Triggers for Referral to Specialized Palliative Care in Advanced Neurologic and Neurosurgical Conditions

Abstract

Background and Objectives

To systematically review the literature for the most suitable trigger criteria for referral to specialist palliative care services in life-limiting and life-threatening neurologic and neurosurgical conditions.

Methods

Literature searches were conducted in Ovid MEDLINE and EMBASE (1990–December 2020). To be included, studies must have trigger/referral criteria clearly outlined, a ≥75% nononcology neurosciences population, and consensus or guidelines documents regarding palliative neurosciences or trigger/referral criteria. We excluded studies that had an oncologic or non-neurosciences population as the main focus of study, trigger and referral criteria not clearly outlined, and no primary or duplicative data. The protocol was registered with PROSPERO (CRD4202013579), and Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines were followed. The American Academy of Neurology Clinical Practice Guidelines Process Manual was used to assess for risk of bias.

Results

Our search identified 1,748 publications, of which 22 articles met the eligibility criteria. Studies were considered in 2 main groups: (A) studies designed specifically to identify trigger criteria for referral to specialized neuropalliative care services (n = 9) and (B) studies that retrospectively reported the reason for referral to specialized palliative care or reflected a consensus statement among people with advanced neurologic illness (n = 13). Overall, the results suggest that several published referral triggers for specialized neuropalliative care are based on expert consensus. However, there is a growing body of literature providing evidence-based condition-specific triggers for multiple sclerosis, parkinsonism, amyotrophic lateral sclerosis, and dementia.

Discussion

There is a growing body of research that outlines evidence-based referral triggers for neuropalliative care. The ambiguity of nomenclature surrounding referral triggers in the current literature and field of neuropalliative care was a limitation to this study. We suggest that condition-specific triggers are likely to be the most effective for identifying the appropriate patients and timing for referral to specialist palliative care. (PROSPERO registration number: CRD42020135791, crd.york.ac.uk/prospero)

The field of neuropalliative care is rapidly maturing with emerging high-quality evidence for both intervention and benefit.¹ Studies clearly demonstrate what has always been known by those in clinical care; the care needs of individuals with life-limiting and life-threatening nononcologic neurologic and neurosurgical conditions, particularly at advanced stages of illness, are substantial and are not being met by current standard care practices.^1-4

Palliative care is defined as “the active holistic care of individuals across all ages with serious health-related suffering because of severe illness and especially of those near the end of life. It aims to improve the quality of life of patients, their families, and their caregivers.”⁵ Although comprehensive, this definition does not provide clarity regarding which neurologic conditions will benefit from the initiation of palliative care or at what time in the illness trajectory it should be initiated. Common nononcologic neurologic and neurosurgical conditions that would benefit from specialist palliative care include Parkinson disease (PD) and related disorders, dementia, many neuromuscular conditions, stroke, and neurocritical care.¹ The care needs in neurologic conditions are high but variable between conditions, and the most appropriate timing for initiation of neuropalliative care is not clear due to the variability both within and between conditions. In addition, palliative care needs of individuals with neurologic conditions differ from those of individuals with other advanced illnesses, in both symptom burden and type and the general trajectory of their illness.³

In addition to the challenge of defining which nononcologic neurologic and neurosurgical conditions to refer to palliative care, clinicians must ascertain the time to initiate neuropalliative care. There is a spectrum of symptomatology and rates of illness progression both between and within the different neurologic conditions, and both sources of variability contribute to prognostic uncertainty. Evidence also indicates that neurologists at all stages of training and practice are not comfortable with their knowledge base in palliative care, suggesting further guidance is needed in identifying the triggers for consultation.^6,7 This discomfort and knowledge variability increases the likelihood that neurologists will initiate neuropalliative care at different times. Clearly identified trigger criteria may help neurologists to develop a deeper understanding of when to refer and what palliative care can offer. Clearly, we need to define the timing for initiation of neuropalliative care and the role of palliative care in the management of nononcologic neurologic and neurosurgical conditions.

Using a systematic review, we aimed to investigate the most appropriate trigger criteria for referral to specialized palliative care services in the neurosciences. We will outline the best evidence supporting the reason for and timing of referral to specialist palliative care in advanced nononcologic neurologic and neurosurgical illnesses.

Methods

Search Strategy and Selection Criteria

The systematic literature searches were conducted in Ovid MEDLINE (1990 to March, 2019) and EMBASE (1990 to December, 2019) electronic bibliographic databases by a clinical librarian with experience in conducting electronic literature searches in consultation with the review authors. No language restrictions were applied. An updated search was performed in July 2019 and December 2020. A combination of search terms (keywords and subject headings: Medical Subject Headings [MeSH] in MEDLINE and Emtree terms in Embase) including, but not limited to, neurology, neurosurgery, neurosurgical procedures, multiple sclerosis (MS), amyotrophic lateral sclerosis (ALS), PD, subarachnoid hemorrhage, traumatic brain injury, terminal care, palliative care, hospice care, and patient referral were used to identify eligible reports. The comprehensive search strategy is provided in eAppendix 1 (links.lww.com/CPJ/A331), and the search strategy was designed using a broader set of search terms related to neurology (i.e., neuro$.af.), which is an alias for all the fields that occur in the source documents and should broadly include all neurology and neurosurgery-related citations. Modifications to the search strategy were made for each database using appropriate for a particular database thesaurus terms, such as MeSH in MEDLINE and Excerpta Medica Thesaurus (Emtree terms) in Embase, and fields. Online issues of the journals and journals' supplements including, but not limited to, BMJ Neurology, Continuum, BMJ Supportive & Palliative Care (Supplement) were manually searched to identify additional relevant studies. The bibliographies of relevant articles were examined for additional published data that met our inclusion criteria. A detailed description of our search strategy can be found in eAppendix 1.

Inclusion criteria were as follows: (1) trigger/referral criteria clearly outlined; (2) data provided in a format that allows for analysis; (3) ≥75% nononcology neurosciences population; and (4) consensus or guidelines document regarding palliative neurosciences or trigger/referral criteria. Studies were excluded that had the following: (1) an oncologic population as main focus of study; (2) a non-neurosciences population as main focus of study (>25% included patients); (3) trigger/referral criteria not clearly outlined, and (4) no primary or duplicative data. Pediatric-focused articles were also considered.

Peer-reviewed original research publications without secondary reporting were considered including randomized controlled trials, observational studies, and case-control, cohort, case series, and qualitative studies. There were no restrictions on socioeconomic, geographic, or ethnic backgrounds of participants. We excluded non–peer-reviewed articles, conference abstracts, study protocols, reviews and books or publications in a language other than English for which we could not obtain a translation.

Standard Protocol Approvals, Registrations, and Patient Consents

This systematic review was performed in accordance with Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) and synthesis without meta-analysis (SWiM) guidelines. The review protocol was registered with PROSPERO (registration number: CRD42020135791, crd.york.ac.uk/prospero). See eTable 1 (links.lww.com/ CPJ/A331) for the PRISMA. Research ethics approval was not required for this systematic review.

Data Extraction and Analysis

Two reviewers (K.M. and T.E.G.) independently screened and selected studies for inclusion (Figure). Articles were screened by title and abstract and then by full text. Disagreements were resolved through discussion between reviewers. There was 1 disagreement between reviewers, which resulted from a difference of opinion regarding the primary reason for excluding a title/abstract. A third reviewer (M.E.J.) focused on resolving disagreement between the first 2 reviewers, contributing to full-text review, classification of level of evidence, and on the approach for grouping studies as group A or group B. Group A studies met protocol criteria. Group B studies did not meet protocol criteria but were included because they contributed valuable information to the research question. Covidence was used to import references, title and abstract screen, and full-text review. Variables for data extraction were predetermined and based on the research objectives. Extracted data included the following: number of participants, study design, proportion of patients with neurologic diagnosis, neurologic/neurosurgical diagnosis, population of interest (i.e., inpatient, outpatient, and hospice), and referral triggers.

Figure. Preferred Reporting Items for Systematic Reviews and Meta-analyses Flowchart Detailing the Screening and Selection of Studies for Inclusion.

A qualitative review of the search results was completed (as per SWiM⁸ reporting guideline). The outcome metric of interest was the existence of evidence-based trigger criteria for referral to neuropalliative care in addition to the trigger criteria themselves. Data synthesis was accomplished through examination of the methods for developing the evidence-based trigger criteria, compilation of the trigger criteria, and classification of the level of evidence for group A studies. We did not examine heterogeneity statistically because the included studies were not designed to measure an intervention or an effect. Heterogeneity was addressed by ordering tables according to subpopulations (e.g., neurologic condition) and direct vs indirect sources of evidence (e.g., groups A and B). The subgrouping of studies was not prespecified in the systematic review protocol and was determined after completing the full-text review. Data are presented using narrative synthesis in the Results section and in tabular format with additional details.

Risk of Bias and Quality Appraisal

Both authors examined each study independently for risk of bias and quality. Specifically, the risk of bias and study quality were graded using the American Academy of Neurology Clinical Practice Guideline Process Manual⁹ using a 4-tiered classification system from Class I to IV (Tables 1 and 2).

Table 1.

Group A Characteristics for Pediatric and Adult Parkinsonism Studies and the American Academy of Neurology Level of Evidence Classifications

Table 2.

Group A Characteristics for Adult Multiple Sclerosis, Neuromuscular Disorders, Dementia, and Neurodegenerative Conditions Studies, and the American Academy of Neurology Level of Evidence Classifications

Data Availability

Data are available on request from qualified investigators (contact corresponding author).

A Note on Terminology

The terms palliative care and hospice are considered together for the purposes of this systematic review and selection of studies. This is because the definitions for these terms differ between and even within nations. In the United States, hospice care refers to palliative care at the end of life, specifically with a 6-month life expectancy.¹⁰ Studies assessing trigger criteria for the last 6 months of life will be separated out from other studies that may investigate trigger criteria earlier in the illness trajectory.

Results

The literature search identified 1,748 publications (Figure—PRISMA flow diagram of study selection). Based on defined inclusion/exclusion criteria, 1,505 studies were excluded from the review. All studies meeting the inclusion criteria were published in English. The most common reason for exclusion was because the referral or trigger criteria were not clearly outlined (62), no original research data were provided in the publication (23), or the study focus was not neurologic (5). Four studies retrospectively reported the most common symptoms at the time of referral for specialized palliative care services but did not report the reason for consultation, so were also excluded. The remaining 241 studies were selected for full-text review.

After full-text review, 22 studies were included in the systematic review and were considered in 2 main groups: (A) studies designed specifically to identify trigger criteria for referral to specialized palliative care services among people with advanced neurologic illness (group A, 9 studies) and (B) studies in which referral triggers were not the main focus but that retrospectively reported the reason for referral to specialized palliative care services or reflected a consensus statement (group B, 13 studies). These groups were chosen because the group A studies provide direct evidence addressing the aim of this systematic review, and the group B studies provided indirect evidence (actual reasons that people were reviewed although identification of these reasons was not the primary goal of the research) pertaining to the aim of this systematic review. The 22 studies contained data for a total of 2,907 patients (group A = 1,318 patients, group B = 1,589 patients) from a mixture of outpatient, inpatient, and hospice settings. More general trigger criteria were identified for neurodegenerative conditions as a whole, and condition-specific triggers were identified for parkinsonism, MS, ALS, and dementia. These triggers will be outlined in more detail further. There were no evidence-based referral triggers identified for neurosurgical conditions.

Group A: Triggers for Palliative Care Referral

See Tables 1 and 2 for group A study characteristics and American Academy of Neurology level of evidence classifications. Studies included in group A investigated trigger criteria for the following conditions only: pediatric neurology, parkinsonism, MS, motor neuron disease (MND), dementia, and neurodegenerative conditions in general.

One Class I study investigated triggers for palliative care referral in pediatric neurology inpatients¹¹ (305 general neurology pediatric inpatients, Class I). This study was a 3-phased case-controlled pilot study aimed at identifying unmet palliative care needs using a quantifiable and objective method based on review of the electronic medical record. In phase 1, the authors retrospectively screened 164 pediatric neurology inpatient charts for the number of emergency room visits, number of hospital admissions, and number of problems on the electronic medical record problem list. In phase 2, all patients admitted to pediatric inpatient neurology over a 6-month period were prospectively screened for patients with ≥6 admissions in the previous 365 days. Seven children were identified, and structured interviews were used to determine their palliative care needs. In phase 3, age-matched and sex-matched controls with <6 admissions in the previous 365 days were also interviewed. Using these criteria, they were unable to distinguish between pediatric neurology patients with and without unmet palliative care needs.

One adult study considered referral triggers for neurodegenerative conditions in general¹² (300 inpatients, outpatients, and hospice patients, Class II). This study focused on factors signaling the last 6 months of life, which would be pertinent to hospice referral in the United States. This retrospective study used a principal components analysis to determine the timing and presence of triggers in the last 2 years of life. The triggers examined were based on previous consensus-based triggers suggested in the literature. They showed that overall there is a high burden of triggers in the last 2 years of life and peaking in the last 6 months of life. Some triggers could be considered together, and 4 factors explained 64% of the variance. These factors were as follows: factor 1: deterioration in physical function, dysphagia, significant complex symptoms, and pain; factor 2: weight loss and respiratory symptoms; factor 3: recurrent infections and cognitive decline; and factor 4: aspiration pneumonia. The most important triggers seemed to vary based on condition with different factors being associated with survival in different conditions.

Parkinsonism

Four studies investigated referral triggers for people living with parkinsonism (PD = 3,^13-15 atypical parkinsonism = 1¹⁶). One retrospective study assessed the factors most frequently triggering admission to hospital among people with atypical parkinsonism admitted to an inpatient palliative care unit. These factors were swallowing difficulties, gait instability/falls, spasticity/pain, and impaired communication¹⁶ (38 palliative care unit inpatients, Class II).

Two studies prospectively developed referral tools aimed at helping clinicians to appropriately identify and trigger referrals to palliative care. These 2 tools include the RADboud indicators for PAlliative Care Needs in Parkinson's Disease¹³ (RADPAC-PD; 51 health care providers, no patients, Class I) and the Needs Assessment Tool: Parkinson's Disease¹⁵ (NAT:PD; Richfield, personal communication, 50 outpatients and 28 caregivers, Class I—unpublished data). For the RADPAC-PD tool, a mixed-methods approach was used to develop trigger criteria. The semistructured interviews and focus groups suggested 2 important time points in a patient's trajectory of illness for intervention: (1) moment to initiate advance care planning; and (2) actual start of the palliative phase.¹³ The Delphi process was used to determine the most important triggers to suggest the appropriate time for advance care planning and for the start of the palliative phase. The main risk of bias in this study is that the triggers used were initially developed based on expert opinion. For the NAT:PD tool, the Needs Assessment Tool: Progressive Disease in Cancer was adapted for use in PD and then evaluated its validity and reliability in a broad population of people living with PD.¹⁵ Adaptations were based on a systematic review of the literature and an expert panel. Both these tools are completed by clinicians thereby reducing the burden on patients and caregivers. The RADPAC-PD applies across patient-care settings,¹³ whereas the NAT:PD was validated in an outpatient population at all stages of disease.¹⁵ Finally, the trigger criteria in neurodegenerative conditions were assessed using a retrospective chart review and principal component analysis.¹² They also looked at specific diagnoses and suggested that aspiration pneumonia may be a useful trigger for referral in PD and that aspiration pneumonia and recurrent infections and cognitive decline may be useful triggers for referral in atypical parkinsonism.¹²

No studies assessed trigger criteria specific for hospice eligibility, but a 2015 study investigated symptoms changes signaling the last 6 months of life in PD.¹⁴ This study was a retrospective chart review of Veterans Affairs outpatients and inpatients with PD and used a within-subjects repeated-measures design to show that patterns of symptom change near the end of life may identify the last 6 months of life and thereby serve as triggers for referral to hospice¹⁴ (339 patients, outpatients and inpatients, Class II). They showed that below normal weight and a decrease to ≤2 or fewer dopaminergic medication prescriptions heralded the last 6 months of life.

Multiple Sclerosis

A 2-round online Delphi, followed by an expert workshop, was used to identify the most appropriate timing for when and why to refer to specialist palliative care for people with MS¹⁷ (58 health care providers, no patients, Class II). The Delphi was used to identify triggers for referral, which were supplemented by the expert workshop. The workshop participants cared for more severely affected people with MS than did the Delphi participants, which is the likely explanation for discrepancies between Delphi consensus and workshop consensus results. Overall, an expanded disability status scale (EDSS) ≥6–8, need for nursing home care or hospital admission, need for psychosocial support, need for support for relatives, communication on disease progression, uncontrolled pain, complex symptoms, 2 symptoms on HOPE scale scoring >3, and decreasing cruising radius were identified as triggers for referral.¹⁷

Neuromuscular Conditions

One single-centered retrospective chart review study assessed for predictors of early death after diagnosis with a neuromuscular condition to identify potential triggers for earlier referral to palliative care¹⁸ (86 outpatients, Class III). The most frequently occurring symptoms associated with an early death were early-onset dyspnea, early-onset dysphagia, and early-onset pain.¹⁸ Statistical modeling showed that older age at onset of neuromuscular disorder, early-onset of dyspnea, dysphagia constipation, and impaired mobility were associated with earlier death and that the combination of these factors predicts poor survival.¹⁸ This study used objective criteria and outcomes to reduce the chance of bias but was completed at a single tertiary care referral center, and data analysis was retrospective, which increases the chance of bias.

Dementia

A novel screening tool was developed for identifying people living with late-stage dementia who would benefit from a palliative care referral¹⁹ (91 general medical inpatients, Class I). This tool was established in the context of a broader clinical trial, and it relied on a computer algorithm search of the electronic health record to enable efficient identification of potential study participants. The algorithm was based on International Statistical Classification of Diseases codes. After the identification of potential study participants, a manual chart review confirmed eligibility.¹⁹ Although this approach was effective and successfully identified patients with late-stage dementia, it is not easily generalizable to other health care settings and requires a data warehouse and research team to enact.

Group B: Retrospective Identification of Reasons for Referral

See Tables 3 and 4 for group B study characteristics. The group B studies were not designed to evaluate trigger criteria for referral to specialized palliative care services, and thus, the level of evidence and risk of bias pertaining to our research question could not be reasonably established. Studies included in group B investigated trigger criteria for the following conditions only: neurology in general, stroke, parkinsonism, MS, MND, neurocritical care, and neurodegenerative conditions in general.

Table 3.

Group B Characteristics for General Adult Studies and Adult MND and MS Studies

Table 4.

Group B Characteristics for Adult Parkinsonism, Neuro-ICU, and Progressive Neurodegenerative Condition Studies

Three adult publications considered referral triggers for neurodegenerative conditions as a group. The first study completed a randomized controlled interventional study to evaluate the effects of neuropalliative care on patients with progressive neurologic disorders and their caregivers²⁰ (quality of life and satisfaction with care provided; 151 outpatients, 140 caregivers). The intervention was an individual consultation with specific care providers for the patient and family caregiver for 45–60 minutes either in the institution or in the patient's home. Triggers for referral to palliative care included the diagnosis of a progressive neurologic disease (MS, PD, and MND), age older than 18 years, a palliative performance scale ≤70 points, and a Mini-Mental State Examination ≥24.²⁰ The second study completed a randomized controlled trial of fast-track vs standard-track palliative care intervention for people and their caregivers affected by progressive neurodegenerative diseases²¹ (50 outpatients and caregivers). They used clearly outlined inclusion criteria that were adapted from the Gold Standards Framework prognostic indicator guidance.²¹ For referral, patients with MND had to have severe disease, patients with MS had to have an EDSS ≥8.5, and patients with PD required a Hoehn and Yahr score ≥4.²¹

The third study describes the development of a neurologic care pathway for meeting the palliative care needs of people living with neurologic conditions including MS, MND, Huntington disease, and PD²² (no patients, care pathway description). The goal of the pathway is to support timely referrals and coordination of care throughout the illness journey. Triggers for referral were taken from the Macmillan Gold Standard Framework (National Health Services, End of Life Care Programme, 2008). These included consideration of specialist palliative care referral when: (1) there are intractable symptoms causing not only suffering, especially pain—but also nausea or breathlessness, (2) there are difficulties with care coordination/management of complex care needs, (3) lifespan is likely to be limited, (4) there is issue of communication and competence, and (5) there is a need for care planning and advance decisions.²²

Parkinsonism

In an outpatient population, a retrospective chart review of all patients attending clinic over a 1-year period was completed and included a description of the referral criteria that were used²³ (96 outpatients). Ninety-six patients were included, and the most common reasons for referral, in order from highest to lowest were as follows: (1) psychosocial support, (2) complex symptom management, (3) advance care planning, (4) caregiver support, (5) goals of care, (6) spiritual support, (7) social/family issues, and (8) care transitions.²³ Sources of bias in this study included that it was single centered and retrospective. The same research group went on to perform a randomized controlled trial assessing integrated outpatient palliative care with standard care for patients with PD and related disorders²⁴ (210 outpatients and 275 caregivers). The study included 210 patients and 175 caregivers and assessed patient quality of life and caregiver burden after 6 months. The inclusion criteria were a diagnosis of PD or related disorders and moderate to high scores on the NAT:PD modified for PD.²⁴

Multiple Sclerosis

A delayed intervention randomized controlled trial was conducted for people with advanced MS²⁵ (52 inpatients and outpatients). Patients deemed to have palliative care needs were randomized to standard best practice vs fast-track palliative care intervention. Although the goal of this study was to assess the palliative care intervention itself rather than the reasons for referral, the reasons for referral were clearly outlined. The triggers were not specifically evidence based but included the following: pain and symptom control, complex psychosocial issues, end-of-life care planning, hydration and nutrition, advance directives and competency, and consent issues.²⁵

Neuromuscular Conditions

Experiences with the integration of a palliative care specialist into the ALS outpatient clinic in an observational cohort study over a 1-year period was prospectively described²⁶ (500 outpatients). Patients were identified for referral to the palliative care specialist using previously published, but not specifically validated, triggers. Advance care planning and transition points in care were the 2 most common reasons for referral,²⁶ and 54% of referred patients also benefitted from symptom management at the time of consultation.²⁶ Another study retrospectively reported reasons for referral to inpatient palliative care among inpatients with ALS at 2 hospitals²⁷ (24 inpatients). Just less than half of patients with an acute nonelective hospitalization also had a palliative care consultation. The main reasons for consultation included the following: goals of care, pain, hospice information/referral, dyspnea, and excessive oral secretions.²⁷

Stroke

An Australian research group investigated factors associated with referral to inpatient palliative care among patients admitted to a stroke care unit²⁸ (57 inpatients). Although they were not testing or validating referral triggers, they reported the characteristics of referrals and factors influencing referral in detail. Of 544 admissions to the stroke unit, 57 patients were included in the analysis. The principal reasons for referral from the stroke care unit to the palliative care unit included the following: end-of-life care (71%), discharge planning (9.7%), request to transfer to the palliative care unit (8.1%), referral to community palliative care services (4.8%), and symptom management (3.2%).²⁸ Older age, sex, and stroke type were associated with referral to palliative care. Men were referred to palliative care at younger ages than women, and primary intracerebral hemorrhage, they made up 35% of referrals,²⁸ but only make up 10%–15% of strokes.

The American Heart Association/American Stroke Association published a Scientific Statement specifically focusing on end-of-life care in stroke²⁹ (no patients, consensus statement). One PICO question within the scientific statement was specific to referral to palliative care and was entitled “Role of palliative care specialists.” They state that, “patients with stroke were more functionally impaired at the time of consultation, more likely to die in the hospital, and had fewer traditional symptom burdens than other diagnoses.”²⁹ The recommendations made by the committee are outlined in Table 5 and were graded as Class IIb, level of evidence B by the authors.²⁹

Table 5.

Referral Criteria General and Per Diseases, Based on Group A Studies

Neurocritical Care

One study retrospectively identified all patients admitted to neurocritical care for greater than 24 hours and who received a palliative care consultation over an 8-month period³⁰ (25 inpatients). The main focus of this study was to provide a qualitative description of the clinical issues addressed. They showed that patients receiving a palliative care consultation were more likely to be older, were female individuals, and belonged to White race.³⁰ Consultations occurred late and patients with palliative care consultations had longer intensive care unit (ICU) stays. The most common reasons for referral were as follows: (1) clarifying goals of care, (2) family support, (3) decision-making, (4) discharge planning, and (5) communication.³⁰ The 4 main clinical issues addressed in palliative care consultation included the following: (1) discussing prognosis, (2) eliciting patient and family values, (3) understanding medical options, and (4) identifying conflict.³⁰

A cross-sectional survey of neurointensivists aimed at elucidating their perceptions, practice patterns, and preferences for palliative care referral within the ICU was performed³¹ (424 inpatients). The most common reasons for referral included goals of care discussions, support with treatment decisions, and family/social support.³¹ Palliative care referrals were used more often in difficult cases in which the discussion of withdrawal of life-sustaining therapies was the focus.

Risk of Bias and Quality Appraisal

Among the group A publications (Triggers for Palliative Care Referral), the main sources for risk of bias were that the trigger criteria used were based on expert opinion or consensus. Despite this, the process used to evaluate the trigger criteria was objective and quantifiable, and the expert trigger criteria selected were based on the existing literature. In some publications, data were obtained retrospectively, which increases the risk of information or selection bias. It is possible that not all triggers for referral to specialized palliative care are documented equally in the medical record. Several of the studies were completed at tertiary care centers, which may create a selection bias; however, most advanced neurologic illnesses will be cared for in this type of setting (e.g., ALS or movement disorders). See Tables 1 and 2 for the level of evidence classification for individual studies.

Discussion

Our systematic review shows that many published referral triggers for nononcologic neurologic and neurosurgical conditions are based on expert consensus. There is no evidence to support general trigger criteria applicable to all nononcologic neurologic and neurosurgical conditions. However, recent research provides evidence-based triggers for referral to palliative care in neurodegenerative conditions as a whole with condition-specific triggers being identified for parkinsonism,^13-16 MS,¹⁷ ALS,¹⁸ and dementia.¹⁹ We found that most adult studies focus on diagnosis-specific triggers. In pediatrics, there exist consensus-based triggers for referral to palliative care,¹¹ but there are not yet any specific evidence-based referral triggers. There are also no evidence-based referral triggers identified for neurosurgical conditions.

Given the diversity of symptoms and illness trajectories among neurodegenerative conditions, it is not surprising that condition-specific triggers for referral, and the timing of referral, are likely to be most useful (Table 5). The triggers applied to neurodegenerative conditions, as a whole, are consistent with more general reasons for palliative care referral that have previously been published for medicine.^32,33 However, research has demonstrated that patients with nononcologic neurologic and neurosurgical illnesses have unique palliative care needs. These palliative care needs may occur from the time of diagnosis and throughout the illness trajectory and are not limited to the end of life. Condition-specific referral trigger criteria are likely to be more effective than general trigger criteria because the palliative care needs in neurologic illness are diverse and disease specific. Furthermore, condition-specific trigger criteria may be better able to account for the variability in rates of progression associated with nononcologic neurologic and neurosurgical illnesses. Disease-specific triggers can be tailored to be useful at different time points in the illness trajectory, such as the 2-staged triggers in the RADPAC-PD¹³ and the various referral triggers identified for MS.¹⁷ Although it may be more cumbersome to have condition-specific triggers for referral, more specific tools will be easier to apply with confidence in a clinical setting and are more likely to identify the most appropriate patients. The RADPAC-PD¹³ and NAT:PD¹⁵ show promising results regarding referral triggers but still require peer-reviewed publication demonstrating validity and reliability in a clinical setting.

An important issue that we identified was ambiguity regarding nomenclature. Most studies use terminology such as triggers or reasons for referral; however, some of these studies also focus on needs assessment tools. Within the literature, there is some cross-over regarding the type of tool being developed and the nomenclature used. This is most relevant when considering how to use tools in future studies or daily practice. In general, needs assessment tools are designed to assess the palliative care needs of a person and to track these needs over time to enable clinicians to determine the responsiveness of symptoms to an intervention. By contrast, tools that identify referral triggers are designed to identify who should be referred to palliative care services and when. An example of this ambiguous nomenclature is the NAT:PD,¹⁵ which contains the terms “needs assessment tool” in its title but was designed for identifying appropriate patients with PD to refer to neuropalliative care and has been used in a recent randomized controlled trial for the same reason. We acknowledge that the reason for the tool's title is largely historical because it was adapted from an older validated tool called the Needs Assessment Tool: Progressive Disease in Cancer.³⁴ However, future tools would benefit from carefully considered nomenclature to provide clarity for users.

This systematic review specifically included primary studies investigating referral triggers, but because of ambiguous nomenclature, it is possible that some studies outlining needs assessment tools or measurement of caregiver burden were excluded based on the inclusion criteria. Some examples of scales that may be used, in some clinical settings, as referral triggers but that were not included in this systematic review include the modified Edmonton Symptom Assessment System Scale for PD,³⁵ palliative outcome score-PD,³⁶ or palliative outcome score-MS-S,³⁷ and the PD caregiver burden questionnaire.³⁸ Furthermore, a recent systematic review¹⁰ was not included in our analysis because it did not meet the inclusion criteria. Nonetheless, this systematic review warrants discussion because it synthesized the literature pertaining to predictors of mortality and causes of death in PD and related disorders. The authors suggest guidance for updated hospice eligibility criteria in this population. These criteria include evidence of advanced disease (one of critical nutrition impairments in the previous year, life-threatening complications in the previous year, or motor symptoms poorly responsive to dopaminergic medications), rapid or accelerating motor dysfunction and disability, and advanced dementia meeting hospice referral criteria based on previously established dementia criteria.¹⁰ The above criteria are a very important contribution to the current literature but remain to be evaluated in a clinical setting.

Our systematic review investigated the evidence base for triggering referral to specialist palliative care services from a neurology care setting based on primary studies. Some studies included in the analysis^12,14 used proximity to death as a trigger criterion. It should be noted that there is consensus within the literature that neuropalliative care can and should be used to varying degrees throughout the illness trajectory and not reserved for the end of life. Despite these limitations, the identified triggers are still likely to be of value to this physician group because the triggers also serve as signals to the most appropriate time to incorporate different aspects of palliative care along the illness trajectory.

This systematic review compiled data from referral trigger criteria from more than 2,907 patients. Variable nomenclature poses challenges to the compilation of data and thus efficiency in the progression of useful evidence to direct treatment. That said, there is a growing body of evidence providing evidence-based referral triggers for neuropalliative care. Condition-specific triggers are likely to be the most effective tools for identifying the appropriate people and timing for referral to specialized palliative care. Future studies should continue to develop condition-specific triggers that will be applicable for research studies and clinical use to enhance the care we provide to people living with advanced neurologic conditions. Furthermore, future studies should also focus on development of referral triggers appropriate for use in pediatric neurology and neurosurgery.

Appendix. Authors

Study Funding

No targeted funding reported.

Disclosure

The authors report no disclosures relevant to the manuscript. Full disclosure form information provided by the authors is available with the full text of this article at Neurology.org/cp.