The state of science in the study of cancer symptom clusters

Article Outline

• Abstract
• Introduction
• Methods
• Review
  • Definition of symptom clusters
  • Theoretical frameworks
    • The theory of unpleasant symptoms
    • The symptom management model
    • The Symptom Cluster in Children and Adolescents with Cancer
    • Cytokine-induced sickness behavior
  • Measurement of symptom clusters
    • Most-common symptom approach
    • All-possible symptom approach
  • The influence of symptom clusters on outcomes
    • Functional status
    • Quality of life
    • Mortality
    • Depression
  • Interventions for symptom clusters
    • Interventions with central symptoms
    • Intervention with multiple symptoms
• Conclusion
• Conflict of interest
• Acknowledgement
• References
• Copyright

Abstract 

Purpose

To provide an integrative review of the literature on the science of symptom clusters in patients with cancer and establish implications for future studies.

Methods

Sixty-one articles about cancer symptom clusters were selected for review from results of a search in MEDLINE, CINAHL, PsycINFO, Sociological Abstracts and Cochrane databases from 1950 to 2010.

Results

This review discusses the current research on the definitions, theoretical frameworks, measurements, outcomes, and interventions of symptom clusters in oncology. Although symptom clusters were identified as groups of several related and coexisted symptoms, researchers had different opinion on the least number of and relationships among symptoms in a cluster. Four theoretical frameworks were used, but none of them were specific to guide research in symptom clusters for general cancer population. Most-common symptom approach and all-possible symptom approach had their own characteristics and methods for cluster identification. Functional status and quality of life were major outcomes that were negatively associated with the number or severity of symptom clusters. Interventions with multiple or central symptoms in clusters were two potential ways to improve patients’ symptom experience.

Conclusions

Despite advances in understanding of symptom clusters, further research is needed to define clusters operationally, and to develop appropriate theoretical frameworks. Methods of cluster identification need further comparison to see which offers the best understanding of symptom clusters. More studies with cross-sectional or longitudinal designs are necessary to explore influences of symptom clusters on patient outcomes, and interventions on symptom clusters.

Keywords: Symptom experience, Symptom clusters, Oncology

Back to Article Outline

Introduction 

Symptoms are among the most-common reasons that patients seek healthcare (Rutledge and McGuire, 2004). Clinical experiences and studies have shown that cancer patients often experience multiple concurrent symptoms during disease trajectories (Dodd et al., 2001a, Given et al., 2001, Patrick et al., 2004). Symptom clusters occur, when these multiple concurrent symptoms are related to each other. Compared with single symptoms, symptom clusters have more complicated and synergetic detrimental influences on patient outcomes (Dodd et al., 2001a, Gift et al., 2004, Given et al., 2001). The number of studies investigating symptom clusters has greatly increased since the first paper regarding the effect of symptom clusters on oncology patients’ functional status, published in 2001 (Dodd et al., 2001a). The purpose of this paper is to provide an integrative review of the literature on cancer symptom clusters. The paper will discuss the definition, theoretical framework, measurement, outcome, and intervention for symptom clusters. Suggestions for further research in this field conclude the review.

Back to Article Outline

Methods 

The literature search was conducted in MEDLINE, CINAHL, PsycINFO, Sociological Abstracts and Cochrane Database, using key words: symptom clusters, multiple symptoms, concurrent symptoms, or constellation of symptoms, which were combined with cancer, oncology, neoplasm or tumor. The search was limited to articles published in English, within the timeframe from 1950 to January 2010. Reference lists and bibliographies from other published articles were used to find additional articles to review. A total of 426 abstracts were identified for initial review. Selection was based on the following inclusion and exclusion criteria. Any data-based reports addressing measurements, outcomes, or interventions of multiple symptoms and relationships between symptoms in oncology patients were included in the review. Theoretical articles about the concept of symptom clusters were also included as it is an important issue in symptom cluster research. Articles that did not examine the relationships between and among multiple symptoms in oncology patients were excluded from the review, along with duplicate articles.

A total of 61 articles were identified, of which 57 were data-based and 4 were theoretical. The publication dates of these articles ranged from 1999 to 2010. Among the 57 data-based articles, seventeen were secondary data analyses, and the others were not secondary data analyses; twenty-two were longitudinal research designs, and the others were cross-sectional research designs. Study populations varied. Twenty-six studies included patients with at least two types of cancer; fourteen studies included only breast cancer patients; eight studies included only lung cancer patients; two studies included only patients with brain tumors; two studies only enrolled patients with prostate cancer; another five studies contained patients with ovarian, colorectal, pancreatic cancer, or malignant hematological disorders, or adult survivors of childhood cancer respectively.

Back to Article Outline

Review 

Definition of symptom clusters 

Conceptual clarification is the foundation for building knowledge about symptom clusters in cancer (Dodd et al., 2004). Although most researchers agree that symptoms in a cluster are correlated with each other and coexistent (Barsevick, 2007, Dodd et al., 2001a, Kim et al., 2005), there is still disagreement about some essential elements in the definition of symptom clusters. For instance, different researchers have different understandings of the relationship between symptoms in a cluster. Some have identified the relationship by the correlation between and among symptoms (Gaston-Johansson et al., 1999, Gift et al., 2003). Others have measured the relationship based on the effect of symptoms on outcomes (Fox et al., 2007). Miaskowski et al. (2004) have also suggested that symptoms can be related to each other through a common mechanism or etiology. In addition, researchers disagree about whether a symptom can be shared by several different clusters. Most studies put a symptom exclusively in one cluster, but two studies were found to allow a symptom shared by several clusters (Aprile et al., 2008, Francoeur, 2005). Clarifying the meaning of relationships between and among symptoms in a cluster will be necessary to define the concept of symptom clusters.

Another discrepancy in the definition of symptom clusters is the minimum number of symptoms constituting a cluster. Dodd et al. (2001a) suggested that at least three symptoms constitute a cluster, but Kim et al. (2005) recommended a minimum requirement of only two symptoms. Many data-based studies have since shown that two symptoms clustered have negative influences on patient’s quality of life or functional status (Chen and Lin, 2007, Chow et al., 2007, Fox and Lyon, 2006, Fox and Lyon, 2007, Given et al., 2001, Walke et al., 2007), while others have supported at least three symptoms in a cluster (Bender et al., 2005, Chan et al., 2005). Additionally, it is also not well understood whether all symptoms in a cluster should be presented at the same time (Kim et al., 2008, Molassiotis et al., 2010). These discrepancies reflect different understandings of the concept of symptom clusters. Variations in study designs, cluster identification methods, and characteristics of study samples could also contribute to these discrepancies. Determining the clinical and theoretical significance of symptom clusters might clarify these issues in the definition of symptom clusters.

Theoretical frameworks 

The theory of unpleasant symptoms (TOUSs) has been utilized frequently in symptom cluster research (Chan et al., 2005, Fox and Lyon, 2006, Fox and Lyon, 2007, Fox et al., 2007, Gift et al., 2004, Gift et al., 2003, Hoffman et al., 2007). TOUS has three main reciprocal components: symptoms, influential factors, and performance (Lenz et al., 1997). According to the theory, each symptom has four dimensions: intensity, timing, level of distress perceived, and quality. Factors influencing symptoms include physiological, psychological, and situational antecedents. Performance is the consequence of the symptom experience, which includes functional and cognitive activities. This model provides a theoretical framework for research on symptom clusters by indicating multiplicative effects of multiple concurrent symptoms. A potential criticism of TOUS model is its focus on physical rather than psychological symptoms. This limitation is reflected in the studies using this model as their theoretical framework. Three studies strictly following this model only included physical symptoms in cluster analyses (Gift et al., 2004, Gift et al., 2003, Hoffman et al., 2007). By discounting the role of psychological symptoms, TOUS offers a less comprehensive understanding of the nature of symptom clusters than theories that take psychological symptoms into account.

Symptom management model (SMM) (recently renamed as symptom management theory) is based on the premise that effective management of any given symptom or group of symptoms should consider all three components (Dodd et al., 2001b, Humphreys et al., 2008). These three components are: symptom experience, symptom management strategies, and outcomes. Symptom experience comprises perception, evaluation, and response to symptoms. Symptom management involves dealing with negative outcomes through biomedical, professional and self-care strategies. Patient outcomes are the results of symptom experience and management, including functional status, quality of life, costs, and morbidity. Each component can be affected by the others. Although the concept of symptom clusters has been recently introduced into the model, the relationships among these multiple symptoms in a cluster are not addressed completely. This restriction is also manifested in the utilization rate of the model: only three studies found in the review adopted this model as the theoretical framework to guide research (Dodd et al., 2010, Dodd et al., 2001a, So et al., 2009).

The Symptom Cluster in Children and Adolescents with Cancer is the only theoretical framework to specifically address symptom clusters. This model has three components: antecedent, symptom cluster, and outcome (Hockenberry and Hooke, 2007). Personal, environmental, and disease factors are the antecedents that can influence children’s symptom experience. Three most common and related symptoms of pain, sleep, and fatigue are the essential components of this cluster. The consequences of this cluster are physical performance and behavioral changes. This framework might be successful in guiding research into symptom clusters of pain, sleep, and fatigue in children and adolescents with cancer. However, restricting the theoretical framework to one symptom cluster with three specific symptoms limits its usefulness for broader symptom cluster research, as there are many other symptom clusters in children with cancer (Yeh et al., 2008). In addition, concentration on children and adolescents impedes utilization of this framework in adult and senior populations.

Cytokine-induced sickness behavior has been proposed by some authors as a possible explanation for the biological mechanism of symptom clusters (Chen and Tseng, 2006, Cleeland et al., 2003, Francoeur, 2005). Sickness behaviors refer to physiological and behavioral responses that can be induced in animal models after the administration of infectious or inflammatory agents (Hart, 1987, Hart, 1991, Watkins and Maier, 2000, Yirmiya, 1996). Physical changes include fever, pain, and increased activity in the hypothalamic–pituitary–adrenal axis and the autonomic nervous system (Watkins and Maier, 2000). Observed behavior changes consist of decreased activity, appetite loss, somnolence, and cognitive impairment (Yirmiya, 1996). Although sickness behavior represents a potential mechanism to explain some symptom clusters in cancer patients, there are still some limitations in the model as a guide for research. For instance, this model cannot explain many other symptoms that are not included in sickness behaviors (Aprile et al., 2008, Bender et al., 2005, Dodd et al., 2001b). In addition, it is difficult to use this model to guide the research in cases where sickness behaviors are separated into several different clusters (Chow et al., 2007, Molassiotis et al., 2010, Walke et al., 2007, Yeh et al., 2008), because it might be assumed that symptoms caused by the same underlying biomedical mechanism would be grouped in the same cluster.

All of these four theoretical frameworks identify the importance of symptoms in disease trajectories. Each framework has its own strengths and limitations in regard to symptom clusters. As general symptom theories, TOUS and SMM focus on multiple symptoms, but the relationships among these symptoms are not clearly addressed. While Symptom Cluster in Children and Adolescents with Cancer is specifically a symptom cluster theory, its specificity limits its use in a broad range of research. Cytokine-Induced Sickness Behavior provides a possible way to understand the underlying biological mechanism for certain symptom clusters. With the limitation of each theoretical framework, comprehensive theoretical models specifically focusing on cancer symptom clusters are still needed to guide further clinical research.

Measurement of symptom clusters 

There are four main characteristics in the most-common symptom approach (see Table 1). First, researchers often select several most-common symptoms in cluster identification, such as pain, fatigue, insomnia and depression. Second, researchers assume that these most-common symptoms be grouped together as a cluster before empirical studies. Third, generally, symptoms selected in cluster identification constitute a single cluster as the results of analytic techniques used to identify clusters. Fourth, the number of symptoms in a cluster is small, with most having 2–3 symptoms. Although the most-common symptom approach presents a way to understand symptom clusters, the main limitation to this approach is whether it is sufficient to select only the most-common symptoms in cluster identification. Since symptoms selected in cluster identification directly determine cluster results, adding or deleting any symptom could change the cluster result. If there is no sound theoretical foundation for selecting only the most-common symptoms in cluster identification, results from this approach might be neither reasonable nor reliable. Table 2 gives an overview of the 20 studies that have used a most-common symptom approach.

Table 1. Comparison of most-common symptom approach and all-possible symptom approach.

Table 2. Studies identifying symptom clusters in cancer patients by most-common symptom approach.

The cluster identification method used most in this approach is clustering by correlations between symptoms (Chan et al., 2005, Dodd et al., 2001a, Fox and Lyon, 2006, Fox and Lyon, 2007, Fox et al., 2007, Gaston-Johansson et al., 1999, Hoffman et al., 2007, Miaskowski and Lee, 1999, So et al., 2009). The correlation is usually calculated by correlation coefficients. For example, Fox and Lyon (2007) explored the relationship between pain, fatigue, and depression in 76 patients with ovarian cancer, and found fatigue and depression were grouped as a cluster because these two symptoms were correlated significantly with each other. The majority of these studies further supported symptom clusters by showing the influence of these clusters on patient outcomes, such as QOL and functional status (Dodd et al., 2001a, Fox and Lyon, 2006, Fox and Lyon, 2007, Fox et al., 2007, Gaston-Johansson et al., 1999, So et al., 2009). Although most studies identified a single symptom cluster, symptoms involved in these single clusters varied at different studies (Chan et al., 2005, Fox and Lyon, 2006, Gaston-Johansson et al., 1999, Hoffman et al., 2007, So et al., 2009). The main reason for this might be that researchers chose different most-common symptoms according to different types of cancer and treatment. In addition, small sample sizes (Chan et al., 2005, Fox and Lyon, 2006) and using individual items from QOL instruments as proxy measures for patients’ symptoms (Chan et al., 2005, Dodd et al., 2001a, Fox and Lyon, 2006, Fox and Lyon, 2007) may also decrease the evidence of significance from this method.

Yet another method of symptom cluster identification, clustering by concurrent multiple symptoms, was used in five studies (Given et al., 2001, Given et al., 2001, Liu et al., 2009, Reyes-Gibby et al., 2006, Wilmoth et al., 2009). By this way, the identification of clusters does not need any statistical analyses, but only needs the co-occurrence of selected related symptoms. Three of these studies further identified the synergistic effect of multiple concurrent symptoms on patient outcomes (Given et al., 2001, Given et al., 2001, Liu et al., 2009). Compared with patients who reported no pain, fatigue or insomnia, those reporting two or three symptoms had a higher risk of lower functional status (Given et al., 2001a). Patients with more symptoms also experienced more additional symptoms or more severe symptoms than those who reported only one or neither symptom (Given et al., 2001, Liu et al., 2009). However, as this method identifies a cluster only by the concurrent characteristic of selected symptoms, it is difficult to exclude the possibility that some unselected concurrent symptoms may be related to these selected symptoms, and thus should also be included in the cluster.

Some studies further explored the nature of symptom clusters or how symptoms were related to each other in a cluster by mediation effects and interaction effects. Two studies investigated the mediation effect between symptoms. The mediation effect indicates that the effect of one symptom on another can be adjusted by a mediator (Baron and Kenny, 1986). In a study of 84 cancer patients with pain, Beck et al. (2005) showed that sleep disturbance was a mediator between pain and fatigue, and that a 35% effect of pain on fatigue was accounted by sleep disturbance. Another study of two symptoms of fatigue and depression used functional status as a mediator, and found fatigue had a direct and an indirect influence on depression by the mediator of functional status (Barsevick et al., 2006). However, both of the studies were cross-sectional, and one used secondary data. With these study designs, it is difficult to assume the causal relationships of the mediation model, especially when there is not a strong theoretical framework to guide analyses (Polit and Beck, 2004). In addition, neither of these studies mentioned the control of other variables, such as age, co-morbidities, or other symptoms. If these variables had an influence on the relationship between symptoms, the mediation effect would change.

Three other studies explored interaction effects within symptom clusters. The interaction effect postulates that “the differing effect of one independent variable on the dependent variable depends on the particular level of another independent variable” (Cozby, 1997). In a study of 268 cancer patients in the initial phase of palliative radiation, Francoeur (2005) identified interaction effects in a cluster of pain, fatigue, and depressive affect. When pain and fatigue were lower, the depressive affect was lower. When pain was high, even lower fatigue could yield a higher depressive affect. Two other studies examined a same symptom cluster of pain, fatigue and insomnia. However, one study found a three-way interaction effect (Hoffman et al., 2007), while the other did not find significant interactions (Dodd et al., 2001a). The main reason for the inconsistency could come from the different dependent variables in both regression models, with one being gender, and another functional status. In addition, different study populations: one study focused on lung cancer patients, the other on general cancer patients, might also explain the inconsistency.

Another method is to identify subgroups of patients with similar symptom experience based on a specific symptom cluster. Unlike the previous presented methods, this method clusters patients together instead of symptoms. Miaskowski et al. (2006) studied 191 patients with cancer, and identified four subgroups of patients using cluster analysis: high fatigue and low pain, low fatigue and high pain, all symptoms low, and all symptoms high. This study further found that the subgroup of patients who reported low levels of all symptoms reported the best functional status and QOL. The findings from this study were further confirmed by two recent investigations either cross-sectionally (Pud et al., 2008) or longitudinally (Dodd et al., 2010). Although several distinct subgroups of patients with similar symptom experience were identified in these studies, the differences in most demographic and clinical characteristics among these subgroups have not yet been demonstrated (Dodd et al., 2010, Miaskowski et al., 2006, Pud et al., 2008). However, the findings might still benefit clinical practice by giving subgroups of patients different interventions based on their diverse symptom experiences.

In the most-common symptom approach, researchers assume several most-common symptoms might be clustered together prior to empirical studies. Methods of cluster identification include clustering by correlations between symptoms, by concurrent symptoms, by mediation effect, by interaction effect, and by subgroups of patients with similar symptom cluster profiles. Among them, mediation effect and interaction effect can also be used to explore the nature of symptom clusters. Identifying subgroups of patients based on a cluster may provide an easy way to translate the results from this method to clinical practice because patients with similar symptom profiles are identified. Since the results of symptom clusters are based mainly on the symptoms selected into cluster identification, extra effort should be paid to the rationale of including and excluding symptoms in identification process.

Compared with the most-common symptom approach, more studies recently have used the all-possible symptom approach to explore cancer symptom clusters. The main characteristics of this approach are opposite to those of the most-common symptom approach (see Table 1). First, researchers often target all potential symptoms that cancer patients might experience to identify clusters rather than simply select the most-common symptoms. Second, researchers obtain results of symptom clusters after statistical analysis, instead of assuming clusters before empirical studies. Third, the number and type of symptom clusters are more than those in the most-common symptom approach. For instance, researchers found not only the fatigue cluster, which is the focus in the most-common symptom approach, but also emotional clusters, gastrointestinal clusters and others (Bender et al., 2008, Chen and Lin, 2007). Fourth, the number of symptoms in a cluster is greater than that in the common-symptoms approach, with more than 4 clusters in many studies. These distinct characteristics in the all-possible symptom approach make the methods of cluster identification also different from those in the most-common symptom approach. Table 3 provides a summary of a total of 37 studies that have used an all-possible symptom approach.

Table 3. Studies identifying symptom clusters in cancer patients by all-possible symptom approach.

Note: Cluster identifying methods were not presented clear.

In the all-possible symptom approach, the method used most to identify symptom clusters is clustering symptoms by underlying factors or components. Literature in this review indicates that factor analysis, principle component analysis (PCA), and cluster analysis are among the most likely to cluster symptoms (see Table 3). Each of these methods can reduce the number of symptoms by putting several related symptoms under one group that is relatively independent of the other groups. For instance, Gleason et al. (2007) identified two symptom clusters: language and mood, from 12 symptoms in 66 patients with newly diagnosed primary or metastatic brain tumors. However, various cluster results have existed across studies because of different study populations, questionnaires, and statistical methods (see Table 3). In addition, it is possible that symptom clusters found in this way might not have a rational explanation because cluster results are based on factors or components from statistical procedures. In order to prevent this disadvantage, researchers have to adjust cluster results until they have clinical significance. This adjustment, although still conducted by statistical methods, is based on researchers’ understanding of symptom clusters, which might be a threat to the objectivity of the results and thus also cause discrepancies in cluster results.

Clustering by a temporal pattern of symptoms provides a way to understand changes in symptom severity over time. Wang et al. (2006) conducted a longitudinal study of 64 patients who had locally advanced lung cancer and had undergone concurrent chemoradiation therapy (CXRT). Four symptom cluster patterns appeared during CXRT: steady increase: pain and sore throat; early increase: nausea and vomiting; early/late increase: fatigue, lack of appetite, drowsiness, sleep disturbance, dry mouth, and distress; and minimal change: sadness, difficulty remembering, and others. Although other longitudinal studies have been conducted to identify symptom clusters’ changes (Chow et al., 2007, Gift et al., 2003, Gleason et al., 2007, Jarden et al., 2009, Kim et al., 2009a, Kim et al., 2008, Molassiotis et al., 2010, Reyes-Gibby et al., 2006), this is the only study identifying cluster results based on the symptom severity changes over time (Wang et al., 2006). The underlying mechanism and clinical meanings of the results from this method are still unclear.

Finding central symptoms provides a possible way of identifying symptom clusters, as well as exploring the nature of symptom clusters in the all-possible symptom approach. By calculating the strength of association between every two symptoms in 300 patients with colorectal cancer, Aprile et al. (2008) found six central symptoms: fever, dehydration, fatigue, anorexia, pain, and weight loss, among 25 most frequent symptoms. These six central symptoms were linked strongly to at least 5 of the 25 most frequent symptoms. However, compared to most other studies using patient-reported symptoms questionnaires, this study recorded symptoms using National Cancer Institute Common Toxicity Criteria. Considering the different understanding of symptoms between clinicians and patients (Basch et al., 2006, Bruner, 2007), it might be difficult to compare this result recorded by clinicians with those reported by patients.

Investigating causal connections among symptoms offers another possibility to examine symptom clusters. Olson et al. (2008) conducted a longitudinal study for 82 cancer patients receiving palliative care. Structural equation modeling (SEM) was used to identify the causal connection among symptoms. As required by SEM, an initial model of exogenous variables, pain, anxiety, nausea, shortness of breath and drowsiness, and endogenous variables, appetite, tiredness (fatigue), depression and well-being, was built before data analyses. The model fit was acceptable. Several causal relationships were reported: drowsiness displayed consistent effects on appetite, tiredness and well-being; anxiety’s effect on well-being changed over time. Although this model explained part of causal connections, some limitations might blur the relationship. For instance, the initial model was not based on published studies, but on researchers’ clinical experience. If a robust initial model could be applied in SEM, exploring causal relationship over time is an interesting way for symptom cluster research.

Clustering by subgroups of patients with similar symptom experience has also been used in the all-possible symptom approach. Most studies used cluster analysis for the identification of patients with similar symptom profiles (Capp et al., 2009, Gwede et al., 2008, Maliski et al., 2008), and one recent study used latent variable mixture modeling (Finnegan et al., 2009). Since more symptoms are included in the all-possible symptom approach, some researchers try to find central symptoms by identifying the most important symptoms in discriminating subgroups of patients (Capp et al., 2009, Ferreira et al., 2008). Ferreira et al. (2008) investigated 115 outpatients with cancer. These patients were separated into two subgroups: a subgroup with multiple and severe symptoms, and a subgroup with fewer symptoms and less severity. Subsequent discriminant analysis revealed that lack of appetite, depression, constipation, and insomnia were the most important symptoms for discrimination between the two subgroups.

In the all-possible symptom approach, any potential symptoms that cancer patients experience might be included in cluster identification, and symptom clusters are only determined after statistical analyses. The most-common method is clustering by underlying factors or components. Others include clustering by temporal pattern of symptoms, clustering by central symptoms in clusters, clustering by causal connection of symptoms, and clustering by subgroups of patients with similar symptom profiles. Since symptom clusters identified in this approach are determined mainly by statistical methods, researchers should pay attention to the clinical significance of cluster results. However, researchers’ understanding of the clinical meanings may also decrease the objectivity of this approach.

The influence of symptom clusters on outcomes 

There is preponderance of data that has demonstrated the negative associations between functional status and the number or severity of symptom clusters (Barsevick et al., 2006, Chen and Lin, 2007, Chen and Tseng, 2006, Dodd et al., 2001a, Ferreira et al., 2008, Fox et al., 2007, Gift et al., 2004, Given et al., 2001, Walke et al., 2007). Miaskowski et al. (2006) found that patients who reported low levels of all four symptoms (pain, sleep disturbance, fatigue, and depression) reported the best functional status. This result was confirmed by Pud et al.’s (2008) cross-sectional investigation and Dodd et al.’s (2010) longitudinal research. A study further found that symptom clusters explained 17.4–19.8% of the variance in functional status (Suwisith et al., 2008). Although relationships between symptom clusters and functional status have been identified, studies vary in terms of type of clusters that have impacts on functional status (see Table 2, Table 3). In addition, the longitudinal influence of specific symptom clusters on cancer patients’ functional status is still unclear.

Relationships between QOL and symptom clusters are similar to the connection between functional status and symptom clusters. Several studies identified that patients in the multiple and severe symptom cluster groups were more likely to have poor QOL (Dodd et al., 2010, Ferreira et al., 2008, Finnegan et al., 2009, Gwede et al., 2008, Miaskowski et al., 2006, Pud et al., 2008, So et al., 2009). In addition, one specific cluster, the cluster of fatigue and depression, appears to have a consistent negative influence on QOL, and explains variance in QOL from 29% to 42% across different studies (Fox and Lyon, 2006, Fox and Lyon, 2007, Gaston-Johansson et al., 1999). However, because only two other studies investigated the impact of other specific clusters (Fox et al., 2007, Ridner, 2005), such as physical and emotional clusters, on QOL, it is still not clear the association between other specific clusters and QOL. Furthermore, although three studies used a longitudinal design (Dodd et al., 2010, Gwede et al., 2008, Kenne Sarenmalm et al., 2008), the temporal effect of specific symptom clusters on QOL in cancer patients is not well-developed.

According to SMM, mortality is one of the outcome indictors in symptom experience (Dodd et al., 2001b). In a longitudinal study to identify symptom clusters in 112 patients newly diagnosed with lung cancer, Gift et al. (2003) found death 6–19 months after diagnosis was predicted not only by age, stage of cancer at diagnosis, but also by symptom severity at 6 months post-diagnosis. However, this study was a secondary data analysis and only physical symptoms were assessed in the cluster identification. In another study of patients with advanced colorectal carcinoma (CRC), although the authors did not explore symptom clusters, mortality was found to be associated with the simultaneous occurrence of multiple symptoms from chemotherapies (Delaunoit et al., 2004). The findings from the two studies suggest the possible influence of multiple symptoms on death rates, but the effect of symptom clusters on mortality is not fully understood.

Although most studies have explored depression as one of the symptoms in a cluster (Barsevick et al., 2006, Chow et al., 2007, Ferreira et al., 2008), two studies examined the impact of symptom clusters on depression (Breen et al., 2009, Francoeur, 2005). In a cross-sectional study, several symptom clusters, such as pain and fatigue, pain and fever, and nausea and fever, had interaction effects on depressive affect (Francoeur, 2005). A recent study found that general malaise, nutritional, and gastrointestinal clusters were independent predictors of depression (Breen et al., 2009). However, as cluster results and sample populations are different in the two studies, it is difficult to compare the findings. In addition, using single-item measurements for symptoms in one of the studies might be threat to the result validity (Francoeur, 2005). Moreover, whether depression should be studied as a symptom in clusters or as an outcome of symptom clusters needs further exploration.

The influence of symptom clusters on patients’ outcomes is a significant indictor to assess the importance of symptom clusters research. Four outcomes are explored in symptom cluster research. Functional status and QOL are always negatively associated with the number or severity of symptom clusters. Severity of certain symptom clusters predicts death in some cancer patients (Gift et al., 2003). Depression is explored as the outcome of several symptom clusters. Based on the review here, both cross-sectional and longitudinal effect of symptom clusters on cancer patients’ outcomes are still not investigated thoroughly. Understanding the effect of symptom clusters on cancer patients’ outcomes will not only help to support the science of symptom clusters, but also help to develop the intervention for symptom clusters.

Interventions for symptom clusters 

The identification of central symptoms in a cluster mentioned previously provides a possible method of intervention based on symptom clusters. Kwekkeboom et al. (2008) conducted a pilot study to relieve cancer pain by progressive muscle relaxation (PMR) and analgesic imagery interventions, and effects of these interventions on other concurrent symptoms were explored as well. Both PMR and analgesic imagery interventions produced great improvements in cancer pain. Patients who achieved a meaningful improvement in pain with analgesic imagery reported fewer concurrent symptoms than those who did not. Although this was a pilot study and the research purpose did not involve symptom clusters, it suggests that control of the central symptoms benefits other multiple concurrent symptoms.

Intervention with multiple related symptoms simultaneously is another way for the management of symptom clusters. One randomized clinical trial explored the effect of an exercise-based multimodal intervention for treatment-related symptom clusters over time in forty-two patients with cancer (Jarden et al., 2009). A symptom assessment tool was completed weekly during hospitalization, and at three and six months after treatment. Five clusters were identified by the principle component analysis, and then the comparison of the symptom cluster intensity scores longitudinally between groups was performed by general estimate equations. Symptom intensity over time in the intervention group was reduced significantly for all clusters except the affective cluster, while the reduction was not observed in the control group. This study provides evidence for an effective intervention in reducing the intensity of symptom clusters. However, using an invalidated questionnaire to measure symptoms is a major limitation of the study. The sample size should also be taken into account as a threat to the study, especially when the complex statistical methods require a large sample size.

The studies mentioned above suggest two possible methods of intervention in symptom cluster management. Because of the limited research in symptom cluster management, it is not clear which of these two methods might be more effective for certain symptom clusters. With the development of the science of symptom clusters, it is also possible that there are other approaches to treat symptom clusters. In addition, whether the outcomes of symptom clusters could be improved by intervention has not been examined. If the effect of interventions on outcomes of symptom clusters could be identified, the intervention would have more clinical significance.

Back to Article Outline

Conclusion 

As clinical experience and understanding of symptom clusters have grown, it is apparent that the science of symptom clusters is important in oncology. Despite all the advances in the understanding of symptom clusters in oncology, many questions remain unanswered. Further research is needed to define a symptom cluster operationally, and to develop appropriate theoretical frameworks for the research into symptom clusters in oncology. Methods of cluster identification need further comparison to see which one offers the best evidence of symptom clusters. More studies with cross-sectional or longitudinal designs and diversified study populations are necessary to explore the influence of symptom clusters on patient outcomes. The methods and effects of intervention for symptom clusters require extra examination. Understanding the complex symptom experience of cancer patients might provide a scientific foundation for clinical healthcare aimed at improving health outcomes in this large patient population.

Back to Article Outline

Conflict of interest 

There is no conflict of interest.

Back to Article Outline

Acknowledgement 

The author appreciates very much the support from her qualifying examination committee: Dr. Sarah Kagan, Dr. Deborah W. Bruner, and Dr. Lorraine Tulman at School of Nursing University of Pennsylvania.

Back to Article Outline

References 

1. Aprile G, Ramoni M, Keefe D, Sonis S. Application of distance matrices to define associations between acute toxicities in colorectal cancer patients receiving chemotherapy. Cancer. 2008;112(2):284–292
   • View In Article
   • CrossRef
2. Baron RM, Kenny DA. The moderator–mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology. 1986;51(6):1173–1182
   • View In Article
   • MEDLINE
   • CrossRef
3. Barsevick AM. The elusive concept of the symptom cluster. Oncology Nursing Forum. 2007;34(5):971–980
   • View In Article
4. Barsevick AM, Dudley WN, Beck SL. Cancer-related fatigue, depressive symptoms, and functional status: a mediation model. Nursing Research. 2006;55(5):366–372
   • View In Article
   • MEDLINE
   • CrossRef
5. Basch E, Iasonos A, McDonough T, Barz A, Culkin A, Kris MG, et al. Patient versus clinician symptom reporting using the National Cancer Institute Common Terminology Criteria for Adverse Events: results of a questionnaire-based study. Lancet Oncology. 2006;7(11):903–909
   • View In Article
6. Beck SL, Dudley WN, Barsevick A. Pain, sleep disturbance, and fatigue in patients with cancer: using a mediation model to test a symptom cluster. Oncology Nursing Forum. 2005;32(3):542
   • View In Article
7. Bender CM, Engberg SJ, Donovan HS, Cohen SM, Houze MP, Rosenzweig MQ, et al. Symptom clusters in adults with chronic health problems and cancer as a comorbidity. Oncology Nursing Forum. 2008;35(1):E1–E11
   • View In Article
8. Bender CM, Ergyn FS, Rosenzweig MQ, Cohen SM, Sereika SM. Symptom clusters in breast cancer across 3 phases of the disease. Cancer Nursing. 2005;28(3):219–225
   • View In Article
9. Breen SJ, Baravelli CM, Schofield PE, Jefford M, Yates PM, Aranda SK. Is symptom burden a predictor of anxiety and depression in patients with cancer about to commence chemotherapy?. The Medical Journal of Australia. 2009;190(7 Suppl.):S99–104
   • View In Article
10. Bruner DW. Should patient-reported outcomes be mandatory for toxicity reporting in cancer clinical trials?. Journal of Clinical Oncology. 2007;25(34):5345–5347
    • View In Article
11. Capp A, Inostroza-Ponta M, Bill D, Moscato P, Lai C, Christie D, et al. Is there more than one proctitis syndrome? A revisitation using data from the TROG 96.01 trial. Radiotherapy and Oncology. 2009;90(3):400–407
    • View In Article
12. Chan CW, Richardson A, Richardson J. A study to assess the existence of the symptom cluster of breathlessness, fatigue and anxiety in patients with advanced lung cancer. European Journal of Oncology Nursing. 2005;9(4):325–333
    • View In Article
13. Chen ML, Lin CC. Cancer symptom clusters: a validation study. Journal of Pain and Symptom Management. 2007;34(6):590–599
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (143 KB)
    • CrossRef
14. Chen ML, Tseng HC. Symptom clusters in cancer patients. Supportive Care in Cancer. 2006;14(8):825–830
    • View In Article
15. Cheung WY, Le LW, Zimmermann C. Symptom clusters in patients with advanced cancers. Support Care Cancer. 2009;17(9):1223–1230
    • View In Article
    • CrossRef
16. Chow E, Fan G, Hadi S, Filipczak L. Symptom clusters in cancer patients with bone metastases. Supportive Care in Cancer. 2007;15(9):1035–1043
    • View In Article
17. Cleeland CS, Bennett GJ, Dantzer R, Dougherty PM, Dunn AJ, Meyers CA, et al. Are the symptoms of cancer and cancer treatment due to a shared biologic mechanism? A cytokine-immunologic model of cancer symptoms. Cancer. 2003;97(11):2919–2925
    • View In Article
    • CrossRef
18. Cozby PC. Methods in Behavioral Research. sixth ed.. Mountain View: Mayfield Publishing Company; 1997;
    • View In Article
19. Delaunoit T, Goldberg RM, Sargent DJ, Morton RF, Fuchs CS, Findlay BP, et al. Mortality associated with daily bolus 5-fluorouracil/leucovorin administered in combination with either irinotecan or oxaliplatin: results from Intergroup Trial N9741. Cancer. 2004;101(10):2170–2176
    • View In Article
    • CrossRef
20. Dodd MJ, Cho MH, Cooper BA, Miaskowski C. The effect of symptom clusters on functional status and quality of life in women with breast cancer. European Journal of Oncology Nursing. 2010;14(2):101–110
    • View In Article
21. Dodd MJ, Miaskowski C, Paul SM. Symptom clusters and their effect on the functional status of patients with cancer. Oncology Nursing Forum. 2001;28(3):465–470
    • View In Article
22. Dodd MJ, Janson S, Facione N, Faucett J, Froelicher ES, Humphreys J, et al. Advancing the science of symptom management. Journal of Advanced Nursing. 2001;33(5):668–676
    • View In Article
    • MEDLINE
    • CrossRef
23. Dodd MJ, Miaskowski C, Lee KA. Occurrence of symptom clusters. Journal of the National Cancer Institute Monographs. 2004;32:76–78
    • View In Article
24. Ferreira KA, Kimura M, Teixeira MJ, Mendoza TR, da Nobrega JC, Graziani SR, et al. Impact of cancer-related symptom synergisms on health-related quality of life and performance status. Journal of Pain and Symptom Management. 2008;35(6):604–616
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (246 KB)
    • CrossRef
25. Finnegan L, Campbell RT, Ferrans CE, Wilbur J, Wilkie DJ, Shaver J. Symptom cluster experience profiles in adult survivors of childhood cancers. Journal of Pain and Symptom Management. 2009;38(2):258–269
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (305 KB)
    • CrossRef
26. Fox SW, Lyon D. Symptom clusters and quality of life in survivors of ovarian cancer. Cancer Nursing. 2007;30(5):354–361
    • View In Article
    • CrossRef
27. Fox SW, Lyon D, Farace E. Symptom clusters in patients with high-grade glioma. Journal of Nursing Scholarship. 2007;39(1):61–67
    • View In Article
28. Fox SW, Lyon DE. Symptom clusters and quality of life in survivors of lung cancer. Oncology Nursing Forum. 2006;33(5):931–936
    • View In Article
29. Francoeur RB. The relationship of cancer symptom clusters to depressive affect in the initial phase of palliative radiation. Journal of Pain and Symptom Management. 2005;29(2):130–155
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (284 KB)
    • CrossRef
30. Gaston-Johansson F, Fall-Dickson JM, Bakos AB, Kennedy MJ. Fatigue, pain, and depression in pre-autotransplant breast cancer patients. Cancer Practice. 1999;7(5):240–247
    • View In Article
    • MEDLINE
    • CrossRef
31. Gift AG, Jablonski A, Stommel M, Given CW. Symptom clusters in elderly patients with lung cancer. Oncology Nursing Forum. 2004;31(2):202–212
    • View In Article
32. Gift AG, Stommel M, Jablonski A, Given W. A cluster of symptoms over time in patients with lung cancer. Nursing Research. 2003;52(6):393–400
    • View In Article
    • MEDLINE
    • CrossRef
33. Given B, Given C, Azzouz F, Stommel M. Physical functioning of elderly cancer patients prior to diagnosis and following initial treatment. Nursing Research. 2001;50(4):222–232
    • View In Article
    • MEDLINE
    • CrossRef
34. Given C, Given B, Azzouz F, Kozachik S, Stommel M. Predictors of pain and fatigue in the year following diagnosis among elderly cancer patients. Journal of Pain and Symptom Management. 2001;21(6):456–466
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (82 KB)
    • CrossRef
35. Glaus A, Boehme C, Thurlimann B, Ruhstaller T, Hsu Schmitz SF, Morant R, et al. Fatigue and menopausal symptoms in women with breast cancer undergoing hormonal cancer treatment. Annals of Oncology. 2006;17(5):801–806
    • View In Article
36. Gleason JF, Case D, Rapp SR, Ip E, Naughton M, Butler JM, et al. Symptom clusters in patients with newly-diagnosed brain tumors. Journal of Supportive Oncology. 2007;5(9):427–433436
    • View In Article
37. Gwede CK, Small BJ, Munster PN, Andrykowski MA, Jacobsen PB. Exploring the differential experience of breast cancer treatment-related symptoms: a cluster analytic approach. Supportive Care in Cancer. 2008;16(8):925–933
    • View In Article
38. Hadi S, Fan G, Hird AE, Kirou-Mauro A, Filipczak LA, Chow E. Symptom clusters in patients with cancer with metastatic bone pain. Journal of Palliative Medicine. 2008;11(4):591–600
    • View In Article
    • CrossRef
39. Hart BL. Behavior of sick animals. Veterinary Clinics of North America: Food Animal Practice. 1987;3(2):383–391
    • View In Article
40. Hart BL. The behavior of sick animals. Veterinary Clinics of North America: Food Animal Practice. 1991;21(2):225–237
    • View In Article
41. Hockenberry M, Hooke MC. Symptom clusters in children with cancer. Seminars in Oncology Nursing. 2007;23(2):152–157
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (121 KB)
    • CrossRef
42. Hoffman AJ, Given BA, von Eye A, Gift AG, Given CW. Relationships among pain, fatigue, insomnia, and gender in persons with lung cancer. Oncology Nursing Forum. 2007;34(4):785–792
    • View In Article
43. Humphreys J, Lee KA, Carrieri-Kohlman V, Puntillo K, Faucett J, Janson S, et al. Theory of symptom management. In:  Smith M,  Liehr P editor. Middle Range Theory for Nursing. second ed.. New York: Springer Publishing Company; 2008;p. 145–158
    • View In Article
44. Jarden M, Nelausen K, Hovgaard D, Boesen E, Adamsen L. The effect of a multimodal intervention on treatment-related symptoms in patients undergoing hematopoietic stem cell transplantation: a randomized controlled trial. Journal of Pain and Symptom Management. 2009;38(2):174–190
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (297 KB)
    • CrossRef
45. Kenefick AL. Patterns of symptom distress in older women after surgical treatment for breast cancer. Oncology Nursing Forum. 2006;33(2):327–335
    • View In Article
46. Kenne Sarenmalm E, Ohlen J, Jonsson T, Gaston-Johansson F. Coping with recurrent breast cancer: predictors of distressing symptoms and health-related quality of life. Journal of Pain and Symptom Management. 2007;34(1):24–39
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (173 KB)
    • CrossRef
47. Kenne Sarenmalm E, Ohlen J, Oden A, Gaston-Johansson F. Experience and predictors of symptoms, distress and health-related quality of life over time in postmenopausal women with recurrent breast cancer. Psycho-oncology. 2008;17(5):497–505
    • View In Article
    • CrossRef
48. Kim E, Jahan T, Aouizerat BE, Dodd MJ, Cooper BA, Paul SM, et al. Changes in symptom clusters in patients undergoing radiation therapy. Supportive Care in Cancer. 2009;17(11):1383–1391
    • View In Article
49. Kim E, Jahan T, Aouizerat BE, Dodd MJ, Cooper BA, Paul SM, et al. Differences in symptom clusters identified using occurrence rates versus symptom severity ratings in patients at the end of radiation therapy. Cancer Nursing. 2009;32(6):429–436
    • View In Article
    • CrossRef
50. Kim H, Barsevick AM, Tulman L, McDermott PA. Treatment-related symptom clusters in breast cancer: a secondary analysis. Journal of Pain and Symptom Management. 2008;36(5):468–479
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (158 KB)
    • CrossRef
51. Kim H, McGuire DB, Tulman L, Barsevick AM. Symptom clusters: concept analysis and clinical implications for cancer nursing. Cancer Nursing. 2005;28(4):270–282quiz 283–274
    • View In Article
52. Kuo T, Ma F. Symptom distresses and coping strategies in patients with non-small cell lung cancer. Cancer Nursing. 2002;25(4):309–317
    • View In Article
    • CrossRef
53. Kwekkeboom KL, Wanta B, Bumpus M. Individual difference variables and the effects of progressive muscle relaxation and analgesic imagery interventions on cancer pain. Journal of Pain and Symptom Management. 2008;36(6):604–615
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (217 KB)
    • CrossRef
54. Lenz ER, Pugh LC, Milligan RA, Gift A, Suppe F. The middle-range theory of unpleasant symptoms: an update. ANS: Advance in Nursing Science. 1997;19(3):14–27
    • View In Article
55. Liu L, Fiorentino L, Natarajan L, Parker BA, Mills PJ, Sadler GR, et al. Pre-treatment symptom cluster in breast cancer patients is associated with worse sleep, fatigue and depression during chemotherapy. Psycho-oncology. 2009;18(2):187–194
    • View In Article
    • CrossRef
56. Maliski SL, Kwan L, Elashoff D, Litwin MS. Symptom clusters related to treatment for prostate cancer. Oncology Nursing Forum. 2008;35(5):786–793
    • View In Article
57. Miaskowski C, Cooper BA, Paul SM, Dodd M, Lee K, Aouizerat BE, et al. Subgroups of patients with cancer with different symptom experiences and quality-of-life outcomes: a cluster analysis. Oncology Nursing Forum. 2006;33(5):E79–E89
    • View In Article
58. Miaskowski C, Dodd M, Lee K. Symptom clusters: the new frontier in symptom management research. Journal of National Cancer Institute Monographs. 2004;32:17–21
    • View In Article
59. Miaskowski C, Lee KA. Pain, fatigue, and sleep disturbances in oncology outpatients receiving radiation therapy for bone metastasis: a pilot study. Journal of Pain and Symptom Management. 1999;17(5):320–332
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (229 KB)
    • CrossRef
60. Molassiotis A, Wengstrom Y, Kearney N. Symptom cluster patterns during the first year after diagnosis with cancer. Journal of Pain and Symptom Management. 2010;
    • View In Article
61. Olson K, Hayduk L, Cree M, Cui Y, Quan H, Hanson J, et al. The changing causal foundations of cancer-related symptom clustering during the final month of palliative care: a longitudinal study. BMC Medical Research Methodology. 2008;8:36
    • View In Article
    • CrossRef
62. Patrick DL, Ferketich SL, Frame PS, Harris JJ, Hendricks CB, Levin B, et al. National Institutes of Health State-of-the-Science Conference Statement: symptom management in cancer: pain, depression, and fatigue, July 15–17, 2002. Journal of National Cancer Institute Monographs. 2004;32:9–16
    • View In Article
63. Polit DF, Beck CT. Nursing Research: Principles and Methods. seventh ed.. Philadelphia: Lippincott Williams & Wilkins; 2004;
    • View In Article
64. Pud D, Ben Ami S, Cooper BA, Aouizerat BE, Cohen D, Radiano R, et al. The symptom experience of oncology outpatients has a different impact on quality-of-life outcomes. Journal of Pain and Symptom Management. 2008;35(2):162–170
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (388 KB)
    • CrossRef
65. Reyes-Gibby CC, Aday LA, Anderson KO, Mendoza TR, Cleeland CS. Pain, depression, and fatigue in community-dwelling adults with and without a history of cancer. Journal of Pain and Symptom Management. 2006;32(2):118–128
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (134 KB)
    • CrossRef
66. Reyes-Gibby CC, Chan W, Abbruzzese JL, Xiong HQ, Ho L, Evans DB, et al. Patterns of self-reported symptoms in pancreatic cancer patients receiving chemoradiation. Journal of Pain and Symptom Management. 2007;34(3):244–252
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (137 KB)
    • CrossRef
67. Ridner SH. Quality of life and a symptom cluster associated with breast cancer treatment-related lymphedema. Supportive Care in Cancer. 2005;13(11):904–911
    • View In Article
68. Rutledge DM, McGuire C. Evidence-based symptom management. In:  Yarbro CH,  Frogge MH,  Goodan M editor. Cancer Symptom Management. third ed.. Sudbury: Jones & Bartlett; 2004;p. 3–14
    • View In Article
69. So WK, Marsh G, Ling WM, Leung FY, Lo JC, Yeung M, et al. The symptom cluster of fatigue, pain, anxiety, and depression and the effect on the quality of life of women receiving treatment for breast cancer: a multicenter study. Oncology Nursing Forum. 2009;36(4):E205–E214
    • View In Article
70. Suwisith N, Hanucharunkul S, Dodd M, Vorapongsathorn T, Pongthavorakamol K, Asavametha N. Symptom clusters and functional status of women with breast cancer. Thai Journal of Nursing Research. 2008;12(3):153–164
    • View In Article
71. Walke LM, Byers AL, Gallo WT, Endrass J, Fried TR. The association of symptoms with health outcomes in chronically ill adults. Journal of Pain and Symptom Management. 2007;33(1):58–66
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (137 KB)
    • CrossRef
72. Walsh D, Rybicki L. Symptom clustering in advanced cancer. Support Care Cancer. 2006;14(8):831–836
    • View In Article
    • MEDLINE
    • CrossRef
73. Wang S, Tsai CM, Chen BC, Lin CH, Lin CC. Symptom clusters and relationships to symptom interference with daily life in Taiwanese lung cancer patients. Journal of Pain and Symptom Management. 2008;35(3):258–266
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (181 KB)
    • CrossRef
74. Wang XS, Fairclough DL, Liao Z, Komaki R, Chang JY, Mobley GM, et al. Longitudinal study of the relationship between chemoradiation therapy for non-small-cell lung cancer and patient symptoms. Journal of Clinical Oncology. 2006;24(27):4485–4491
    • View In Article
75. Watkins LR, Maier SF. The pain of being sick: implications of immune-to-brain communication for understanding pain. Annual Review of Psychology. 2000;51:29–57
    • View In Article
    • MEDLINE
    • CrossRef
76. Wilmoth MC, Coleman EA, Wahab HT. Initial validation of the symptom cluster of fatigue, weight gain, psychologic distress and altered sexuality. Southern Online Journal of Nursing Research. 2009;9(3):7
    • View In Article
77. Yamagishi A, Morita T, Miyashita M, Kimura F. Symptom prevalence and longitudinal follow-up in cancer outpatients receiving chemotherapy. Journal of Pain Symptom Management. 2009;37(5):823–830
    • View In Article
78. Yeh C, Chiang Y, Chien L, Lin L, Yang C, Chuang H. Symptom clustering in older Taiwanese children with cancer. Oncology Nursing Forum. 2008;35(2):273–281
    • View In Article
79. Yirmiya R. Endotoxin produces a depressive-like episode in rats. Brain Research. 1996;711(1–2):163–174
    • View In Article
    • MEDLINE
    • CrossRef