trials support the understanding that patients who receive EPC can attain improved QOL without any loss of quantity of life. While the two trials conducted in samples with mixed diseases did not find a survival benefit from palliative care, the studies conducted exclusively in patients with metastatic cancer showed that patients receiving EPC lived longer. EPC may have different effects on survival in different populations of patients. However, survival was not a primary end point in these studies and the findings have to be replicated.

Potential-mediating factors

Palliative care interventions have been shown to impact several variables, some of which have been associated with increased survival. Although data are currently lacking on the mechanisms by which EPC may lead to longer survival, the potential-mediating factors can be organized into a hypothesized model illustrated in Figure 1.

Figure 1.

This framework is informed by the Health Decision Model, which combines decision analysis, behavioral decision theory, and health beliefs in a unifying theory to explain patients’ health decisions and resulting behaviors that influence short- and long-term medical outcomes. ¹² For example, a patient’s willingness or desire to seek aggressive care at the EOL will be determined in part by his or her preexisting perceptions regarding the efficacy of such interventions for prolonging survival, regardless of actual statistical data provided by an oncologist. Health beliefs, such as acceptance of diagnosis, perceptions of illness severity and likelihood of cure as well as satisfaction with patient–physician relationship, make a substantial contribution to patient cooperation with treatment plans.

We hypothesize that early, integrated palliative care (EPC) directly targets the components of the Health Decision Model by assisting patients with decision analysis, assessing preexisting health beliefs and preferences and providing detailed and accurate information regarding illness severity and prognosis. As a result of this intervention, patients may experience improved QOL and mood, develop a more realistic understanding of disease and goals of therapy, and enhance adaptive coping behaviors, all of which in turn, can impact treatment adherence and EOL decisions. Survival may be mediated not only by positive changes in QOL and mood but also by more proximal disease-specific health behaviors, such as optimal use of anticancer therapy and hospice services at the EOL.

Drawing on the existing literature of patients with cancer and other medical populations, we review the impact of EPC on components of this model and discuss the relationship to a potential survival benefit.

Improved QOL and control of physical symptoms

Palliative care aims to prevent and relieve suffering from physical and emotional symptoms and promote the best possible QOL. Health-related quality of life (HRQOL) encompasses both physical well-being, including the severity and distress caused by physical symptoms, and the ability to function physically, psychologically, socially, and spiritually. QOL has been identified as an independent prognostic factor for survival in patients with many types of cancer, including NSCLC. Individual physical symptoms have also been associated with survival, although this finding is less robust across studies. ^13,14 By directly impacting both overall QOL and ameliorating physical symptoms, EPC could potentially influence survival.

HRQOL and survival

Multiple studies have demonstrated a strong association between patient-reported QOL, measured at the time of diagnosis, and survival in samples of individuals with cancer. HRQOL has been found to independently predict survival in studies of individual cancers and in large meta-analyses which have pooled studies across different cancer samples. In particular, the physical and functional domains of QOL have been consistently associated with survival. ^13,15

Including HRQOL and the experience of physical symptoms, measured at diagnosis, has added prognostic value in patients with lung cancer in multiple studies. ^{16 –19} A recent study of more than 2000 patients with NSCLC demonstrated a clear association between overall QOL, measured at a single time point within 6 months of diagnosis, and overall survival (p < 0.001). This association remained significant after controlling for performance status, age, and disease stage. ¹⁹ In a study of more than 500 elderly patients with NSCLC, self-reported QOL and independence with intermediate activities of daily living, such as driving and preparing meals, were also independently associated with survival. ¹⁸

In addition to the value of a single measurement of baseline QOL, research has suggested the utility of monitoring change in QOL as a predictor of survival in patients with advanced NSCLC. Eton and colleagues hypothesized that change in patient-reported HRQOL and physical symptoms (using the Functional Assessment of Cancer Therapies-Lung) and trial outcome index (TOI), which incorporates physical well-being, functional well-being, and common physical symptoms, would be related to survival in a study of 599 patients with NSCLC receiving first-line chemotherapy. In this sample, patients with higher baseline physical well-being showed better response to chemotherapy and lower risk of death. Higher baseline TOI and direction of change in TOI over time were also associated with the risk of disease progression. These relationships remained significant after accounting for performance status and stage of disease. ¹⁶ The authors suggest that measurement of physical well-being and TOI add value at diagnosis by improving our ability to predict survival. Further, following the TOI longitudinally may prompt ongoing discussions about continuing or changing treatment plans. ¹⁶ Evaluating QOL at the time of diagnosis of cancer and early in the treatment course may also help clinicians to identify subpopulations of patients with high potential to benefit from interventions such as EPC.

Physical symptoms and survival in NSCLC

Patients with advanced lung cancer experience frequent and severe physical symptoms. In a community sample of patients with locally advanced lung cancer referred for consideration of radiation, the great majority experienced fatigue (80%), dyspnea (73%), loss of appetite (65%), and chest pain at distressing levels. The frequency and intensity of symptoms increased as patients approached death (p = 0.001). ²⁰

Some researchers have found associations between particular physical symptoms and survival, although results have not been consistent across all studies. In two studies of patients with NSCLC, patient report of pain independently predicted survival in a multivariate model that accounted for overall QOL and performance status. ^17,21 Rather than examining individual physical symptoms, a more clinically relevant approach may be to consider clusters of symptoms that occur together. The great majority of patients with advanced lung cancer (79% in the sample described above) experience more than three disabling physical symptoms at one time. ^20,22,23

Palliative care: impact on QOL and physical symptoms: focus on RCTs including NSCLC patients

QOL encompasses more than physical symptoms, it includes not only physical well-being but also functional, psychological, social, and spiritual well-being. Researchers have hypothesized that to impact overall QOL in a clinically meaningful way, interventions may need to be sufficiently comprehensive to address all of these domains.

Rummans and colleagues designed an intervention for patients with advanced cancer referred to radiation therapy. Patients participated in groups conducted by an interdisciplinary team of providers with expertise in physical, psychological, social, and spiritual functioning. Patients in the intervention group were able to maintain or improve their QOL at 4 weeks, whereas patients in the usual care group had a significant decline in QOL during this period. ²⁴ Interestingly, palliative care, particularly EPC delivered by an interdisciplinary team, may address these core elements of QOL. By incorporating the expertise of physicians, nurses, social workers, and chaplains, a palliative care team may be able to target physical symptoms and functioning comprehensively and flexibly throughout the course of illness. Given the relationship between QOL, symptom burden, and survival discussed above, EPC may have the potential to affect survival outcomes.

The majority of well-designed, adequately powered studies of palliative care that have specified overall QOL as a primary outcome have shown improvement. ^1,5,24 However, palliative care interventions have demonstrated mixed results regarding improvement in physical symptoms. Of the two RCTs of EPC in patients with advanced cancer, both the Project ENABLE and Temel et al. studies demonstrated significant improvements in QOL. The Temel et al. study, which used a model of outpatient palliative care visits, found significant improvement on the Functional Assessment of Cancer Therapy–trial outcomes index (FACT-TOI) scale, which incorporates subscales for physical well-being, functional well-being, and a cluster of common symptoms experienced by patients with lung cancer. In contrast, the Project ENABLE study, which used a nurse-led telephone intervention with fewer opportunities for direct assessment and management of symptoms, did not detect significant improvement in physical symptoms.

Other RCTs of palliative care also point to the importance of providing direct patient care and ideally longitudinal outpatient follow-up to achieve clinically meaningful improvement in physical symptoms. For example, in one study assessing referrals to an inpatient palliative care service, the investigators hypothesized that the average length of stay in the hospital (5 days following initial palliative care consultation) was not long enough to impact the symptoms significantly. ⁸ An RCT examining the effect of referral to an inpatient palliative care service, which focused on chronically ill patients (22% with cancer diagnoses), utilized a model in which the palliative care physician made recommendations to the inpatient attending. Study investigators observed, based on data gathered from the same institution, that recommendations to prescribe opiates and antidepressants were implemented by the inpatient attending less than 20% of the time. ²⁵

Further, in an RCT of an outpatient palliative care intervention conducted with patients diagnosed with life-limiting illnesses (approximately 1/3 of whom had cancer), the palliative care intervention was associated only with improvements in dyspnea and anxiety and failed to impact pain and depression significantly. In this study, primary care providers prescribed the recommended opiates in 8% of patients and antidepressants in 18% of patients. These data highlight the importance of palliative care clinicians providing direct patient care and improving communication and care coordination with the primary medical team. The investigators also suggest that to improve overall QOL, an intervention may need to target pain and depression. ²⁶

Additionally, a randomized controlled pilot study was recently conducted that evaluated the impact of an EPC intervention implemented by an oncology nurse practitioner compared with usual care. Specifically, the nurse practitioner met with patients who had been newly diagnosed with metastatic cancer to explain the benefits and role of hospice and assist with completion of advanced directives. Outcome measures included QOL measured by the Functional Assessment of Cancer Therapy–General (FACT-G) scale. This study closed early to accrual following the results of the Project ENABLE study discussed previously which demonstrated significant improvement in QOL in patients with advanced cancer who were randomized to a nurse-led intervention. ⁹ The investigators concluded that it was unethical to have a control group without access to EPC. In the small sample of patients accrued between 2008 and 2009 (12 in the intervention group and 14 in the control group), investigators found a statistically significant benefit in the emotional domain of QOL as measured by the FACT-G in the group of patients who received EPC. None of the other domains of QOL were significantly different between groups. The investigators suggest that an intervention led by a nurse practitioner embedded in an oncology team may have the potential to improve the emotional well-being of patients with advanced cancer. ²⁷

Moore and colleagues assessed the effectiveness of nurse-led follow-up compared to standard oncology care in 203 patients with lung cancer, treated at a specialist cancer hospital and three local cancer units in England, who had completed the active phase of their treatment and were expected to survive at least 3 months. Although not targeted at patients with metastatic disease, the intervention was designed to improve QOL and physical symptoms. The nursing intervention incorporated regular telephone assessments by clinical nurse specialists, weekly open access clinics, and emphasized close communication and coordination with the primary oncology team. Outcome measures included QOL, overall survival, symptom-free survival, progression-free survival, and resource use. Overall QOL did not differ significantly between groups, although the intervention group reported decreased dyspnea at 3 months and improved emotional well-being at 12 months. Nurses recorded progression of symptoms in the intervention group sooner than doctors who provided standard care (p = 0.01), although there were no differences in the rates of objective progression of disease or survival. ²⁸ Additionally, an RCT of a home care nursing program was conducted in a sample of 164 patients on oral chemotherapy for colorectal or breast cancer expected to live more than 6 months. Patients in the intervention group experienced significantly decreased toxicity due to chemotherapy, although QOL did not differ between groups. ²⁹

Potential mechanisms

EPC may impact survival by improving patient’s well-being, comprehensively targeting QOL, symptom burden, and depression, by assisting with treatment decision making, and by increasing social support. While multiple studies demonstrate the independent relationship between increased survival and HRQOL, including the domain of physical symptoms, the direction of this relationship is not yet clear. Self-reported HRQOL may reflect aspects of the patient’s physical well-being and functioning that are not detected by physician measures such as performance status. Additionally, patients may detect changes in physical symptoms sooner and at milder levels of distress than their oncologists. ^13,14 Self-reported measures may enhance the clinician’s understanding of the patient’s experience and capture variables that may also be associated with survival. However, there is the still the question of whether improving QOL directly leads to longer survival.

Through comprehensive medical management, EPC may promote survival by improving management of symptoms due to medical comorbidities. Conditions such as congestive heart failure and COPD can affect the ability of the patients to tolerate chemotherapy, radiation, or surgery. Given frequent oncology appointments and physical debilitation, patients with advanced lung cancer may make fewer visits to their primary care providers. Oncologists tend to focus on cancer treatment and may overlook other comorbid conditions. Palliative care clinicians focus more explicitly on the assessment and treatment of all symptoms that cause suffering which often include symptoms from other illnesses. However, even though patients with NSCLC typically have medical comorbidities, almost all of these individuals tend to die as a result of their cancer rather than these competing causes, given the short median survival.

Symptom management provided by EPC could also impact survival by helping to alleviate the side effects of chemotherapy, enabling patients to tolerate more chemotherapy. Greer et al., in a subanalysis of the Temel et al. study, examined this question, although the authors did not find support for this hypothesis. Specifically, patients assigned to EPC and standard oncology care received equal numbers of overall chemotherapy regimens and had equal intervals between first-line, second-line, and third-line chemotherapy. ³⁰ Yet, given that the investigators did not measure actual relative dose intensity of chemotherapy or dose delays or reductions, it is possible that they were unable to detect clinically meaningful differences in how much chemotherapy was received. Better adherence to a chemotherapy regimen has been shown to improve survival in patients with metastatic NSCLC. ³¹ Managing the side effects of chemotherapy may have allowed patients to receive chemotherapy with fewer dose reductions and delays.

Finally, changes in QOL may have biologic significance. An intervention that improves QOL and reduces chronic stress may affect immune function and has the potential to impact tumor growth and recurrence. ^{13,32 –34} This hypothesis will be more fully discussed in the next section.

Decreased depression

Depression is common among patients with advanced cancer. Elevated depressive symptoms have been reported to occur in more than 1/3 of patients with advanced lung cancer. ³⁵ Lo et al. prospectively examined a group of patients with metastatic lung and gastrointestinal cancers, observing that greater than 20% reported at least moderately severe symptoms of depression that persisted over time and increased in the final months of life. Depressive symptoms in the final months of life were associated with high physical symptom burden, hopelessness, lower social support, and low spiritual well-being. ³⁶

Although previous research has been mixed, more recent studies and systematic meta-analyses identify depression as an independent and potentially clinically meaningful predictor of overall mortality in patients with cancer. The relationship between depression and mortality is not explained by other known risk factors for cancer-related mortality such as tumor stage, age, or performance status. ^37,38

One recent secondary analysis of a randomized trial of supportive–expressive therapy of patients with metastatic breast cancer showed that decreased depressive symptoms over the first year was associated with longer overall survival (54 months vs. 25 months) when measured out to 14 years. This effect was not mediated by the specific intervention. Authors highlight the need to measure depression at multiple moments to distinguish between a potentially healthy adjustment reaction and depression. ³⁹

EPC has the potential to treat depression, representing another possible mechanism by which EPC prolongs survival. Pirl and colleagues examined this question through a secondary analysis of the RCT of EPC versus standard care in 151 patients with metastatic NSCLC described above. ¹¹ Depression was assessed at baseline and at 12 weeks with the Patient Health Questionnaire-9, an instrument that has been used extensively to screen for depression in the medically ill. Patients who met criteria for major depressive syndrome at baseline had shorter median survival (5.4 months vs. 10 months, p < 0.001). Patients assigned to the EPC group had significantly greater improvement in depressive symptoms at 12 weeks than the standard oncology care group. Interestingly, patients in the EPC group were not prescribed more antidepressants or referred more frequently to mental health clinicians. Regardless, the observed survival benefit from EPC appeared to be independent of improvement in depressive symptoms. ⁴⁰ These data do not support the hypothesis that treating depression is the main cause of increased survival in patients with metastatic NSCLC receiving earlier palliative care. Impacting depression alone may not be enough to contribute to a survival benefit in this population.

Multiple mechanisms have been proposed to explain the relationship between depression and survival in patients with cancer. Similar to improvements in QOL, decreased depression may impact health behaviors and therefore impact adherence with chemotherapy, influence other behaviors linked to wellness such as exercise, and affect choices made about care at the EOL. From a biological perspective, depression has been shown to chronically activate the hypothalamic–pituitary axis. Chronic high levels of stress have been associated with higher baseline levels of cortisol and decreased adaptive immunity. ^{32 –34,41,42} Interestingly, in both animal and human studies, high levels of chronic stress have been shown to lead to a shift in the profile of T helper cells, which may lead to decreased antitumor immunity. The profile of T helper cells has also been shown to have prognostic value beyond TNM staging in colorectal cancer. ⁴³

A small study conducted in patients with cervical cancer examined the relationship between a phone therapy intervention, improved QOL, and biomarkers including the profile of T helper cells. Patients whose QOL improved after the intervention were found to have decreased morning serum cortisol and the hypothesized shift in T helper immune stance. ³³

Impact on illness understanding, goal setting, and patterns of care

Many patients with advanced cancer have poor understanding of their prognosis and overestimate their likelihood of survival compared to the estimates of their physicians. ⁴⁴ An accurate understanding of prognosis and goals of care may be associated with patients and caregivers, making treatment choices that are consistent with their wishes. ^{45 –47} EPC, which aims to heighten the understanding of prognosis and goals of care, may result in more informed decisions about care throughout the course of illness. These decisions might lead to care that could potentially lengthen survival.

Although the primary purpose of palliative care is assistance with decision making, few studies have addressed whether palliative care improves illness understanding. Temel et al. found that one third of patients with newly diagnosed metastatic lung cancer endorsed the inaccurate belief that their cancer is curable. Involvement in EPC was associated with improved illness understanding. Specifically, over time, patients were more likely to become accurate in their perception that their cancer was not curable. Importantly, patients in the intervention group who had accurate prognostic understanding of their illness were less likely to receive chemotherapy at the EOL. ¹¹ Additionally, a greater number of patients in the intervention group had their resuscitation preferences documented in the outpatient electronic medical record. ^10,11 A large, multisite prospective cohort study of patients with advanced cancer found that patients who had a discussion about their EOL care with their physician were less likely to receive aggressive care and experienced better QOL during their final week of life. ⁴⁸ Other trials specifically assessing the impact of EPC have also observed increased documentation of advanced directives in the intervention group, although that finding has not been consistent across all studies. ⁸

Chemotherapy at the EOL

A significant percentage of patients with advanced cancer continue to receive chemotherapy very late in their course of illness, close to death, and the rates of receiving aggressive chemotherapy at the EOL are increasing over time. ⁴⁹ Retrospective evaluation of the surveillance, epidemiology, and end results (SEER) Medicare database of nearly 8000 patients aged greater than 65 with advanced NSCLC showed no improvement in overall survival in patients who received chemotherapy during their last 2 weeks of life. ⁴⁹ Moreover, many believe that there may be a point in a patient’s course of decline when IV chemotherapy could be more likely to hasten death with its toxicity rather than to extend life. Stopping chemotherapy earlier before reaching this potential turning point might result in patients living longer.

In a retrospective chart review conducted at a Veteran’s Administration hospital, investigators examined patterns of care from 2002 to 2008 and found a trend toward more aggressive care at the EOL including more patients receiving chemotherapy during the last month of life. Consultation with the interdisciplinary palliative care team more than 2 weeks before death (available in the inpatient and outpatient setting) was associated with a trend toward decreased chemotherapy administration close to the EOL. ²⁴ In the trial by Temel and colleagues, patients assigned to EPC were significantly less likely to receive IV chemotherapy close to the EOL compared to those in the standard care group, although the likelihood of receiving oral chemotherapy was similar between groups. In addition, the interval between last chemotherapy infusion and time of death was longer for patients who received EPC. ³⁰

Hospice

Researchers investigated the association between hospice and survival in a retrospective cohort analysis of the SEER Medicare database including patients with six representative diseases, such as advanced lung cancer, in which physicians predicted death within the next 3 years. Across diagnostic categories, the investigators observed a small but significant survival benefit associated with referral to hospice and longer hospice stays. Of note, in the lung cancer group, those referred to hospice survived for an additional 39 days (p < 0.0001). ⁵⁰ An examination of the SEER database in almost 8000 patients over 65 years old with advanced NSCLC revealed that patients receiving chemotherapy at the EOL were less likely to be referred to hospice. This study also showed a trend toward longer survival with hospice, particularly for patients whose hospice stay was ≥4 days. ⁵¹

Hospice could extend survival by avoiding the risk of overtreatment with chemotherapy with potentially high rates of toxicity. Additionally, hospice may provide improved monitoring and treatment of symptoms and increased social support, factors that have also been associated with improved survival. ⁵⁰

Several RCTs of palliative care have found longer hospice stays in the intervention group, although these findings have not been consistent across RCTs. A trial of referrals to an interdisciplinary inpatient palliative care team showed significantly longer median hospice stays (24 days vs. 12 days), p = 0.04 in the intervention group. ⁸ Temel and colleagues also found that patients with metastatic NSCLC who received EPC were referred to hospice earlier in the course of their illness. ¹¹

Increased social support

Palliative care focuses on addressing the needs of both patients and caregivers, which may also have the potential to increase social support. In addition to psychological support, social support encompasses practical domains such as financial assistance and accompaniment to appointments. Higher levels of support may be related to psychological concerns such as the perception of being a burden to others. One of the benefits of earlier referral to hospice includes greater opportunity for supporting caregivers, and researchers have postulated that this may be one contributing factor to the finding that hospice may extend the length of survival. ⁵⁰ Before hospice, EPC may also increase the support for patients and families.

Researchers have attempted to explore the relationship between social support and survival by examining the role of marital status. An examination of the SEER Medicare database of 25,000 lung cancer patients has shown an association between increased survival in married patients compared with single, separated, or widowed patients, a finding which may be particularly strong in men. ⁵² A significant survival advantage for married lung cancer patients, regardless of gender, was also identified in examination of a large Norweigian cohort. ⁵³ In the head and neck cancer population, investigators have found that partnered males have a survival advantage compared to nonpartnered males. ⁵⁴

A more recent examination of the Mayo clinic lung cancer database did not demonstrate a clear relationship between marital status, survival, and QOL. When considering only patients with locally advanced NSCLC, widowed and single patients had slightly decreased survival compared to married and divorced patients. The researchers found that patients who are not partnered may receive less aggressive cancer-directed therapy, but these data were not analyzed by stage of disease. ⁵⁵ Patients with earlier stage NSCLC might have improved survival with more aggressive treatment, whereas chemotherapy at the EOL in metastatic disease could possibly hasten death. More recent research suggests that the relationship between marital status and survival might also be accounted for by other variables such as gender in the lung cancer population. ⁵⁶

In addition to decreasing burden on caregivers, palliative care clinicians are also providing direct social support to patients adjunctively to standard oncology care. Counseling and helping patients to cope with their illness throughout its course are integral parts of EPC.

In one RCT of EPC, caregivers reported improvement in their QOL, which could translate into greater social support for patients. This intervention was designed for dyads of patients and their caregivers and focused on improving problem-solving abilities. Patients in this trial had advanced cancer and were simultaneously enrolled in cancer clinical trials. The intervention did not influence QOL of patients but did promote higher QOL for caregivers compared to usual care. The impact appeared to build over time, and the authors concluded that it might have more utility if provided earlier in the course of illness. ⁵⁷

Finally, social support, mood, and QOL likely interact and overlap with one another. QOL scales incorporate the domains of psychological and social functioning. Interestingly, one potential mediator of the persistence of depression in patients with advanced cancer is “attachment security.” In other words, how much patients can trust in their connections to others may impact their experience of depression. ³⁶ Promoting social support and decreasing caregiver burden have the potential to impact patient well-being, address practical barriers to receiving care, contribute to shared understanding of illness, and affect patterns of care at the EOL. These complex relationships have yet to be fully investigated.

Summary and implications

Providing palliative care concurrent with cancer-directed care appears to be feasible and acceptable to patients with advanced cancer close to the time of diagnosis. ^9,11,58 Early, integrated palliative care has been associated with significant improvements in mood, QOL, and health care utilization. Preliminary data from one RCT also suggest a clinically meaningful survival benefit of nearly 3 months compared to standard oncology care in patients with NSCLC. ¹¹ However, it is not yet clear how EPC may promote survival in this population. This review explored potential hypotheses that are consistent with the central focus of palliative care on prevention and relief of suffering, better decision making, and coordination of care across the continuum of illness (see Figure 1). We hypothesize that EPC has the potential to impact overall survival by directly affecting the patient’s well-being and experience of suffering, increasing social support, improving understanding of illness and informing decision making, which subsequently contribute to less aggressive care at the EOL and earlier referral to hospice.

Interestingly, although many trials of palliative care interventions have shown improvement in some of the hypothesized mediators of survival, not all of these trials had an observed survival benefit. Affecting only one component, such as decreasing depression, may not be sufficient to explain the benefit linked to EPC. Palliative care is designed to impact multiple aspects of the patient and family’s experience, and we hypothesize that the potential survival benefit is multifactorial. Some of these factors could be biological with decreasing levels of chronic stress leading to changes in adaptive immunity; social with increased social support leading to better personal care; or treatment related with more accurate understanding of illness leading to less aggressive care at the EOL. Additionally, many of these factors are associated with each other, further complicating the identification of which factor or combination of factors might lead to improved survival. To date, there have been no mediation analyses of EPC trials that have supported any of the hypotheses outlined in this review. More research is clearly needed.

Conclusion and directions for future research

Although improved survival could be a result of EPC in patients with metastatic NSCLC, this finding should not overshadow all of the other benefits such as improved QOL, better understanding of prognosis, and less aggressive care at the EOL. Improved survival has not been the primary goal of palliative care delivered early in the course of disease or otherwise.

Future trials of EPC in patients with metastatic NSCLC would ideally have large enough samples to include survival as an end point. Any improvement in survival from EPC is likely multifactorial. Investigating these complex mediating relationships require very large samples. In addition to mediation, investigators should also analyze any differences in the effect of palliative care on patient subgroups.

To replicate findings, earlier palliative care interventions must also be clearly described and standardized. Once standardized, individual components of the interventions can be examined to assess their impact on outcomes. Focusing on smaller intervention components could help to identify underlying mechanisms.

Research on EPC in metastatic NSCLC is still in its nascent stage. However, this research has the potential to change cancer care. Similar to the trials of anticancer therapies, a sufficient evidence base is required for modifying standard practice. Although it is not the primary goal of EPC, including survival as an outcome and identifying the underlying mechanisms of a potential survival benefit might further facilitate its acceptance and dissemination.