Referral Criteria to Palliative Care for Patients with Heart Failure: A Systematic Review

Abstract

Background:

Heart failure patients have significant symptom burden, care needs, and often a progressive course to end-stage disease. Palliative care referrals may be helpful but it is currently unclear who and when patients should be referred. We conducted a systematic review of the literature to examine referral criteria for palliative care among heart failure patients.

Methods:

We searched Ovid, MEDLINE, Ovid Embase, and PubMed databases for articles in the English language from the inception of databases to January 17, 2019 related to palliative care referral in heart failure patients. Two investigators independently reviewed each citation for inclusion and then extracted the referral criteria. Referral criteria were then categorized thematically.

Results:

Of the 1199 citations in our initial search, 102 articles were included in the final sample. We identified 18 categories of referral criteria, including 7 needs-based criteria and 10 disease-based criteria. The most commonly discussed criterion was “Physical or emotional symptoms” (n=51 [50%]), followed by “Cardiac stage” (n=46 [45%]), “Hospital utilization” (n=38 [37%]), “Prognosis” (n=37 [36%]) and “Advanced cardiac therapies” (n=36 [35%]). Under cardiac stage, 31 (30%) articles suggested NYHA functional class ≥III and 12 (12%) recommended NYHA class ≥IV as cutoffs for referral. Prognosis of ≤1 year was mentioned in 21 (21%) articles as a potential trigger; few other criteria had specific cutoffs.

Conclusions:

This systematic review highlighted the lack of consensus regarding referral criteria for the involvement of palliative care in heart failure patients. Further research is needed to identify appropriate and timely triggers for palliative care referral.

Introduction

Heart failure is a global pandemic that affects 26 million people¹. Despite significant scientific advancements, in the majority of cases, it still remains a progressive incurable disease². Patients with advanced heart failure often suffer from a multitude of physical and psychological sequelae³ and encounter many complex decision-making needs necessitating sensitive communication throughout the unpredictable disease trajectory.

Palliative care is a philosophy of care that utilizes a holistic, multidisciplinary approach that seeks to improve quality of life through relief of suffering and stress by addressing the various (physical, psychosocial, spiritual) needs of patients and families when faced with life-threatening illness⁴. The World Health Organization has recommended “integrating palliative care into all relevant global disease control and health system plans” and “developing guidelines and tools on integrated palliative care across disease groups and levels of care, addressing ethical issues related to the provision of comprehensive palliative care”⁵. Over the past few years, there has been an increased recognition of the role that palliative care can play in improving health outcomes for patients with advanced heart failure^6–8. Various organizations in the U.S. and worldwide, such as the American College of Cardiology, International Society for Heart and Lung Transplantation, and European Society of Cardiology, have made recommendations supporting early integration of palliative care into heart failure management^{7, 9, 10}. Currently, palliative care involvement for patients with heart failure remains suboptimal, with referrals often being made too late in the disease course^11–13. This is partly attributed to the lack of standardized referral criteria, along with other factors such as heart failure’s unpredictable disease trajectory, variable healthcare providers’ attitudes and beliefs about palliative care, and variable access to specialty palliative care^{2, 11}.

Unlike the field of oncology where there has been a fair amount of literature supporting the idea that timely and appropriate palliative care referrals can be beneficial, such evidence appears to be lacking when it comes to heart failure. A clearer understanding of the current criteria used to initiate a palliative care referral would be helpful towards a) improving the rate of referral of appropriate patients for palliative care interventions in a timely manner, b) allowing programs to establish clear and effective measures for quality improvement purposes, c) providing researchers a uniformity of standard for which to design future trials involving heart failure and palliative care, and d) paving the way for policymakers to help set up guidelines for optimal clinical care including facilitating benchmarking between services providing cardiac care, as well as shaping the evolution of palliative care programs. We conducted a systematic review of the literature to examine referral criteria for palliative care among patients with heart failure. This would represent a first step towards developing a standardized set of appropriate and timely referral criteria.

Methods

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Literature Search

The institutional review board at MD Anderson Cancer Center provided approval to proceed without the need for full committee review. Our clinical librarian searched Ovid, MEDLINE, Ovid Embase, and PubMed databases for publications in the English language from the inception of databases to January 17, 2019. The following concepts were searched using subject headings and keywords as needed, “palliative care”, “palliative medicine”, “terminal care”, “hospice care”, “end-of-life care”, “advance care planning”, “referral and consultation”, “referral”, “consultation”, “patient transfer”, “delivery of health care, integrated”, “heart failure”, and “myocardial failure” etc. The search terms were combined by “or” if they represented similar concepts, and by “and” if they represented different concepts. The majority of case reports and conference abstracts were excluded. The complete MEDLINE Search strategy is detailed in Appendix I (Supplemental Material).

After the initial search, two investigators (Y.C., H.K.) independently reviewed the title and abstract of each citation for inclusion. We included all original studies, reviews, systematic reviews, guidelines, editorials, commentaries, and letters, and excluded duplicates, case reports, and conference abstracts. If either of the two investigators coded that article as related to referral criteria for palliative care then that publication was included for further review. After the preliminary round, we then retrieved the full text of each article of interest and again two investigators (Y.C., H.K.) independently reviewed each of the articles and coded each publication as to whether it contained referral criteria for palliative care for heart failure patients. Among the 211 articles, 80 (%), 22 (%) and 109 (%) had 2, 1 and 0 reviewers endorsement for inclusion, respectively. The rate of agreement was 189/211=90% (Kappa 0.79, P<0.001) A senior investigator (D.H.) then reviewed the discrepancies and provided arbitration for the 22 publications in question which were ultimately all included.

Data Collection

All articles were reviewed by a palliative care physician (Y.C.) who extracted the referral criteria. These criteria were then categorized based on their themes. Our categorization of the referral criteria can also be further classified under two major domains (Table 1): needs-based criteria (i.e., patient/caregiver supportive care concerns) and disease-based criteria (i.e., diagnosis, prognosis, progression, or utilization of healthcare resources), which are modeled after the oncology palliative care literature¹⁴. A second investigator (H.K.) provided a secondary review (of all articles) for consistency. There was disagreement on the categorization for only 2 of 102 articles. With the aid of a third investigator (D.H.), we arrived at an overall consensus. This systematic review follows the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guideline for reporting¹⁵ where applicable.

Table 1.

Palliative care referral criteria for heart failure

Needs-based criteria	N(%)	Domain
Physical or emotional symptoms2, 4, 6, 11, 12, 16–59 Mood disturbance, dyspnea, refractory pain, weight loss, ESAS score >7, fatigue, refractory angina, frailty, insomnia	51 (50)	Needs
Functional decline4, 12, 18, 21, 22, 26–28, 33–36, 38–40, 45, 53, 54, 60–67 Worsening quality of life, care dependency, cognitive decline, 6-minute walk test < 300 meters, validated performance status scale‡	26 (25)	Needs
Decision support2, 11, 17–19, 23, 28–31, 35, 38, 40, 45, 47, 49, 55, 67, 68 Advance care planning or goals of care, hospice referral, care coordination, request for physician aid-in-dying	19 (19)	Needs
Psychosocial11, 17, 18, 28–33, 35, 38, 40, 42, 47, 67 Social needs, supportive counseling, spiritual issues, financial concerns	15 (15)	Needs
Caregiver distress11, 18, 27, 30, 32, 35, 38, 42, 45, 48, 54, 55, 69	13 (13)	Needs
Cardiac cachexia4, 12, 21, 22, 27, 29, 34, 60–62, 65, 66	12 (12)	Needs
Request (by patient/family/care team) for palliative care4, 12, 18, 30, 38, 40, 45, 54, 55	9 (9)	Needs
Disease-based criteria	N(%)	Domain
Cardiac stage4, 16–38, 60–67, 70–83 Persistent NYHA stage III or IV or AHA stage C or D symptoms	46 (45)	Disease
Hospital utilization2, 4, 6, 12, 16–22, 25–30, 32, 33, 37, 38, 40, 44, 54, 55, 61–64, 67, 72, 76, 81, 84–88 Repeated hospital admissions, intensive care unit admission, prolonged length of stay	38 (37)	Disease
Prognosis4, 6, 12, 18–20, 25, 27, 29, 30, 33, 35–38, 40, 41, 46, 59, 60, 67, 69, 71, 74, 79, 86, 87, 89–97 “Surprise question,”* life expectancy < 1 year, validated risk score†	37 (36)	Disease
Advanced cardiac therapies2, 4, 6, 12, 17, 22, 27, 34, 38, 40, 43, 60–62, 65, 66, 68, 72, 73, 76, 79, 83–85, 98–108 Left ventricular assist device, dependence on inotropes or other intravenous therapy, heart transplant, implantable cardioverter-defibrillator, ineligible for or uninterested in advanced therapies	36 (35)	Disease
Disease progression4, 12, 21, 27, 30, 32–36, 40, 41, 46, 54, 56, 59–66, 69, 73, 75, 79, 81, 82, 85, 93, 108, 109 Inability to tolerate GDMT, frequent exacerbations despite GDMT, EF < 35%, Peak oxygen consumption < 14 cc/kg/min or < 60% predicted, EF < 20%, lack of further treatment options	33 (32)	Disease
Medical complications of heart failure4, 12, 16, 17, 19, 21, 25, 27, 34, 36, 38, 40, 41, 43, 46, 56, 62, 64–66, 72, 75, 85, 87, 88, 96, 109 Cardiorenal syndrome, cardiac arrest, hyponatremia, recurrent malignant arrhythmias, implantable device shock, persistent hypotension, persistent tachycardia, multi-organ failure, persistent S3 gallop	27 (26)	Disease
End of life17, 19, 22, 27–29, 35, 38, 42, 49, 61, 85, 99, 108, 110 Withdrawal of life-prolonging interventions, patient refusing treatment	15 (15)	Disease
Comorbidities4, 12, 18, 26, 34, 36, 38, 56, 67, 72, 77, 87, 96	13 (13)	Disease
New diagnosis of heart failure27, 30, 40, 43, 111	5 (5)	Disease
High risk procedures12, 17, 38, 40, 110 Cardiac surgery, percutaneous aortic valve replacement	5 (5)	Disease
Other12, 30, 32, 36, 44, 64, 78, 109 Advanced age, need for specialty assessment, specific further medical information, takes up cardiac team’s time	8 (8)	N/A

Sub-categories are listed beneath each category in order of frequency from highest to lowest.

Abbreviations: Edmonton Symptom Assessment Scale (ESAS), New York Heart Association (NYHA), American Heart Association (AHA), Guideline-Directed Medical Therapy (GDMT), Seattle Heart Failure Model (SHFM), North American Evaluation Study of Congestive Heart Failure and Pulmonary Artery Catheterization Effectiveness (ESCAPE), Get With The Guidelines-Heart Failure (GWTG-HF), Enhanced Feedback for Effective Cardiac Treatment (EFFECT), CardioVascular Medicine Heart Failure (CVM-HF), Palliative Performance Scale (PPS), World Health Organization (WHO)

^*“Surprise question” asked to clinicians is “Would you be surprised if this patient were to die within the next year?”

^†Elevated B-type natriuretic peptide, ESCAPE risk score > 50%, SHFM score > 85%, EFFECT score >33% 1-year mortality, GWTG-HF score > 58, CVM-HF index score > 12

^‡Karnofsky performance status < 40, low PPS score, WHO performance status score 2/3

Statistical Analysis

We summarized the data using descriptive statistics, including counts, frequencies, and percentages.

Results

Literature Search

Our literature search identified 1199 articles. Three hundred eighty-seven were excluded as they represented duplicates, case reports, or conference abstracts. A total of 812 articles were reviewed, with a final sample of 102 (12.5%) publications included (Figure 1). The characteristics of the included articles are summarized in Table 2. A majority of the articles were from the United States (n=51 [50%]) and were published after 2010 (n=76 [75%]). The most prevalent article type were narrative reviews (n=45 [44%]). There were also 17 (17%) retrospective studies, 12 (12%) prospective non-randomized studies, 4 (4%) qualitative studies, and 4 (4%) prospective randomized trials. Majority of the articles were published in either cardiology (n=42 [41%]) or palliative care journals (n=38 [37%]).

Figure 1.

Study Flowchart

Table 2.

Publication Characteristics

Article characteristics	N (%)
Article type
Editorials	6 (6)
Practice Guidelines	4 (4)
Letters	3 (3)
Prospective non-randomized studies	12 (12)
Prospective randomized trials	4 (4)
Qualitative studies	4 (4)
Retrospective studies	17 (17)
Review-narrative	45 (44)
Review-systematic	2 (2)
Trial design	5 (5)
Publication year
2003–2010	26 (25)
2011–2018	76 (75)
Journal type
Cardiology journals	42 (41)
Palliative care journals	38 (37)
General medical journals	8 (8)
Nursing journals	8 (8)
Other	6 (6)
Country
United States	51 (50)
United Kingdom	28 (27)
Canada	10 (10)
Sweden	2 (2)
The Netherlands	2 (2)
Ireland	2 (2)
Australia	1 (1)
China	1 (1)
Czech Republic	1 (1)
Denmark	1 (1)
Germany	1 (1)
Italy	1 (1)
Spain	1 (1)

Categories of Referral Criteria

We identified 18 categories of referral criteria, which were broadly grouped according to “needs-based” (n=7) and “disease-based” (n=10) (Table 1).

The most commonly discussed criterion was “Physical or emotional symptoms” (n=51 [50%]), followed by “Cardiac stage” (n=46 [45%]), “Hospital utilization” (n=38 [37%]), “Prognosis” (n=37 [36%]), and “Advanced cardiac therapies” (n=36 [35%]); most articles mentioned more than one criteria category (n=78 [76%]). Ninety-two of the articles (90%) did not make an identifiable distinction between inpatient palliative care referral criteria and outpatient palliative care referral criteria. Only 2 (2%) articles mentioned inpatient, 4 (4%) articles outpatient, and 4 (4%) articles with both inpatient and outpatient.

Physical and emotional symptoms to trigger referral

This was the most prevalent category with 51 articles (50%). The most commonly discussed symptom was “mood disturbance” (n=15 [15%]), followed by “dyspnea” (n=11 [11%]). There were 3 articles (3%) that discussed using the Edmonton Symptom Assessment Scale (ESAS), a validated symptom assessment tool, as a trigger for palliative care referral with a cutoff score of 7 or above (ESAS worst possible severity score is 10). However, 33 publications (32%) did not outline any specific symptoms to be considered as criteria, nor assessment scales or cutoffs.

Heart failure stage to trigger referral

Cardiac stage (NYHA functional class, AHA stages) was the next most prevalent category (n=46 [45%]). 43 articles (42%) recommended using NYHA functional class to trigger a referral. Thirty-one publications (30%) indicated NYHA functional class III and/or IV as referral criteria for palliative care consult and 12 articles (12%) indicated NYHA class IV only as a trigger to referral. Only 7 articles mentioned using AHA/ACC stages for referral (stages C/D, n=3 [3%]; stage D only, n=4 [4%]).

Hospital utilization as referral criteria

Thirty-eight articles (37%) comprise this category. Among these, eight (8%) articles specifically stated ≥2 hospitalizations in 12 months as a criterion, 3 (3%) articles referenced ≥2 admissions in the last 6 months, and 5 (5%) articles mentioned intensive care unit (ICU) admission as a possible referral criterion.

Prognosis as referral criteria

Among the 37 publications (36%) suggesting prognosis as a referral criterion (Table 4), the surprise question (“Would you be surprised if this patient were to die within the next year?”) was mentioned as a trigger for palliative care referral in 14 (14%) articles. Other investigators also suggested the use of mortality risk scores (n=10 [10%]). Seven (7%) articles listed life expectancy <1 year as referral criteria, whereas 4 articles (4%) were nonspecific in terms of defining prognosis.

Table 4.

Number of articles that recommended palliative care referral based on prognosis

Sub-category	N (% = N/37)
“Surprise question”*	14 (38)
Validated risk score†	10 (27)
Life expectancy < 1 year	7 (19)
Elevated BNP	4 (11)
Unspecified	4 (11)

Abbreviations: Seattle Heart Failure Model (SHFM), North American Evaluation Study of Congestive Heart Failure and Pulmonary Artery Catheterization Effectiveness (ESCAPE), Get With The Guidelines-Heart Failure (GWTG-HF), Enhanced Feedback for Effective Cardiac Treatment (EFFECT), CardioVascular Medicine Heart Failure (CVM-HF), B-type natriuretic peptide (BNP)

^*“Surprise question” asked to clinicians is “Would you be surprised if this patient were to die within the next year?”

^†ESCAPE risk score > 50%, SHFM score > 85%, EFFECT score >33% 1-year mortality, GWTG-HF score > 58, CVM-HF index score > 12

Advanced cardiac therapies as referral criteria

This category had 36 articles (35%) and included insertion of left ventricular assist device (LVAD) (n=18 [18%]), dependence on inotropes or other intravenous therapy (n=11 [11%]), consideration for heart transplants (n=5 [5%]), and implantable cardioverter-defibrillator placement (n=4 [4%]) as potential triggers. Twelve (12%) articles indicated that patients who were ineligible for or uninterested in advanced therapies should be considered for palliative care referral.

Disease progression as referral criteria

This category had a total of 33 articles (32%) and included heart failure exacerbations (n=11 [11%]), inability to tolerate goal directed medical therapy (n=8 [8%]), and low left ventricular ejection fraction (n=7 [7%]). Five articles (5%) specified EF <35% and 2 (2%) suggested EF <20% as triggers.

Medical complications of heart failure

This category had a total of 27 (26%) publications. Complications warranting referral to palliative care included cardiorenal syndrome (n=11 [11%]), cardiac arrest (n=9 [9%]), hyponatremia (n=7 [7%]), implantable device shocks (n=7 [7%]), recurrent malignant arrhythmias (n=6 [6%]), persistent hypotension (n=5 [5%]), persistent tachycardia (n=3 [3%]), and multi-organ failure (n=3 [3%]).

Functional Decline

“Functional decline” was mentioned in 26 (25%) articles with few highlighted specific criteria other than functional decline. Four articles list specific performance scales as possible referral criteria (Karnofsky performance scale < 50, n=2 [2%]; Palliative performance scale, n=1 [1%]; WHO performance status 2/3, n=1 [1%]). Two publications (2%) listed the 6-min walk test < 300m as a potential trigger for referral. Care dependency was also referenced by 7 articles (7%). Four publications (4%) mentioned cognitive decline.

Decision Support

This category included 19 articles (19%), with 16 (16%) highlighting need for advance care planning/goals of care as potential referral criteria. Assistance for hospice referral (n=4 [4%]), care coordination (n=2 [2%]), and request for physician aid in dying (n=2 [2%]) were also mentioned as potential triggers.

Other Referral Criteria

We identified nine other referral categories that were mentioned less commonly. Five of the nine categories were mentioned in ≤15% of the articles (“End of life”, n=15 [15%]; “Psychosocial”, n=15 [15%]; “Caregiver distress”, n=13 [13%]; “Comorbidities”, n=13 [13%]; and “Cardiac cachexia”, n=12 [12%]). The other four categories each had less than 10% of articles (“Request by patient/family/care team for palliative care”, n=9 [9%]; “New diagnosis of heart failure”, n=5 [5%]; “High risk procedures”, n=5 [5%]; and “Other”, n=8 [8%]). Please refer to Table 1 for specific subcategories, if available, for each of the above categories listed.

Discussion

We conducted a systematic review of the literature in order to examine what criteria have been considered for initiating a palliative care referral in heart failure. With our final sample of 102 articles, we discovered a wide array of reasons for which to refer heart failure patients to specialist palliative care. A large number of categories for referral criteria were identified, covering 18 different themes. This observation suggests that these articles recognized a multitude of opportunities along the disease trajectory to refer patients; however, it also underscores that there was no clearly established consensus. The most commonly noted criteria category was physical or emotional symptoms, followed by cardiac stage, hospital utilization, prognosis, advanced cardiac therapies, and disease progression. We found no clear universal consensus on referral criteria for patients with heart failure. As such, our goal was to map all the criteria found without preference for one set of criteria over another. Some investigators proposed that a palliative care referral should be considered based on a patient’s palliative care needs, while others support referrals based on disease trajectory or other factors. Findings from this study would facilitate further research to identify which of these criteria should be considered to trigger a referral.

“Physical or emotional symptoms” was the most prevalent referral category highlighting the increasing awareness of the high symptom burden among heart failure patients, and the role of palliative care in symptom management¹¹. Utilizing a holistic and inter-disciplinary approach⁴, specialist palliative care can provide support for patients and their families concurrently with cardiology teams. However, despite acknowledging the need for referral for symptom management, only a few articles have explicitly described which symptoms and how they should be assessed, as well as the cutoff for referral for specialist palliative care^{4, 24}. For example, 3 randomized controlled trials of antidepressants in patients who develop depression in the setting of heart failure have been negative, and a rigorous symptom-based, collaborative-care management approach did not significantly improve outcomes for patients with heart failure¹¹². Further research is needed.

Categories such as “Cardiac stage”, “Advanced cardiac therapies”, “Disease progression”, and “Medical complications of heart failure” were also commonly mentioned as potential criteria for palliative care referral. However, the specifics remain elusive (ex., inability to tolerate goal directed medical therapy) with only a few mentioning definitive referral cutoffs (ex., LVEF <35%)^{4, 73}. When a cutoff was proposed, the timing of palliative care referral seemed to be variable, as patients with stage D/NYHA class III-IV heart failure have a 1-year mortality rate of > 50%, and those on intravenous inotrope therapy have a median survival of 461 to 584 days^{113, 114}.

Among the articles that discussed prognosis as a referral criterion, a majority suggested a cutoff of 1 year as potential trigger, with multiple investigators recommending the use of the “surprise question” (“Would you be surprised if this patient were to die within the next year?”) for screening. However, prognostication of survival in heart failure is particularly challenging because of its unpredictable trajectory. The sensitivity, specificity and overall accuracy of the surprise question may depend on the patient population and clinician experience¹¹⁵. Quantitative risk models have poor performance at prospectively identifying individuals who will die in the next year¹¹⁶. And data shows that clinicians are equally inaccurate while also bringing additional biases to prognostication¹¹⁷.

Although multiple guidelines for patients with heart failure supported the integration of palliative care for patients with heart failure, there appeared to be no consensus between these professional bodies on when patients should be referred. For example, the 2013 ACC/AHA Guidelines for the Management of Heart Failure stated, “Consider palliative care or hospice care throughout the hospital stay, before discharge, at the first visit after discharge, and during follow-up in selected patients. Palliative and supportive care are effective for patients with symptomatic advanced heart failure to improve quality of life”⁷. The 2013 ISHLT Guidelines for Mechanical Circulatory Support suggested that a palliative care referral should considered during evaluation for mechanical circulatory support implantation⁹. The 2016 ESC Guidelines for the Diagnosis and Treatment of Acute and Chronic Heart Failure recommended that a palliative care approach “should be introduced early in the disease trajectory and increased as the disease progresses ”¹⁰. The European Association for Palliative Care Task Force recently published an expert position statement which supports the use of need-based criteria instead of time-based criteria for referral¹¹⁸.

Despite guidelines recommending early integration of palliative care into heart failure management, many of the suggested referral criteria identified in our study involved patients with advanced heart failure far into the disease trajectory. This could be in part due to primary palliative care being provided by non-palliative care specialists (as specialty palliative care is typically reserved for more complex and higher level palliative needs). Furthermore, throughout our systematic review, there were very few articles (10%) proposing a clear distinction between criteria meant for inpatient versus outpatient specialist palliative care referral. Since outpatient referrals facilitate earlier access to palliative care services, such timeliness is threatened by the lack of consensus and clarity in this area. Importantly, palliative care is most effective when provided earlier to allow for timely symptom management, longitudinal patient education and monitoring, psychological interventions, and support advance care planning^119–121; however, more randomized trials are needed to confirm this in the heart failure population. Similar to palliative care for cancer patients, referral criteria would ideally be focused on: 1) earlier timing of appropriate referral (and not rely solely on prognostication) and 2) consideration of needs-based criteria rather than disease-based criteria^{14, 122, 123}. Given limited available palliative care resources, and considering the large volume of potential heart failure patients, it will be impossible for specialist palliative care providers to see all the potential heart failure patients who are triggered by the 18 referral categories identified. Consequently, finding the most effective and eventually validated referral criteria at the appropriate time will be a critical factor to ensure that those with the most complex palliative needs are referred at the ideal time to specialist palliative care providers in order to optimize care. Further research is needed to identify the right time, the right reason/indication for palliative care referral for heart failure patients.

Our systematic review has several limitations. First, a large proportion of articles were narrative reviews based on expert opinion. There were very few empiric studies in the literature that examined how triggered referral using specific criteria impacted patient outcomes compared to routine referral. In a randomized clinical trial, Rogers et al. compared routine outpatient palliative care referral versus usual care for 150 patients with high Evaluation Study of Congestive Heart Failure and Pulmonary Artery Catheterization Effectiveness risk scores⁶. They found that timely palliative care involvement was associated with a significant improvement in quality of life, depression and anxiety. However, another randomized trial by Hopp et al. comparing inpatient palliative care referral to usual care for hospitalized patients with stage III/IV heart failure and a 1 year mortality risk of >=33% based on the Enhanced Feedback for Effective Cardiac Treatment score reported no significant difference in the proportion of patients choosing comfort-oriented care⁷⁴. Further research is needed to examine outcomes associated with automatic referral and our study represents only a first step to developing standardized referral criteria. Second, we did not conduct a hand search or search grey literature, thus, may not have captured all available data. Third, we included certain eligibility criteria of randomized trials as possible referral criteria, although, they may not have been specifically designed for this.

We identified 18 categories of palliative care referral criteria for patients with heart failure. These were broadly classified as either need-based criteria or disease-based criteria. This diversity highlights the many reasons to involve palliative care and at the same time the lack of consensus on who and when patients should be referred. As we move towards a more personalized approach to medical care, it is important to refer the right patient to the right service at the right time. Theoretically, this would help to triage the limited available palliative care resources, while concurrently standardizing appropriate and timely palliative care access. Referral criteria are being examined in cancer patients but few studies have been conducted in heart failure patients. Further research is needed to delineate appropriate and timely triggers for palliative care referral for heart failure patients.

Table 3.

Number of articles that recommended palliative care referral for physical or emotional symptoms

Sub-category	N (% = N/51)
Mood disturbance	15 (29)
Dyspnea	11 (22)
Fatigue	3 (6)
Refractory pain	4 (8)
Weight loss	4 (8)
Edmonton Symptom Assessment Scale > 7	3 (6)
Refractory angina	2 (4)
Insomnia	1 (2)
Frailty	1 (2)
Unspecified	33 (65)

Table 5.

Number of articles that recommended palliative care referral based on need for advanced cardiac therapies

Sub-category	N (% = N/36)
Consideration of left ventricular assist device	18 (50)
Ineligible for or uninterested in advanced therapies	12 (33)
Dependence on inotropes or other intravenous therapy	11 (31)
Consideration of heart transplant	5 (14)
Consideration of implantable cardioverter-defibrillator placement	4 (11)
Left ventricular assist device-associated complications	1 (3)
Rejection after transplant	1 (3)

CLINICAL PERSPECTIVE.

What is new?

• This systematic review identified eighteen categories of potential triggers for specialist palliative care referral, including 7 needs-based and 10 disease-based criteria.

• Our findings highlight the lack of consensus on who and when patients with heart failure should be referred to specialist palliative care, which may explain why palliative care access is currently highly variable and often delayed.

What are the clinical implications?

• This study serves as a first step towards developing a set of criteria to facilitate timely palliative care referral for patients with heart failure, with the long-term goals of standardizing palliative care access, personalizing care choices, improving patients’ quality of life, and triaging the limited palliative care resources.

Non-standard Abbreviations and Acronyms

ESAS

Edmonton Symptom Assessment Scale

NYHA

New York Heart Association

AHA

American Heart Association

GDMT

Guideline-Directed Medical Therapy

SHFM

Seattle Heart Failure Model

ESCAPE

North American Evaluation Study of Congestive Heart Failure and Pulmonary Artery Catheterization Effectiveness

GWTG-HF

Get With The Guidelines-Heart Failure

EFFECT

Enhanced Feedback for Effective Cardiac Treatment

CVM-HF

CardioVascular Medicine Heart Failure

PPS

Palliative Performance Scale

WHO

World Health Organization

BNP

B-type natriuretic peptide