Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2024 Jan 17.
Published in final edited form as: Support Care Cancer. 2022 May 11;30(8):6889–6899. doi: 10.1007/s00520-022-07125-z

Symptom Clusters in Outpatients with Cancer Using Different Dimensions of the Symptom Experience

Carolyn S Harris 1, Kord M Kober 1, Bruce Cooper 1, Yvette P Conley 2, Anand A Dhruva 3, Marilyn J Hammer 4, Steven Paul 1, Jon D Levine 3, Christine A Miaskowski 1,3
PMCID: PMC10792845  NIHMSID: NIHMS1956804  PMID: 35543816

Abstract

Purpose:

Relatively few studies have evaluated for symptom clusters across multiple dimensions. It is unknown whether the symptom dimension used to create symptom clusters influences the number and types of clusters that are identified. Study purposes were to describe ratings of occurrence, severity, and distress for 38 symptoms in a heterogeneous sample of oncology patients (n=1329) undergoing chemotherapy; identify and compare the number and types of symptom clusters based on three dimensions (i.e., occurrence, severity, and distress); and identify common and distinct clusters.

Methods:

A modified version of the Memorial Symptom Assessment Scale was used to assess the occurrence, severity, and distress ratings of 38 symptoms in the week prior to patients’ next cycle of chemotherapy. Symptom clusters for each dimension were identified using exploratory factor analysis.

Results:

Patients reported an average of 13.9 (±7.2) concurrent symptoms. Lack of energy was both the most common and severe symptom while “I don’t look like myself” was the most distressing. Psychological, gastrointestinal, weight gain, respiratory, and hormonal clusters were identified across all three dimensions. Findings suggest that psychological, gastrointestinal, and weight gain clusters are common while respiratory and hormonal clusters are distinct.

Conclusions:

Psychological, gastrointestinal, weight gain, hormonal, and respiratory clusters are stable across occurrence, severity, and distress in oncology patients receiving chemotherapy. Given the stability of these clusters and the consistency of the symptoms across dimensions, use of a single dimension to identify these clusters may be sufficient. However, comprehensive and disease-specific inventories need to be used to identify distinct clusters.

Keywords: cancer, chemotherapy, symptoms, symptom clusters

INTRODUCTION

Patients receiving chemotherapy report between 10 [1] to 14.5 [2] concurrent symptoms. While these data fostered symptom clusters’ research [3, 4], progress in this area of scientific inquiry is limited by multiple unanswered questions [5–7]. One question is whether the symptom dimension (i.e., occurrence, severity, distress) impacts the number and types of symptom clusters that are identified. As highlighted in one systematic review of symptom clusters in patients receiving adjuvant chemotherapy [7], less than half of the 23 studies evaluated for symptom clusters across two or more symptom dimensions. A second question that warrants investigation is the determination of which clusters are common and distinct across various types of cancer [5]. The answers to these questions will guide clinical assessments and inform mechanistic-based studies.

Nine cross-sectional studies evaluated for symptom clusters in heterogeneous samples receiving chemotherapy [8–16]. Six studies used a single symptom dimension to identify the clusters [8–10, 12, 15, 16], two used two or more dimensions [11, 13], and one did not report the dimension used in the analysis [14]. Across these nine studies, the number of clusters varied from three to eight. While a psychological cluster was the only common one across seven of these studies [8–10, 12, 13, 15, 16], none of them contained the same symptoms. This variability in both the types of clusters and symptoms within the clusters is related to heterogeneity in the symptom inventories used; number of symptoms evaluated; timing of the assessments; and statistical methods used. Because of these differences, one cannot determine if the number and types of symptom clusters vary based on the dimensions used to create the clusters. In addition, these data suggest that the only common cluster, in samples with heterogeneous types of cancer, is a psychological one.

While we previously evaluated for symptom clusters across two or more symptom dimensions in patients with breast [17], gastrointestinal [18], gynecological [19], or lung [20] cancer using exploratory factor analysis (EFA), we have not used EFA to evaluate for symptom clusters in the entire sample. In addition, we recently reported on the results of a network analysis (NA) of symptom clusters in the combined sample [13]. A comparison of the number and types of symptom clusters that were identified for each type of cancer diagnosis to those that are identified for the combined sample, as well as a comparison of findings using different analytic approaches [5], will allow for the generation of hypotheses related to common and unique symptom clusters in oncology patients.

Therefore, the purposes of this study were to describe ratings of occurrence, severity, and distress for 38 symptoms in a heterogeneous sample of oncology patients undergoing chemotherapy and identify and compare the number and types of symptom clusters based on three symptom dimensions (i.e., occurrence, severity, and distress). In addition, an evaluation of common and distinct symptom clusters was done for the total sample compared to four distinct types of cancer (i.e., breast [17], gastrointestinal [18], gynecological [19], lung [20]) and for two different methods (i.e., EFA, NA[13]).

METHODS

Patients and Settings

This analysis is part of a larger study that evaluated symptom clusters in oncology outpatients receiving chemotherapy [13, 17–20]. Eligible patients were ≥18 years of age; had a diagnosis of breast, lung, gastrointestinal, or gynecologic cancer; had received chemotherapy within the preceding four weeks; were scheduled to receive at least two additional cycles of chemotherapy; were able to read, write, and understand English; and gave written informed consent. Patients were recruited from two Comprehensive Cancer Centers, one Veteran’s Affairs hospital, and four community-based oncology programs. Of the 1343 patients enrolled, 1329 patients had complete Memorial Symptom Assessment Scale (MSAS) data.

Procedures

Eligible patients were approached during their first or second cycle of chemotherapy and provided written informed consent. Patients completed questionnaires in their home and returned them in a postage paid envelope, six times over two cycles of chemotherapy. Data from the enrollment assessment (symptoms in the week before the patient’s second or third cycle of chemotherapy) were used in these analyses. Medical records were reviewed for disease and treatment information. This study was approved by the Committee on Human Research at the University of California, San Francisco.

Instruments

Patients completed a demographic questionnaire, Karnofsky Performance Status (KPS) scale [21], and Self-Administered Comorbidity Questionnaire [22]. Toxicity of each patient’s chemotherapy regimen was rated using the MAX2 index [23, 24].

A modified version of the 32-item MSAS was used to evaluate the occurrence, severity, and distress of 38 common symptoms associated with cancer and its treatment [25]. Six common symptoms were added: hot flashes, chest tightness, difficulty breathing, abdominal cramps, increased appetite, and weight gain. Using the MSAS, patients reported whether they had experienced each symptom in the past week. If they had experienced the symptom, they were asked to rate its severity and distress. Severity was measured using a four-point Likert scale (i.e., 1 = slight, 2 = moderate, 3 = severe, 4 = very severe). Distress was measured using a five-point Likert scale (i.e., 0 = not at all, 1 = a little bit, 2 = somewhat, 3 = quite a bit, 4 = very much). The validity and reliability of the MSAS are well established [25].

Data Analysis

Descriptive statistics and frequency distributions were calculated for the demographic and clinical characteristics, as well as symptom occurrence rates and severity and distress ratings using the Statistical Package for the Social Sciences Version 27 (IBM Corporation, Armonk, NY). EFA was used to identify symptom clusters using Mplus Version 8.6 [26].

For the EFA, factor loadings were considered meaningful if the loading was ≥0.40 [26]. In addition, factors were considered to be adequately defined if at least two items (i.e., symptoms) had loadings of ≥0.40 [27]. Items were allowed to load on two factors (i.e., cross-load) if they fell within our preset criteria of ≥0.40. For the EFA of the occurrence items, tetrachoric correlations were used to create the matrix of associations [26]. For the EFAs of the severity and distress ratings, polychoric correlations were used to create the matrix of associations. The simple structure for the occurrence, severity, and distress EFAs were estimated using the method of unweighted least squares with geomin (i.e., oblique) rotation. The unweighted least squares estimator was selected to achieve more reliable results with the dichotomous (i.e., occurrence) and ordinal (i.e., severity, distress) items [26].

The EFA for severity was done using severity ratings that included a zero (i.e., 0, 1, 2, 3, 4). If the patient indicated that they did not have the symptom, a severity score of zero was assigned. The EFA for distress was done using distress ratings that included a zero (did not have the symptom) and the original ratings shifted from 1 (not at all) to 5 (very much). The initial EFA analyses were done using severity and distress ratings that did not include zero (i.e., 1, 2, 3, 4, 5). However, the pairwise missingness (i.e., 1-covariance coverage for each of the item pairs) was over 90% and the estimation failed to converge.

Factor solutions were estimated for two through five factors. The factor solution with the greatest interpretability and clinical meaningfulness was selected given that it met the criteria set for evaluating simple structure (i.e., size of item loadings, number of items on a factor). Then, each factor solution was examined to determine a clinically appropriate name for the symptom cluster. Clusters were named based on the symptoms with the highest factor loadings and the majority of the symptoms within the cluster.

Differences in Number and Types of Clusters

To evaluate percent agreement among the symptoms within the same cluster using occurrence, severity, and distress ratings, previous studies by our group [17–20, 28–31] and others [32, 33] used the criteria proposed by Kirkova and Walsh [34]. They suggested that to be in agreement with each other, at least 75% of the symptoms in the cluster should be present including the prominent and most important symptom (i.e., symptom with the largest factor loading).

While Kirkova and Walsh [34] used the term “stability” to describe these criteria, the definition and use of stability within symptom cluster research is inconsistent [7] and has led to the subjective application of these criteria. Therefore, in this study, the term stability is used to describe whether or not the same clusters are identified across dimensions and/or studies. In contrast, consistency is used to describe whether the specific symptoms within a cluster remain the same across symptom dimensions (i.e., percent agreement among the symptoms within the cluster).

RESULTS

Demographic and Clinical Characteristics

Of the 1329 patients in this study, 77.8% were female, 69.9% were White, 64.4% were married or partnered, and had a mean age of 57.3 (±12.3) years (Table 1). While the majority (60.4%) reported a mean household annual income of ≥$70,000, only 35.1% were currently employed. Most patients were well-educated (16.2 ±3.0 years), exercised on a regular basis (70.9%), and had never smoked (64.7%). Patients had 2.4 (±1.4) comorbid conditions and an average KPS score of 80.1 (±12.4). On average, patients reported 13.9 (±7.2) concurrent symptoms before their second or third cycle of chemotherapy.

Table 1.

Demographic and Clinical Characteristics of the Patients (n=1329)

Characteristic Mean SD
Age (years) 57.3 12.3
Education (years) 16.2 3
Body mass index (kilograms/meters squared) 26.2 5.7
Karnofsky Performance Status score 80.1 12.4
Number of comorbidities out of 13 2.4 1.4
Self-administered Comorbidity Questionnaire score 5.5 3.2
Time since cancer diagnosis (years) 2 3.9
Time since diagnosis (median) 0.42
Number of prior cancer treatments (out of 9) 1.6 1.5
Number of metastatic sites including lymph node involvement (out of 9) 1.2 1.2
Number of metastatic sites excluding lymph node involvement (out of 8) 0.8 1
MAX2 Index of Chemotherapy Toxicity score (0 to 1) 0.17 0.08
Mean number of MSAS symptoms (out of 38) 13.9 7.2
n (%)
Gender
 Female 1033 77.8
 Male 295 22.2
Ethnicity
 White 917 69.9
 Black 95 7.2
 Asian or Pacific Islander 161 12.3
 Hispanic, Mixed, or Other 139 10.6
Married or partnered (% yes) 843 64.4
Lives alone (% yes) 283 21.6
Child care responsibilities (% yes) 286 22
Care of adult responsibilities (% yes) 95 7.9
Currently employed (% yes) 462 35.1
Income
 < $30,000 219 18.4
 $30,000 to < $70,000 252 21.2
 $70,000 to < $100,000 199 16.7
 ≥ $100,000 520 43.7
Exercise on a regular basis (% yes) 922 70.9
Current or history of smoking (% yes) 462 35.3
Type of cancer
 Breast 534 40.2
 Gastrointestinal 407 30.6
 Gynecological 233 17.5
 Lung 155 11.7
Type of prior cancer treatment
 No prior treatment 323 25
 Only CTX, surgery, or RT 543 42
 CTX and surgery, or CTX and RT, or surgery and RT 257 19.9
 CTX and surgery and RT 169 13.1
Cycle length 558 42.1
 14 days 671 50.6
 21 days 97 7.3
 28 days
Emetogenicity of the chemotherapy regimen
 Minimal/low 259 19.5
 Moderate 810 61
 High 258 19.4
Antiemetic regimen
 None 92 7.1
 Steroid alone or serotonin receptor antagonist alone 265 20.4
 Serotonin receptor antagonist and steroid 618 47.7
 NK-1 receptor antagonist and two other antiemetics 321 24.8
Open in a new tab

Symptom Prevalence

Lack of energy was the most common symptom (Table 2). Mean severity ratings were calculated in two ways (i.e., with and without zeros). When zeros were included in the calculation, lack of energy was the most severe symptom. In the “without zeros” analyses, hair loss was rated as the most severe symptom. “I don’t look like myself” was the most distressing symptom.

Table 2.

Occurrence Rates and Severity and Distress Ratings for Symptoms Prior to Chemotherapy

Symptomsa Occurrence Ratesb Severity Ratings with Zerosc Severity Ratings without Zerosd Distress Ratingse
n % Mean SD Mean SD Mean SD
Lack of energy 1106 83.2 1.67 1.01 2.02 0.72 1.79 1.14
Difficulty sleeping 918 69.1 1.38 1.13 2.01 0.76 1.79 1.11
Pain 803 60.4 1.14 1.10 1.92 0.73 1.77 1.10
Feeling drowsy 801 60.3 1.04 1.01 1.75 0.70 1.16 1.05
Hair loss 728 54.8 1.35 1.49 2.49 1.12 1.88 1.34
Numbness/tingling in hands/feet 694 52.2 0.94 1.09 1.84 0.81 1.52 1.18
Worrying 692 52.1 0.94 1.06 1.85 0.74 1.63 1.04
Difficulty concentrating 690 51.9 0.79 0.90 1.55 0.64 1.48 1.07
Change in the way food tastes 656 49.4 1.04 1.23 2.12 0.89 1.72 1.26
Nausea 631 47.5 0.82 1.04 1.76 0.81 1.65 1.12
Feeling sad 612 46.0 0.77 0.97 1.71 0.71 1.50 1.06
Dry mouth 603 45.4 0.77 1.00 1.73 0.75 1.23 1.12
Constipation 578 43.5 0.84 1.12 1.98 0.83 1.70 1.17
Feeling irritable 549 41.3 0.69 0.95 1.70 0.72 1.46 1.03
Lack of appetite 549 41.3 0.78 1.07 1.92 0.79 1.28 1.11
Feeling nervous 505 38.0 0.59 0.88 1.62 0.68 1.41 0.98
“I don’t look like myself” 503 37.8 0.80 1.18 2.15 0.93 1.98 1.22
Changes in skin 482 36.3 0.68 1.03 1.91 0.81 1.64 1.19
Feeling bloated 440 33.1 0.58 0.93 1.79 0.73 1.54 1.07
Cough 433 32.6 0.45 0.75 1.42 0.62 1.02 1.08
Hot flashes 423 31.8 0.58 0.98 1.87 0.81 1.42 1.16
Dizziness 416 31.3 0.46 0.79 1.51 0.69 1.24 0.98
Sweats 415 31.2 0.53 0.92 1.77 0.78 1.29 1.09
Problems with sexual interest or activity 397 29.9 0.71 1.24 2.47 0.98 1.87 1.28
Diarrhea 393 29.6 0.54 0.95 1.87 0.81 1.46 1.13
Shortness of breath 357 26.9 0.44 0.82 1.67 0.71 1.51 1.04
Increased appetite 344 25.9 0.44 0.83 1.75 0.68 0.91 1.11
Weight gain 337 25.4 0.39 0.76 1.58 0.70 1.37 1.33
Weight loss 335 25.2 0.38 0.76 1.56 0.71 0.96 1.17
Itching 330 24.8 0.41 0.82 1.71 0.74 1.28 1.07
Abdominal cramps 299 22.5 0.40 0.84 1.87 0.75 1.61 1.08
Mouth sores 278 20.9 0.34 0.76 1.70 0.74 1.46 1.06
Difficulty breathing 265 19.9 0.32 0.72 1.64 0.72 1.63 1.13
Chest tightness 237 17.8 0.27 0.64 1.54 0.67 1.42 1.00
Swelling of arms or legs 194 14.6 0.27 0.74 1.91 0.83 1.62 1.16
Problems with urination 187 14.1 0.24 0.68 1.79 0.80 1.51 1.21
Difficulty swallowing 183 13.8 0.23 0.66 1.73 0.82 1.64 1.15
Vomiting 164 12.3 0.21 0.66 1.80 0.90 1.74 1.18
Open in a new tab

Occurrence Clusters

Five-factor solution was selected for the occurrence EFA (Table 3). Psychological cluster had six symptoms and worrying had the highest factor loading. Gastrointestinal cluster had 11 symptoms and lack of appetite had the highest factor loading. Weight gain cluster had two symptoms and weight gain had the highest factor loading. Hormonal cluster had two symptoms and hot flashes had the highest factor loading. Respiratory cluster had four symptoms and difficulty breathing had the highest factor loading.

Table 3.

Comparison of Symptom Clusters Prior to Initiation of Chemotherapy Using Ratings of Occurrence, Severity, and Distressa

Cluster Symptoms Occurrence Severity Distress
Psychological symptom cluster Worrying 0.864 0.866 0.875
Feeling sad 0.855 0.850 0.872
Feeling nervous 0.744 0.750 0.760
Feeling irritable 0.626 0.569 0.574
Difficulty concentrating 0.549 0.517 0.560
“I don’t look like myself” 0.458 − 0.427
Total number of symptoms in this cluster 6/6 5/6 6/6
Gastrointestinal symptom cluster Lack of appetite 0.784 0.774 0.770
Weight loss 0.679 0.658 0.680
Nausea 0.663 0.624 0.612
Change in the way food tastes 0.612 0.690 0.677
Vomiting 0.546 0.538 0.525
Difficulty swallowing 0.513 0.517 0.503
Abdominal cramps 0.455 0.472 0.444
Diarrhea 0.433 0.483 0.455
Dry mouth 0.431 0.472 0.474
Constipation 0.430 − −
Dizziness 0.404 − −
Mouth sores − 0.420 −
Total number of symptoms in this cluster 11/12 10/12 9/12
Weight gain symptom cluster Weight gain 0.921 0.875 0.914
Increased appetite 0.785 0.746 0.736
Total number of symptoms in this cluster 2/2 2/2 2/2
Hormonal symptom cluster Hot flashes 0.883 0.907 0.920
Sweats 0.670 0.728 0.647
Total number of symptoms in this cluster 2/2 2/2 2/2
Respiratory symptom cluster Difficulty breathing 1.037 1.032 1.035
Shortness of breath 0.716 0.763 0.741
Chest tightness 0.689 0.614 0.628
Cough 0.457 0.430 0.427
Total number of symptoms in this cluster 4/4 4/4 4/4
Open in a new tab

Severity Clusters

Five-factor solution was selected for the severity EFA (Table 3). Psychological cluster had five symptoms and worrying had the highest factor loading. Gastrointestinal cluster had 10 symptoms and lack of appetite had the highest factor loading. Weight gain cluster had two symptoms and weight gain had the highest factor loading. Hormonal cluster had two symptoms and hot flashes had the highest factor loading. Respiratory cluster had four symptoms and difficulty breathing had the highest factor loading.

Distress Clusters

Five-factor solution was selected for the distress EFA (Table 3). Psychological cluster had six symptoms and worrying had the highest factor loading. Gastrointestinal cluster had nine symptoms and lack of appetite had the highest factor loading. Weight gain cluster had two symptoms and weight gain had the highest factor loading. Hormonal cluster had two symptoms and hot flashes had the highest factor loading. Respiratory cluster had four symptoms and difficulty breathing had the highest factor loading.

Stability and Consistency

Five stable clusters were identified across all three symptom dimensions (Table 3). Across all five clusters, the symptom with the highest factor loading was the same across all three dimensions. In terms of consistency, for psychological cluster, consistency ranged from 83.3% (severity) to 100% (occurrence, distress). For gastrointestinal cluster, consistency ranged from 75.0% (distress) to 91.7% (occurrence). For weight gain, hormonal, and respiratory clusters, consistency was 100% across the three dimensions.

DISCUSSION

Findings from this study provide new information on the occurrence, severity, and distress of 38 symptoms in a large, heterogeneous sample of oncology patients. In the week prior to their second or third cycle of chemotherapy, patients reported on average 13.9 symptoms. Consistent with previous studies of patients receiving chemotherapy, lack of energy was the most common and severe symptom [8, 9, 15]. However, as noted previously [18, 19, 35], the most common symptoms are not always the most distressing. Hair loss was rated as the most severe symptom when zeros were not included in the mean severity scores, while “I don’t look like myself” was the most distressing. Based on these findings, to have a more complete picture of the impact of individual symptoms, multiple dimensions of the symptom experience warrant evaluation.

Using findings from the literature, as well as our previous EFAs for breast [17], gastrointestinal [18], gynecological [19], and lung [20] cancers, and our NA for the entire sample [13], the remainder of this discussion describes the common and distinct symptom clusters (Table 4).

Table 4.

Comparison of Symptom Clusters Across Cancer Types and Analytic Methods Using Ratings of Occurrence, Severity, and Distress

Symptom dimension Symptom cluster EFA n=1329 NAa n=1328 Breastb n=534 GIc n=399 GYNd n=232 Lunge n=145
Occurrence Psychological X X X X X X
GI X X X X
Epithelial/GI X
Epithelial X
Nutritional X X
Weight change X X X
Weight gain X
Hormonal X X X X
Respiratory X X X
Lung CA-specific X
CTX related X X
Sickness behavior X X
Pain and abdominal X
Severity Psychological X X X X X X
GI X X X
GI/epithelial X
Epithelial/GI X
Epithelial X
Nutritional X X
Weight change X X X
Weight gain X
Hormonal X X X X
Respiratory X X X
Lung CA-specific X
CTX related X X
Sickness behavior X
Distress Psychological X X Not assessed X Not assessed
Psychological/GI X
GI X X X
GI/epithelial X
Epithelial X
Nutritional X
Weight change X X
Weight gain X
Hormonal X X X
Respiratory X X X
CTX related X X
Open in a new tab

Psychological Cluster

Consistent with two reviews that reported that a psychological cluster was one of the most common clusters in patients receiving chemotherapy [6, 7], this cluster was identified across all three symptom dimensions. Therefore, it is not surprising that a psychological cluster was identified in our previous studies of four types of cancer [17–20] as well as in our NA [13]. In this cluster, the most consistent symptoms across dimensions, cancer types, and analytic methods were: worrying, feeling sad, feeling nervous, and feeling irritable. Taken together, these findings suggest that a psychological cluster is stable across various cancer types and can be identified using any symptom dimension. Given its stability, psychological symptoms need to be routinely assessed in all oncology patients.

Gastrointestinal Cluster

Across studies of patients receiving chemotherapy [6, 7], a gastrointestinal cluster was identified repeatedly using ratings of occurrence, severity, and distress. Given chemotherapy affects rapidly dividing cells, its impact on the gastrointestinal tract results in a constellation of symptoms [36]. While nausea, vomiting, and diarrhea are the most consistent symptoms within this cluster [6, 7], in the current study, lack of appetite, weight loss, nausea, change in the way food tastes, vomiting, difficulty swallowing, diarrhea, abdominal cramps, and dry mouth were consistent across the three dimensions.

When compared with our previous studies of individual types of cancer [17–20], as well as the NA of the total sample [13], the names of this cluster, as well as the specific symptoms were not consistent. For example, abdominal cramps was the only symptom that was consistent across these studies and dimensions. In addition, the “gastrointestinal” cluster identified in patients with gynecological or lung cancer included multiple symptoms related to the epithelium (e.g., changes in skin, itching). This variability has a number of plausible explanations, including: differential effects of specific chemotherapy regimens on the gastrointestinal mucosa; differential effects of the cancer itself (e.g., colon cancer versus breast cancer) on the gastrointestinal tract; differential perceptions of a specific symptom in terms of its severity versus its distress; and/or variations in the relationships among various symptoms that are associated with specific types of cancer (e.g., feeling bloated in gastrointestinal cancers). Despite these variations, given the identification of a gastrointestinal cluster across multiple independent samples [8, 9, 12, 32, 33, 37–39], this cluster can be considered stable. Additional research is warranted to determine the specific factors that contribute to subtle variations in the consistency of symptoms within this cluster.

Weight Gain Cluster

In the current study, a weight gain cluster was identified that included weight gain and increased appetite across all three symptom dimensions. However, across previous studies with heterogeneous cancer types [10, 13, 15, 16], as well as in our own studies with specific cancer diagnoses [17–20], this cluster was highly variable both in terms of stability and consistency. For example, in a study of patients with hematologic malignancies [10], lack of appetite, taste changes, and nausea were included in an appetite cluster. In another study of older cancer patients with a variety of solid tumors [15], lack of appetite, change in the way food tastes, constipation, weight loss, and “I do not look like myself” were identified as a nutrition cluster. In our work [13, 17–20], weight gain was the only consistent symptom across cancer types, analytic methods, and dimensions.

Variability, in both stability and consistency, across studies may be due to differences in the types of chemotherapy received, medications patients are taking, and/or the location of tumors in or near the digestive system. Another factor that may contribute to variability is the symptom assessment instrument that was used. In our [13, 17–20] and one of the aforementioned studies [10], modified versions of the MSAS were used that included multiple symptoms related to appetite and nutrition. Studies that use an instrument with fewer symptoms will not be able to identify a weight- or nutrition-related cluster. Given that changes in nutritional status can lead to a variety of comorbidities (e.g., diabetes) [40], comprehensive nutritional assessments are a vital component of cancer care.

Respiratory Cluster

Respiratory cluster, that included difficulty breathing, shortness of breath, chest tightness, and cough, was found across all three dimensions. In our previous studies, a respiratory cluster was identified in the total sample using NA [13] and in patients with gynecological [19] and lung [20] cancer across two or more dimensions; but not in patients with breast [17] or gastrointestinal [18] cancers. In addition, across two studies that evaluated for symptom clusters in a heterogeneous sample [15, 38], only one identified a respiratory cluster [38]. The inconsistent identification of this cluster suggests that it may be unique to certain cancer types. These differences may be related to tumor locations and/or conditions that are more common to specific diagnoses (e.g., ascites, pleural effusion).

Hormonal Cluster

Hormonal cluster was identified that included hot flashes and sweats across all three symptom dimensions. In another study that compared symptom clusters that were identified in younger (<60 years) and older (≥60 years) patients receiving chemotherapy [15], a hormonal cluster was identified in only the younger group. The identification of this cluster in younger patients supports the hypothesis that this cluster may emerge during/following cancer treatments that induce menopause [41, 42].

In addition, this cluster may be unique to specific cancer diagnoses. For example, a type of hormonal cluster (i.e., menopausal, vasomotor) was identified in women with breast [39] and ovarian [43] cancer. In addition, among our previous analyses [13, 17–20], a hormonal cluster was identified in the total sample using NA, and in women with breast [17] and gynecological [19] cancer across two or more symptom dimensions. Across all symptom dimensions within these three studies [13, 17, 19], hot flashes and sweats were consistent. Of note, studies that do not use disease-specific or comprehensive symptom inventories will not be able to identify this distinct cluster in patients with breast or gynecological cancers, and perhaps in men with prostate cancer.

Comparison with Network Analysis

Identification of psychological, gastrointestinal, weight gain or nutritional, hormonal, and respiratory clusters using EFA is consistent with our previous NA of the total sample [13]. For both analyses, the symptoms within the psychological, hormonal, and respiratory clusters were relatively consistent across all three symptom dimensions. While both studies identified a gastrointestinal cluster, this cluster was only identified using distress in the NA. While both analytic approaches use measures of correlation to identify clusters, they differ in key ways. In our previous NA [13], symptom clusters were identified using the Walktrap algorithm and all symptoms within the network were retained regardless of the strength of the relationship between and among symptoms. For the EFAs, because the symptoms needed to have a factor loading ≥0.40, 13 to 15 symptoms did not load on one or more clusters. The advantages and disadvantages of various analytic methods need to be explored in future studies with large samples.

A number of limitations warrant consideration. Because our previous studies of patients with breast [17] and lung [20] cancer used only two symptom dimensions (i.e., occurrence, severity) to identify symptom clusters, our evaluation of the stability and consistency of clusters using distress warrants additional research. Given the study’s cross-sectional design, additional research needs to determine which clusters remain stable across dimensions, cancer diagnoses, and/or time. Given that the occurrence and severity of symptoms may be influenced by specific chemotherapy drugs, additional research is warranted on the stability and consistency of symptom clusters across different chemotherapy regimens. While these findings suggest that respiratory and hormonal clusters are distinct clusters that occur with specific types of cancer, the proportions of patients with a gynecological (i.e., 17.5%) or lung (i.e., 11.7%) cancer were relatively small. In addition, our sample was primarily White and well-educated, which limits the generalizability of our findings.

CONCLUSION

Our findings suggest that psychological, gastrointestinal, weight gain, hormonal, and respiratory clusters are stable across occurrence, severity, and distress prior to the start of the next cycle of chemotherapy. Given the stability of these clusters across dimensions and the consistency of the symptoms within the clusters, they can be identified using any dimension of the symptom experience. However, for any single symptom, multiple dimensions of the symptom experience warrant evaluation to assess its full impact on a patient.

In addition, these findings suggest that gastrointestinal, psychological, and nutrition or weight change clusters are common across cancer types. Given the stability of these clusters across diagnoses, future research should explore whether these clusters share common biological mechanisms. Furthermore, additional research is needed to evaluate whether these clusters remain stable over time and across other cancer treatments (e.g., radiation therapy, surgery). Conversely, hormonal and respiratory clusters may be unique to specific cancer types. Symptoms within these distinct clusters need to be assessed in patients with breast, gynecological, or lung cancer in the clinical and research settings.

Funding

This work was supported by a grant from the National Cancer Institute (NCI, CA134900). Ms. Harris is supported by a grant from the American Cancer Society and the National Institute of Nursing Research of the National Institutes of Health (T32NR016920). The content is solely the responsibilities of the authors and does not necessarily represent the official views of the National Institutes of Health. Dr. Miaskowski is an American Cancer Society Clinical Research Professor.

Footnotes

Competing Interests

The authors have no relevant financial or non-financial interests to disclose.

Availability of Data and Material

Available with reasonable request.

Ethics Approval

The study procedures were approved by the Committee on Human Research at the University of California, San Francisco and the Institutional Review Board at each of the study sites. This study was performed in accordance with the ethnical standards as laid down in the 1964 Helsinki Declaration.

Consent to Participate

Informed consent was obtained from all individual participants included in this study.

References

  • 1.Kim JE, Dodd MJ, Aouizerat BE, Jahan T, Miaskowski C (2009) A review of the prevalence and impact of multiple symptoms in oncology patients. J Pain Symptom Manage 37(4):715–36. 10.1016/j.jpainsymman.2008.04.018 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Thiagarajan M, Chan CM, Fuang HG, Beng TS, Atiliyana MA, Yahaya NA (2016) Symptom prevalence and related distress in cancer patients undergoing chemotherapy. Asian Pac J Cancer Prev 17(1):171–6. 10.7314/apjcp.2016.17.1.171 [DOI] [PubMed] [Google Scholar]
  • 3.Dodd M, Miaskowski C, Paul S (2001) Symptom clusters and their effect on the functional status of patients with cancer. Oncol Nurs Forum 28(3):465–470. [PubMed] [Google Scholar]
  • 4.Given CW, Given B, Azzouz F, Kozachik S, Stommel M (2001) Predictors of pain and fatigue in the year following diagnosis among elderly cancer patients. J Pain Symptom Manage 21(6):456–466. [DOI] [PubMed] [Google Scholar]
  • 5.Miaskowski C, Barsevick A, Berger A, et al. (2017) Advancing symptom science through symptom cluster research: Expert panel proceedings and recommendations. J Natl Cancer Inst 109(4). 10.1093/jnci/djw253 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Sullivan CW, Leutwyler H, Dunn LB, Miaskowski C (2018) A review of the literature on symptom clusters in studies that included oncology patients receiving primary or adjuvant chemotherapy. J Clin Nurs 27(3–4):516–545. 10.1111/jocn.14057 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Harris CS, Kober KM, Conley YP, Dhruva AA, Hammer M, Miaskowski CA (2021) Symptom clusters in patients receiving chemotherapy: A systematic review. BMJ Support Palliat Care Preprint. 10.1136/bmjspcare-2021-003325 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Chen ML, Lin CC (2007) Cancer symptom clusters: A validation study. J Pain Symptom Manage 34(6):590–9. 10.1016/j.jpainsymman.2007.01.008 [DOI] [PubMed] [Google Scholar]
  • 9.Chen ML, Tseng HC (2006) Symptom clusters in cancer patients. Support Care Cancer 14(8):825–30. 10.1007/s00520-006-0019-8 [DOI] [PubMed] [Google Scholar]
  • 10.Cherwin CH, Perkhounkova Y (2017) Distress-based gastrointestinal symptom clusters and impact on symptom interference and quality of life in patients with a hematologic malignancy receiving chemotherapy. J Pain Symptom Manage 53(4):751–758. 10.1016/j.jpainsymman.2016.11.005 [DOI] [PubMed] [Google Scholar]
  • 11.Karabulut N, Erci B, Ozer N, Ozdemir S (2010) Symptom clusters and experiences of patients with cancer. J Adv Nurs 66(5):1011–21. 10.1111/j.1365-2648.2009.05254.x [DOI] [PubMed] [Google Scholar]
  • 12.Matzka M, Köck-Hódi S, Jahn P, Mayer H (2018) Relationship among symptom clusters, quality of life, and treatment-specific optimism in patients with cancer. Support Care Cancer 26(8):2685–2693. 10.1007/s00520-018-4102-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Papachristou N, Barnaghi P, Cooper B, et al. (2019) Network analysis of the multidimensional symptom experience of oncology. Sci Rep 9:1–11. 10.1038/s41598-018-36973-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Yamagishi A, Morita T, Miyashita M, Kimura F (2009) Symptom prevalence and longitudinal follow-up in cancer outpatients receiving chemotherapy. J Pain Symptom Manage 37(5):823–30. 10.1016/j.jpainsymman.2008.04.015 [DOI] [PubMed] [Google Scholar]
  • 15.Yates P, Miaskowski C, Cataldo JK, et al. (2015) Differences in composition of symptom clusters between older and younger oncology patients. J Pain Symptom Manage 49(6):1025–1034. 10.1016/j.jpainsymman.2014.11.296 [DOI] [PubMed] [Google Scholar]
  • 16.Saeidzadeh S, Perkhounkova Y, Gilbertson-White S, Cherwin CH (2022) The influence of multiple chronic conditions on symptom clusters in people with solid tumor cancers. Cancer Nurs 45(1):E279–E290. 10.1097/ncc.0000000000000915 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Sullivan CW, Leutwyler H, Dunn LB, et al. (2018) Stability of symptom clusters in patients with breast cancer receiving chemotherapy. J Pain Symptom Manage 55(1):39–55. 10.1016/j.jpainsymman.2017.08.008 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Han CJ, Reding K, Cooper BA, et al. (2019) Symptom clusters in patients with gastrointestinal cancers using different dimensions of the symptom experience. J Pain Symptom Manage 58(2):224–234. 10.1016/j.jpainsymman.2019.04.035 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Pozzar RA, Hammer MJ, Cooper BA, et al. (2021) Symptom clusters in patients with gynecologic cancer receiving chemotherapy. Oncol Nurs Forum 48(4):441–452. 10.1188/21.ONF.441-452 [DOI] [PubMed] [Google Scholar]
  • 20.Russell J, Wong ML, Mackin L, et al. (2019) Stability of symptom clusters in patients with lung cancer receiving chemotherapy. J Pain Symptom Manage 57(5): p. 909–922. 10.1016/j.jpainsymman.2019.02.002 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Karnofsky D (1977) Performance scale. In: Kennealey G, Mitchell M (ed) Factors that influence the therapeutic response in cancer: A comprehensive treatise. Plenum Press: New York, NY [Google Scholar]
  • 22.Sangha O, Stucki G, Liang MH, Fossel AH, Katz JN (2003) The Self-Administered Comorbidity Questionnaire: A new method to assess comorbidity for clinical and health services research. Arthritis Rheum 49(2):156–63. 10.1002/art.10993 [DOI] [PubMed] [Google Scholar]
  • 23.Extermann M, Bonetti M, Sledge GW, O’Dwyer PJ, Bonomi P, Benson III AB (2004) MAX2--a convenient index to estimate the average per patient risk for chemotherapy toxicity: Validation in ECOG trials. Eur J Cancer 40(8):1193–8. 10.1016/j.ejca.2004.01.028 [DOI] [PubMed] [Google Scholar]
  • 24.Utne I, Loyland B, Grov EK et al. (2018) Distinct attentional function profiles in older adults receiving cancer chemotherapy. Eur J Oncol Nurs 36:32–39. 10.1016/j.ejon.2018.08.006 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Portenoy RK, Thaler HT, Kornblith AB, et al. (1994) The Memorial Symptom Assessment Scale: An instrument for the evaluation of symptom prevalence, characteristics and distress. Eur J Cancer 30a(9):1326–36. 10.1016/0959-8049(94)90182-1 [DOI] [PubMed] [Google Scholar]
  • 26.Muthén L, Muthén B (2019) Mplus. Muthen & Muthen: Los Angeles, CA [Google Scholar]
  • 27.Brown T (2015) The common factor model and exploratory factor analysis, 2nd edn. The Guilford Press, London [Google Scholar]
  • 28.Han CJ, Reding K, Cooper BA, et al. (2019) Stability of symptom clusters in patients with gastrointestinal cancers receiving chemotherapy. J Pain Symptom Manage 58(6):989–1001. 10.1016/j.jpainsymman.2019.07.029 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Wong ML, Cooper BA, Paul SM, et al. (2017) Differences in symptom clusters identified using ratings of symptom occurrence vs. severity in lung cancer patients receiving chemotherapy. J Pain Symptom Manage 54(2):194–203. 10.1016/j.jpainsymman.2017.04.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Sullivan CW, Leutwyler H, Dunn LB, et al. (2017) Differences in symptom clusters identified using symptom occurrence rates versus severity ratings in patients with breast cancer undergoing chemotherapy. Eur J Oncol Nurs 28:122–132. 10.1016/j.ejon.2017.04.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Pozzar RA, Hammer MJ, Cooper B, et al. (2021) Stability of symptom clusters in patients with gynecologic cancer receiving chemotherapy. Cancer Nurs preprint. 10.1097/ncc.0000000000000988 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Chongkham-ang S, Wonghongkul T, Panuthai S, Pinyokham N, Miaskowski C (2018) Symptom experience and symptom clusters of Thai women with breast cancer receiving chemotherapy. Pac Rim Int J Nurs Res 22(1):43–57. [Google Scholar]
  • 33.Lin DM, Yin XX, Wang N, et al. (2019) Consensus in identification and stability of symptom clusters using different symptom dimensions in newly diagnosed acute myeloid leukemia patients undergoing induction therapy. J Pain Symptom Manage 57(4):783–792. 10.1016/j.jpainsymman.2018.12.329 [DOI] [PubMed] [Google Scholar]
  • 34.Kirkova J, Walsh D (2007) Cancer symptom clusters—A dynamic construct. Support Care Cancer 15(9):1011–1013. 10.1007/s00520-007-0259-2 [DOI] [PubMed] [Google Scholar]
  • 35.Cherwin C, Kwekkeboom K (2016) Prevalence, duration, severity, and distress of chemotherapy-related gastrointestinal symptoms in patients with a hematologic malignancy. Oncol Nurs Forum 43(5):561–71. 10.1188/16.ONF.43-05AP [DOI] [PubMed] [Google Scholar]
  • 36.Cherwin CH (2012) Gastrointestinal symptom representation in cancer symptom clusters: A synthesis of the literature. Oncol Nurs Forum 39(2):157–165. 10.1188/12.ONF.157-165 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Kim S (2018) A longitudinal study of lipid peroxidation and symptom clusters in patients with brain cancers. Nurs Res 67(5):387–394. 10.1097/nnr.0000000000000302 [DOI] [PubMed] [Google Scholar]
  • 38.Molassiotis A, Wengstrom Y, Kearney N (2010) Symptom cluster patterns during the first year after diagnosis with cancer. J Pain Symptom Manage 39(5):847–58. 10.1016/j.jpainsymman.2009.09.012 [DOI] [PubMed] [Google Scholar]
  • 39.Li HJ, Sereika SM, Marsland AL, Conley YP, Bender CM (2020) Symptom clusters in women with breast cancer during the first 18 months of adjuvant therapy. J Pain Symptom Manage 59(2): 10.1016/j.jpainsymman.2019.10.002\. [DOI] [PubMed] [Google Scholar]
  • 40.Aprile G, Basile D, Giaretta R, et al. (2021) The clinical value of nutritional care before and during active cancer treatment. Nutrients 13(4):1196. 10.3390/nu13041196 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Seib C, Porter-Steele J, McGuire A, McCarthy A, Balaam S, Anderson DJ (2017) Menopausal symptom clusters and their correlates in women with and without a history of breast cancer: A pooled data analysis from the Women’s Wellness Research Program. Menopause 24(6):624–634. 10.1097/gme.0000000000000810 [DOI] [PubMed] [Google Scholar]
  • 42.Del Carmen MG, Rice LW (2017) Management of menopausal symptoms in women with gynecologic cancers. Gynecol Oncol 146(2):427–435. 10.1016/j.ygyno.2017.06.013 [DOI] [PubMed] [Google Scholar]
  • 43.Huang J, Gu L, Zhang L, Lu X, Zhuang W, Yang Y (2016) Symptom clusters in ovarian cancer patients with chemotherapy after surgery: A longitudinal survey. Cancer Nurs 39(2):106–16. 10.1097/NCC.0000000000000252 [DOI] [PubMed] [Google Scholar]

RESOURCES