Palliative care in patients with haematological neoplasms: An integrative systematic review

Design and data sources

We performed a systematic review following the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines. ²² The study project was registered at PROSPERO, an international database of prospectively registered systematic reviews (registration #: CRD42016036240). The literature search was conducted in the following databases: MEDLINE PubMed, Cochrane, CINAHL, Scopus and Web of Science. We evaluated all publications dating from the establishment of the respective databases through 30 June 2015.

Eligibility criteria

We included all published studies (in English or Spanish language) containing information on palliative care interventions in adult patients with any haematological malignancy subtype at any disease stage. Given that the main aim was to obtain a better understanding of the relationship between palliative care and onco-haematology departments, no restrictions were placed on study design or assessment measures. However, editorials, letters to the editor, and papers dealing with cancer treatment and intensive care units (ICU) were excluded.

Search strategy and study selection

The search strategy for the MEDLINE database, which used both text words and MeSH/EMTREE terms, is shown in Table 1. For consistency, the search strategy was adapted to the specific language of each database. In addition, a manual search of the reference lists of identified articles was performed. Articles identified in the search were separately screened and assessed for inclusion by two reviewers (D.M-A. and J.P-S.). Next, the retrieved articles were sorted in three stages: first by title, second by abstract and finally by full text. Articles were excluded if they failed to meet the aforementioned inclusion criteria; any disagreements in this regard were resolved by discussion among the members of the research team. The flowchart in Figure 1 illustrates the article search and selection process and reasons for exclusion.

OPEN IN VIEWER

Table 1.

Search strategy.

#1	‘Palliative Care’ [Mesh]	Settings
#2	‘Hospice and Palliative Care Nursing’ [Mesh]
#3	‘Palliative Medicine’ [Mesh]
#4	‘Terminal Care’ [Mesh]
#5	‘Terminal Illness’ [Text Word]
#6	‘Dying’ [Text Word]
#7	‘End of life’ [Text Word]
#8	#1 OR #2 OR #3 OR #4 OR #5 OR #6 OR #7
#9	‘hematological malignanc*’ [Text Word]	Population
#10	‘hematologic neoplasms’ [Mesh]
#11	‘bone marrow neoplasms’ [Mesh]
#12	#9 OR #10 OR #11
#13	8 AND #12

OPEN IN VIEWER

Figure 1.

PRISMA flow diagram of study selection.

Data extraction, data analysis and synthesis of findings

We extracted and synthesized the data using the integrative method described by Whittemore and Knafl, ²³ a method that allows for the synthesis of diverse data sources, including both empirical (qualitative, quantitative and mixed methods) and theoretical studies (case reports, theory model, comments and reflexions). The data analysis process in this methodology integrates the findings of all the included studies. All studies that met the inclusion criteria were treated equally in this synthesis.

We also used the data analysis process proposed by Whittemore and Knafl, ²³ which includes data reduction, display and comparison, as well as conclusion drawing. The first step is to develop a matrix to extract the main characteristics of each article, including author(s); year of publication; country where the study was conducted; study design and methodology; population; setting; sample; topic; main aim and main outcomes. After analysing the matrix data, five themes related to the aims of this review emerged: (1) end-of-life (EOL) care issues; (2) the relationship between onco-haematology departments and palliative care departments and referral characteristics; (3) patient clinical characteristics; (4) experience of patients/families with the disease and care process and (5) home care. Themes that did not fit into one of these five categories were grouped under a 6th category denominated ‘miscellany’. Next, all articles that met the inclusion criteria were classified into one of these six themes and the articles were analysed by topic. Data for each subgroup classification were extracted from all primary data sources, regardless of the type of evidence, and compiled into the matrix. The third step was to compare the extracted data item-by-item to ensure that similar data were compared, categorized and synthesized. The data were managed manually to organize and facilitate the analysis.

Quality appraisal

We followed the PRISMA reporting guidelines to report the characteristics of the studies and their methodological quality. ²² Given the lack of consensus in the literature regarding the exclusion of studies for methodological reasons, ²⁴ we elected not to exclude any study for quality reasons.

EOL care issues

A total of 37 articles reported on EOL care issues. Of these, 15 (40.5%) were qualitative studies that primarily provided information about the opinions, beliefs and experiences of health professionals.^{20,34,35,39–41,43,46,48,51,60,61,79,94,101} In all, 15 studies (40.5%) used a quantitative approach, reporting data on 14,571 patients with haematological malignancy, with most of these 15 studies providing descriptive information about medical care at EOL^{27,36,47,71,76,77,87,109} and/or the cause and place of death.^{37,89,99,106,114,115,116} A variety of other methodologies were used, including a quali-quantitative approach, ⁷ a case report ⁸⁴ and reviews of medical care ¹⁰⁴ or comprehensive needs at EOL. ²⁶ Three reviews analysed the cause and/or place of death.^33,63,75 Notably, none of the aforementioned studies compared care strategies.

One of the most common issues identified in these articles was when EOL care should be initiated in haematological malignancies patients.^34,40,84,101 Interest in this question is not surprising given the complexities involved in identifying EOL stages in these patients. ⁴³ Several studies explained why it is so difficult to identify the EOL stage in haematological malignancies patients, the main reason being the physician’s hope that the patient will respond to a new treatment (even after multiple relapses) and the often rapid decline from stable disease to EOL – a common occurrence in haematological malignancies – which leaves little time to properly adjust the therapy to the new clinical situation. Many authors^{20,26,61,94,99,104,109,115} posit that this ‘hope’ may reflect unrealistic expectations about the potential benefits of the anticancer treatment.

Numerous authors pointed out that patients with haematological malignancies often receive more intensive and aggressive treatment at EOL – including chemotherapy, blood transfusions, antibiotics and ICU admission – than patients with non-haematological neoplasms,^{27,36,37,41,47,60,63,71,76,77,89,106,109,114–116} resulting in poor quality EOL care.^26,35,61,106 The most common place of death for haematological malignancies patients in these studies were acute care hospital wards and ICUs; notably, haematological malignancies patients are referred to hospices later than patients with non-haematological neoplasms.^{7,27,37,63,75,99,106,115,116} However, as one study found, when patients and families are aware that the patient is dying and there is a strong desire to die at home, then doing so is feasible. ³⁹

Given these EOL care issues, several authors^47,51,79 stress that closer collaboration between haematologists and palliative care specialists is necessary to improve EOL care in these patients. The role of the nurse in this process is considered especially important. ³³ Several authors propose specific improvements^46,48,87 including (1) better communication among health care professionals, (2) improved use of hospital resources, (3) education about palliative care for staff and family and (4) increased awareness that a multi-dimensional intervention involves not only medical care, but also emotional and spiritual interventions.

Relationships between onco-haematology departments and palliative care departments and referral characteristics to palliative care

In all, 48 studies reported on the relationship between palliative care and onco-haematology departments and on referrals from onco-haematology departments to palliative care. Of those, 15 (31.2%) were qualitative, with a focus on describing the difficulties involved in identifying natural transitions to palliative care goals, the relationship between the teams, and/or staff training.^{21,39–42,44,51,54,59,61,78,79,81,94} Other mixed methodology studies reported similar information.^45,104

In all, 17 quantitative studies (35.4%) provided data on 11,149 patients with haematological malignancies. The vast majority of these were descriptive studies, providing information about referrals to palliative care teams; in addition, some of these studies also compared the results to practices in patients with non-haematological neoplasms.^{8,19,25,27,49,56,58,67,93,97,100,102,105,109–111,116}

Seven narrative reviews about the referral process and the relationship between palliative care teams and onco-haematology department teams were retrieved.^{6,18,32,33,85,91,108} One systematic review about referral of haematological malignancies patients to palliative care and hospice ⁸⁰ was identified. Three narrative reviews assessed specific issues: (1) the role of palliative care in bone marrow transplantation (BMT), ⁶⁵ (2) a self-review by one author of her own articles ⁸³ and (3) quality of life (QOL) in myelodysplastic syndrome. ⁹⁶ Three case reports^30,84,88 were also identified.

One notable finding regarding the relationship between onco-haematology departments and palliative care departments – reported in multiple studies – was the reluctance of haematologists to refer patients to palliative care. There is general agreement among the authors of the studies in this review that several factors contribute to this reluctance, including different treatment goals; prognostic difficulties; palliative care specialists’ lack of training in haematology and the preference of haematologists to manage palliative care delivered to ‘their’ patients.^{6,18,19,21,27,65,80,85,91,112} Several impediments to palliative care have also been described, including haematologists having difficulties in arranging an appointment for their patients with the palliative care department, ⁴⁴ reimbursement issues^6,19,21 and low patient engagement in advanced care planning. ⁸¹ Nurses in BMT units and haematology wards also describe the divergence between their responsibility to prolong survival and the need to prepare patients for an EOL scenario; in this sense, nurses reported difficulties in shifting priorities between curative and palliative efforts.^33,54,94

Symptom control was the main reason why haematologists refer patients palliative care,^19,44,88 especially when pain control was poor.^45,67,102 In general, hospitalized patients with haematological malignancies are referred for palliative care later than inpatients with non-haematological neoplasms; in addition, haematological malignancies patients are transferred to hospices later, ¹⁰⁰ and with lower performance status, than non-haematological neoplasms patients. ¹¹⁶ The data across the reviewed studies are consistent in showing that the average time between referral to palliative care and death is a matter of days among haematological malignancies patients, whereas it is a question of months in non-haematological neoplasms patients.^{8,19,25,49,58,93,102,109} A recent study ¹⁰⁵ reported a greater integration between onco-haematology departments and palliative care departments; in patients with a survival time ⩾1 year, the likelihood of being referred to a palliative care team was six times greater than in those surviving <1 month from diagnosis, so patients with longer survival are more likely to be referred to palliative care. However, it should be noted that patients with haematological malignancies referred to a palliative home care team may have a greater risk of in-hospital death than other populations.^97,110

The most common causes for the delay in referring haematological malignancies patients to palliative care – apart from the reasons described above – are the inherent prognostic difficulties related to the unpredictable course of haematological malignancies, together with the effect of specific treatments, the absence of reliable prognostic tools and the short interval from potentially curative treatment to death.^{39–42,51,56,61,111}

Many of the studies reviewed here have suggested how the relationship between haematologists and palliative care specialists could be improved, with a particular emphasis on improving training and education^51,78,79 to enhance our understanding of the specific needs of these patients, thereby increasing the likelihood of referrals (thus facilitating an interdisciplinary approach).^{32,39–42,100,112} The potential benefits of early referral in terms of symptom control and support have been described by several authors.^26,30,65 Other authors emphasize the value of palliative care in improving patient QOL,^25,96 and in providing support to the family during the course of the illness and in bereavement.^30,41 For this reason, recommendations also include the need for improved coordination and communication among all members of the interdisciplinary team as well as with other hospital resources, community palliative care services ⁴⁵ and general practitioners ⁴⁴ in order to provide comprehensive care of haematological malignancies patients, ⁵¹ especially in EOL. ⁸⁴

Patient clinical characteristics

In all, 34 studies reported data on the clinical characteristics of haematological malignancies patients. Six (17.6%) qualitative studies were retrieved.^{39–41,51,70,81} In all, 19 studies (55.8%) were quantitative, providing data on 3,696 patients; most of these studies were descriptive in nature, and all reported data on the symptom burden in haematological malignancies according to the disease stage and as it compares to non-haematological neoplasms.^{8,25,27,47,62,64,67,68,76,82,87,90,92,95,98,100,103,113} Four mixed methodology articles were included.^50,66,86,104 In addition, three specific narrative reviews about BMT, ⁶⁵ myelodysplastic syndrome ⁹⁶ and advanced haematological malignancies ¹⁰⁸ were included. One article contained clinical guidelines on the treatment of multiple myeloma, including symptom (mainly pain) management. ⁸⁶ Finally, there was one case report on physical and emotional symptoms in a patient with acute leukaemia. ⁸⁸

Our findings show that the most common symptoms present in patients with haematological malignancies in the palliative stage are those associated with bone marrow failure (e.g. anaemia, thrombocytopenia and neutropenia) and with complications derived from this failure (e.g. fatigue, dyspnoea, bleeding, infections, mucositis, delirium, drowsiness, nausea and vomiting).^{8,25,27,28,47,50,62,64,66–68,82,90,92,95,98,100,102,103,108,113} Two studies^8,92 compared the symptoms experienced by patients with haematological malignancies and non-haematological neoplasms, reporting that symptoms in haematological malignancies patients are often more severe. Pain is frequent in haematological malignancies patients and is the most common reason for referral to palliative care, particularly in patients with multiple myeloma.^81,86,96 Nevertheless, one study found that haematological malignancies patients are less likely than those with non-haematological neoplasms to receive opioids. ⁷⁶ One study ⁸ found that delirium was more prevalent in haematological malignancies than in non-haematological neoplasms patients (41% vs 16%), while another study ¹⁰⁰ found the opposite (14% vs 21%). Bleeding was a common problem in nearly all the studies in this review, a finding that underscores the difficulty in identifying patients at risk of catastrophic bleeding. The use of blood products and routine blood counts in patients with haematological malignancies in the palliative stage remains controversial, as noted by many authors.^{28,64,65,70,96}

One study found that emotional distress in patients with haematological malignancies is often severe. ⁵¹ Several authors even compare the level of emotional distress in haematological malignancies to post-traumatic stress disorder, a finding that underscores its severity.^39–41 Vollmer et al. ⁸⁷ described how the characteristics of haematological malignancies often lead patients to have a strong subjective perception of being close to death, a finding that explains why haematological malignancy tends to be associated with high levels of emotional distress.

Experience of advanced stage patients and their families

A total of 19 articles reported on the subjective experience of patients (and their families) with advanced disease. Of these, 10 (52.6%) were qualitative, mainly providing information about EOL experience.^{34,39–41–43,51,79,81,101} One review evaluated the author’s own previously published qualitative work. ⁸³

Three^82,87,89 of these 19 reports (15.7%) were quantitative studies providing data on 406 patients, describing the distress associated with haematological malignancy and its symptoms. One study was a narrative review of QOL in patients with myelodysplastic syndromes. ⁹⁶ Another three^45,50,66 (15.7%) were mixed methodology studies, and one (5.2%) was a case report. ⁸⁸

Nearly, all of those studies^{34,39–43,82,83,101} found that haematological malignancies patients presented high levels of distress during the course of the disease, with multifactorial causes including physical,^51,66 psychological^50,87 and social ⁵¹ causes. Unsurprisingly, distress was greatest among patients with late-stage disease. Albrecht et al. ⁸⁸ highlighted the importance of early detection of distress, noting the key role of haematology nurses in detecting distress. In general, family members confirm the patients’ perspectives, describing a perceived lack of support from the haematologist, mainly at EOL and bereavement,^{34,39–43,82,83,101} and the important role of nurses at these times.^60,96 In contrast, patients and families describe a positive experience with palliative care teams, ⁸² especially with palliative care nurses.^45,79,95,96

Home care

In all, 12 articles reported data on home care. Of these, nine (75%) were quantitative studies, providing descriptive data on 1,410 patients about the feasibility of managing symptoms at home and how home care may reduce the need for hospital admissions.^{29,31,52,57,62,64,69,85,97,110} Two narrative reviews (16.6%) were also identified.^55,85 However, only one of these studies – which focused on EOL at home – used a qualitative methodology. ⁴²

Two studies^85,110 found that palliative home care decreases hospital admissions, reduces intra-hospital mortality and decreases the use of the emergency services. ⁹⁷ A report from Italy ⁵⁷ found that the cost of home care in haematological malignancies patients with an expected survival of <3 months was lower than the corresponding hospitalization charges, despite an increase in health resource use (health care professional visits, blood transfusions and days of care). By contrast, a study conducted in Greece ⁵² concluded that palliative home care for haematological patients is more expensive than hospital care. These reports notwithstanding, the general consensus among these authors appears to be that home care for patients with haematological malignancies and their families is feasible and should not be limited to EOL care,^{29,31,42,55,57,62,69} even when haemorrhagic complications are present. ⁶⁴

Overall, these 12 studies found that the presence of the following factors can facilitate the dying process at home: (1) the patient’s desire to remain at home and the caregiver’s agreement with this goal, (2) the availability of appropriate resources to offer the home care option, (3) trust in the feasibility of this option and (4) provision of adequate support and proactive intervention by an experienced palliative care team.

Miscellany

In all, 14 articles on a wide variety of palliative care-related topics were included. These articles used heterogeneous methodologies (quantitative and qualitative studies, reviews). Of the 14 studies, 5 (35.7%) concerned prognosis,^{38,53,73,90,107} three (21.4%) reported on QOL,^79,90,98 two (14.2%) investigated survival,^25,74 two (14.2%) explored spiritual issues,^46,48 one (7.1%) described mortality rates ⁷² and one study (7.1%) proposed specific criteria to define haematological malignancies with poor prognosis. ²⁸

Lassaunière et al. ²⁸ proposed several specific criteria (largely based on extensive clinical experience with such patients) to help identify patients that they denominated as ‘advanced’ haematological malignancies patients: age > 60 years, performance status, disease stage and sensitivity or resistance to chemotherapy. Kripp et al. ¹⁰⁷ carried out a retrospective study to determine the variables associated with poor prognosis, finding that all of the following were associated with poor prognosis: ECOG > 2, platelet < 90 × 10⁹/ L, LDH >248 µΙ/L (4.13 µkat/L), albumin < 30 g/L, need for blood transfusions (Yes/No) and opioid-based therapy. Potential prognostic variables for survival described in other studies include physician-assessed clinical prognosis and patient-reported symptom severity, especially drowsiness.^73,90 One study ⁵³ evaluated older patients with high-risk myelodysplastic syndromes or acute myeloid leukaemia to determine the poor prognostic factors associated with referral to palliative care, finding the following variables: female sex, no dependent children and presence of secondary leukaemia. The difficulty of making an accurate prognosis in haematological malignancies was stressed by several authors, underscoring the need for greater efforts to identify prognostic variables and/or to develop new tools.^38,73,107

Other studies describe other aspects related to prognosis and survival. Some authors^74,79 argue that the patient should be provided with realistic information about the disease and the prognosis. In addition, one study ²⁵ found that, in patients with recurrent disease, palliative care was not associated with decreased survival but rather that overall survival was greater in these patients.

Overall, the reported data indicate that QOL declines as the symptom burden increases.^68,98 Moreover, when QOL is poor, coping may also be difficult due to the emotional distress and the frequent presence of serious spiritual issues.^46,48 Finally, one study found that a greater symptom burden (and, consequently, low QOL) was related to a higher risk of death in an acute palliative care unit. ⁷²

Strengths and limitations

This study has several potential limitations. First, we may have overlooked one or more relevant studies although we believe that this is unlikely due to the rigorous and highly sensitive search strategy used in this review, which included a careful manual search of reference lists. Another potential limitation is the heterogeneity of the included studies in terms of methodologies, topics and quality – all of which present challenges for the analysis and synthesis. However, the integrative systematic review methodology minimizes the risk of overlooking relevant information. ¹²⁶

The main strength of this integrative review is the comprehensive search and inclusion of a broad selection of relevant publications on the subject. In addition, because this methodology allowed us to synthesize information from a wide range of sources, we believe that this contributes to a better understanding of the study phenomena from a wide range of viewpoints. Recently, other researchers have successfully applied this method in palliative care to better understand dignity at EOL. ¹²⁷ Perhaps the greatest strength of this review is that it updates current knowledge on a relatively new and controversial subject, thereby providing clinicians and researchers with a valuable reference source to continue developing palliative care interventions to improve the care of patients with haematological malignancies, as well as of their families.

Future research needs

The results of this review confirm the growing interest in assessing the role of palliative care in haematological malignancies and the relationship between onco-haematology departments and palliative care departments. Importantly, nearly 50% of the reviewed articles were published in the last 5 years, a finding that underscores just how topical this subject is. As noted in many of the studies included in this review, future improvement of palliative care in haematological malignancies patients will be guided by a growing evidence base. However, the evidence to date suggests that palliative care in haematological malignancies patients can be improved by the following: (1) the development of reliable predictors of survival to inform decision-making, (2) identification and testing of better strategies to improve communication and cooperation between the palliative care and onco-haematology departments to foster mutual understanding and trust and (3) the development of shared strategies for early palliative care intervention and assessment of the potential benefits or risks of such strategies.