Palliative Care Service Utilization and Advance Care Planning for Adult Glioblastoma Patients: A Systematic Review

Simple Summary Glioblastoma is the most common primary brain malignancy diagnosed in adults, and, despite standard of care treatment, it carries a devastating prognosis with median overall survival of 16–21 months. Advance care planning and palliative care services utilization are important for this patient population due to their cancer and neurocognitive symptoms. We present a systematic review on prevalence of advance care planning, end-of-life services utilization, and experiences among adults with glioblastoma. The findings from our review serve as a foundation for future additional works, particularly prospective studies, that may address gaps in palliative care resource utilization and disparities in advance care planning for adult glioblastoma patients. Abstract Glioblastoma (GBM) has a median overall survival of 16–21 months. As patients with GBM suffer concurrently from terminal cancer and a disease with progressive neurocognitive decline, advance care planning (ACP) and palliative care (PC) are critical. We conducted a systematic review exploring published literature on the prevalence of ACP, end-of-life (EOL) services utilization (including PC services), and experiences among adults with GBM. We searched from database inception until 20 December 2020. Preferred reporting items for systematic reviews guidelines were followed. Included studies were assessed for quality using the Newcastle-Ottawa Scale. The 16 articles were all nonrandomized studies conducted in six countries with all but two published in 2014 or later. ACP documentation varied from 4–55%, PC referral was pursued in 39–40% of cases, and hospice referrals were made for 66–76% of patients. Hospitalizations frequently occurred at the EOL with 20–56% of patients spending over 25% of their overall survival time hospitalized. Many GBM patients do not pursue ACP or have access to PC. There is a dearth of focused and high-quality studies on ACP, PC, and hospice use among adults with GBM. Prospective studies that address these and additional aspects related to EOL care, such as healthcare costs and inpatient supportive care needs, are needed.


Data Extraction and Analysis
We extracted key features from the eligible studies, including aims, design, inclusion and exclusion criteria, patient population, and outcomes. Outcome variables included ASCO Quality Oncology Practice Initiative (QOPI) adherence measures, hospice utilization rates, PC utilization rates, percentages of advance directive documentation, and location of death [19].

Quality Assessment
Three authors (A.W., G.R.C., and C.B.P.) independently used the Newcastle-Ottawa Scale to assess the quality of the studies [20]. Studies were graded as very low, low, medium, or high quality based on selection criteria, risk of bias, and overall study design [21]. No studies were excluded from analysis due to quality.

ACP Conversations and Documentation
The prevalence of any ACP documentation ranged from 4% to 55% (Table 1) [12,29,30,33,34]. The study from our institution by Pollom et al. [12] assessed ACP at different time points in the patients' treatment course and found that 11% of patients had ACP documented before their diagnosis, 29% within six months after diagnosis, 54% at their last follow-up, and among decedents, 55% prior to death. This prevalence was similar to the rates reported by Hemminger et al. [29], where ACP documentation was available in 52% of cases by the third oncology visit, which is the standard recommended by ASCO QOPI's EOL performance measures. Three studies also specified the prevalence of code status documentation, ranging from 36% (via a medical orders for life-sustaining treatment {MOLST} form) to 46% [12,29,30]. No study reported the prevalence of 'do not hospitalize' directives.
Two studies found the prevalence of an ACP or goals of care conversations to be 83% [27,31]. Miranda et al. [31] found that these conversations most frequently discussed EOL planning (including hospice and PC), prognosis, and prognostic understanding. None of the analyzed conversations reported healthcare proxies, family involvement, or information preferences [31].

Healthcare Services Utilization at the EOL
Ten studies reported healthcare services utilization at the EOL (Table 1) [12,[23][24][25][26]29,30,[34][35][36]. Four studies addressed PC services referral and utilization [12,25,26,29], six studies addressed hospice referral and utilization [12,23,24,29,30,34] and four studies addressed other healthcare services utilization, including inpatient hospitalizations [30,[34][35][36]. Among the four studies reporting PC utilization, PC referrals ranged between 39% and 40% [12,26], and PC consult utilization was reported to be between 34% and 36% [25,29]. Median time from diagnosis to PC consult was only measured in one study and found to be 111 days [26]. Notably, the study by Pollom at al. [12] found that among patients with a PC referral, all had this referral at least three days prior to death, and another study by Lin et al. [26] found that the median time from PC referral to death was 33 days.
Hospice referral rates are one of ASCO's QOPI metrics and, in these studies, ranged between 66% and 76% while utilization rates ranged from 38% to 86% [12,30]. The timing of hospice referral and enrollment also varied between studies. Oberndorfer et al. [24] found that while 0% of patients were enrolled in hospice care between six and ten weeks prior to death, hospice enrollment increased to 24% between weeks two and six prior to death and reached 38% in the last two weeks before death. One study by Diamond et al. [23] found that 23% of home hospice GBM patients were enrolled within seven days of death. Kuchinad et al. [30] found that patients were referred to hospice a median of 22 days before death and had a median hospice LOS of 21 days, while Hemminger et al. [29] reported a median time of 18.5 days from hospice enrollment to death and found that nearly 60% of patients had enrolled in hospice within seven days of death.
Lastly, four studies reported hospitalizations at or near the EOL [30,[34][35][36]. The percent of patients with a hospitalization in the last month of life ranged from 37% to 42%, with an average LOS of 8.75 days [30,36]. One study reported that among those hospitalized in the last month of life, 34% had an ICU admission [36]. Two studies discussed hospitalizations and found that 20% to 56% of patients spent over 25% of their overall survival time hospitalized, and notably, up to 20% of patients spent 100% of their overall survival time hospitalized [34,35]. Paszat et al. [34] reported that more than half (56%) of GBM patients diagnosed between 1982 and 1994 in Ontario spent over 25% of their overall survival time hospitalized. Among patients spending over 25% of their overall survival time hospitalized, the median cumulative LOS was 31 days, which was almost double compared to 17 days in those spending less than 25% of survival time hospitalized [35].

Location of Death
Six studies included data on location of death (Table 1) [22,[28][29][30]32,36]. In three of the studies, death was most common in a home setting, ranging from 39% to 64% of deaths [29,30,32]. In the three other studies, however, death in other healthcare facilities (including hospitals and skilled nursing facilities) was more common, with up to 78% patients dying in such facilities [22,28,36]. Flechl et al. [28] found that of the 45% of patients who desired to die at home, 68% of them died there, while 27% of them died in a hospital, and 5% in hospice. Lastly, across studies, hospice, whether inpatient or at home, was listed as the location of death for 12% to 64% patients [22,[28][29][30]32].

Patient-Reported Experiences
Among all studies describing patients' experiences at the EOL, two described patientreported outcomes [25,31]. In the study by Seibl-Leven et al. [25], patients completed a self-assessment using the Palliative Outcome Scale to report their PC concerns [37,38]. While "other symptoms" (defined as symptoms other than pain) were of minor importance to this cohort of GBM patients, familial anxiety and patients' own illness-related anxieties were of high importance. Moreover, when asked to report what problems they have experienced in the recent past, GBM patients most commonly reported experiencing healthrelated impairments (33%), worries about the future (11%), and loss of everyday skills (10%) [25]. On the other hand, Miranda et al. [31] used the Life Priorities Survey [39] to describe patients' goals and priorities during EOL. The survey allows patients to designate which goals and priorities are most important to them using Likert scaling and ranking their top five goals and priorities [39]. While "live as long as possible, no matter what" was the priority most frequently ranked in patient's top three goals (N = 9 of 22 patients), three of the 22 said it was "not important at all" [31]. Lastly, all 22 patients reported that "(being) mentally aware" and "(being) independent" were "somewhat to extremely important" [31].

Caregivers' Experiences
Two studies focused on GBM caregivers' quality of life, challenges faced, and emotions at the end of the patients' life [25,28]. Flechl et al. [28] found that 50% of caregivers reported job restrictions that impaired their caregiving abilities, 29% reported that they felt incompletely prepared for their tasks, and 29% reported financial difficulties as a significant barrier. Financial difficulties were associated with caregiver burnout (60%) and reduced quality of life, with impaired quality of life among caregivers being comparable to that reported in GBM patients themselves [28]. In a study by Seibl-Leven et al. [25], financial difficulties (mean Zarit Burden Interview {ZBI} score 0.76 ± standard deviation 1.08) were of lower concern among caregivers, whereas both insufficient time for self (1.36 ± 1.16) and stress between caring and other responsibilities (1.56 ± 1.30) emerged as common challenges. However, the caregivers indicated that they did not feel overly burdened by the patient's illness-related symptoms (e.g., embarrassment over patient behavior), and they did not endorse a feeling of being "unable to care for relatives any longer" or wishing to leave the "care of relative to someone else" [25].

Study Quality
Using the Newcastle-Ottawa Scale [20,21], five studies were found to be of moderate quality [22,26,[34][35][36] and ten were of low quality [12,[23][24][25]27,[29][30][31][32][33] (Table 1). None were found to be of high quality, because they were not randomized controlled trials. The most common study quality detriment was the "comparability" domain, driven by lack of control groups or matched subjects in study designs. The most common quality strength was in the "outcomes" domain, driven by the use of independent blind assessments or record linking to assess outcomes.

Discussion
This systematic review suggests low rates of ACP documentation ranging from 4% to 55% for adult GBM patient cohorts in different countries in six separate studies, and that the existing literature on ACP, patient-reported experiences beyond symptomatology, and PC services utilization is sparse, varied, and of low quality. Most studies addressed only disparate aspects regarding EOL needs and experiences among adult GBM patients. Although our literature review did not yield any randomized controlled trials that met the pre-specified eligibility criteria, we did find a published protocol for a forthcoming randomized phase III trial evaluating the effect of early PC for GBM patients [40]. Future additional studies with rigorous design, particularly prospective investigations, are warranted.
Our findings continue to suggest that the adult GBM patient population would benefit from more and better ACP. There are already existing groups pursuing novel programs to better address this deficit, and quality measures used in other cancer populations that could be translated into GBM management. For example, Fritz et al. designed and developed an ACP program for physicians from multiple specialties involved in GBM patient care, as well as for a few GBM patients and their caregiver proxies, to discuss topics including financial concerns and proxy needs [41]. Providers caring for GBM patients may potentially also participate in the ASCO QOPI, which was introduced in 2006 and consists of over 150 quality metrics for oncology practices [42]. While QOPI focuses on clinical practices for melanoma, breast, colorectal, gynecological, lymphoma, lung, and prostate cancers, some of the PC and EOL measures are also applicable to neuro-oncology patients. QOPI EOL metrics include documentation of pain and dyspnea assessments, rates and timing of hospice enrollment and PC referrals as well as the proportion of patients who present to emergency departments or are admitted to intensive care units within the last 30 days of life [19]. According to one study of oncology practices that report and submit data on QOPI EOL measures, each subsequent year of participation led to improved rates of symptom documentation and PC referrals as well as increases in favorable performance related to care of pain compared to peer organizations [43]. Some existing studies of GBM patients already report data related to QOPI EOL measures [12,29,30]. Additional future studies with data on QOPI adherence may be beneficial in characterizing the landscape of ACP and EOL care for adult GBM patients. In the future, QOPI measures may also be established specifically for neuro-oncology practices (e.g., neurological deficits) and lead to improved care for patients with brain cancer at the EOL.
Many studies included in this systematic review focused on healthcare resource utilization, such as length of stay and referrals to PC or hospice. Yet, only a minority of these studies explored the proportion of patients referred and the timing of the referral. In these studies, patients were generally referred to hospice near the EOL, within a range of 3-22 days prior to death. Little data were available from these studies about the decisionmaking underlying the timing of these referrals and whether these referrals are based on qualitative scales of functional status or disease and treatment course. This systematic review includes studies from multiple countries and healthcare systems that define in different ways the scope of palliative care and hospice services. For example, hospice is regarded as a Medicare benefit in the United States, while other countries include hospice as an extension of available palliative care resources for patients to enroll in. Overall, there is much room for improvement in palliative and hospice services utilization by patients with aggressive and incurable diseases, such as GBM.
Patient and caregiver-reported experiences and concerns are also important considerations of EOL care. Out of all included studies within this systematic review, only a few refer to these outcomes, each utilizing separate methodologies, such as self-assessment surveys like the Palliative Outcome Scale or interview questions. GBM patients most commonly expressed concern with their health, while caregivers had many worries regarding finances according to Flechl et al. [28]. In other studies involving patients with advanced cancer and even serious cases of bacteremia, common patient-reported concerns also centered on psychological distress and limitations on function [44][45][46]. Future studies on GBM patients may further delineate these emotional and physical concerns through use of rigorous qualitative methods to enhance understanding of EOL experiences and needs.

Limitations
Study limitations include heterogeneity of cohort size and study design, which make reported outcomes difficult to generalize. Selection biases and confounding factors are also inherent to retrospective studies, such as many summarized in this systematic review. In addition, this review is limited by the paucity of studies dedicated solely to adult GBM patients, particularly as many studies do not report outcomes solely for GBM and instead summarize them for patients with all high-grade gliomas as a single cohort.
Furthermore, this systematic review highlights multiple studies that rely on metrics as indicators for quality palliative and goal-concordant care (e.g., rates of ACP documentation and hospice enrollment), which is itself difficult to define and measure [11]. Some have cautioned against focusing on such metrics as the solution to achieving goal-concordant care [47,48], because the metrics may not reflect the quality of palliative care or achieve goalconcordant care. Because measuring outcomes and quality in palliative care still requires standardization via such metrics, this is a limitation inherent to this systematic review.

Conclusions
Adult GBM patients have a poor prognosis and experience an array of debilitating symptoms due to the incurable and concurrent neuro-degenerative nature of their disease. Proactive advance care planning and appropriate use of palliative care resources are critical aspects of high-quality care for these patients and their caregivers, yet our findings suggest relatively low prevalence of both of these components among GBM patients. The field would benefit from rigorous studies, particularly involving prospective cohorts, to inform future improvements in ACP and EOL care for adult GBM patients as well as to explore other pertinent topics (e.g., whether and which social determinants are associated with utilization of PC, high-quality EOL quality metrics).
Funding: This research received no external funding. The APC was funded by the Stanford University School of Medicine Departments of Neurology and Radiology.

Data Availability Statement:
No new data were created or analyzed in this study. Data sharing is not applicable to this article.
OR "advance care" OR "advance health care planning" OR "advance medical planning" OR "advanced care planning" OR "attitude to death" OR "withholding treatment").
Cochrane Library (glioblastoma OR glioma OR astrocyte* OR spongioblastoma OR "GBM") AND ("terminal care" OR "end of life" OR "end of life care" OR "EOL" OR "End-of-life" OR "advance care planning" OR "advance directive*" OR "advanced care" OR "advance care" OR "advance health care planning" OR "advance medical planning" OR "advanced care planning" OR "attitude to death" OR "withholding treatment").