|Home | About | Journals | Submit | Contact Us | Français|
Findings from several studies suggest that oncology patients undergoing active treatment experience multiple symptoms and that these symptoms can have a negative effect on patient outcomes. However, no systematic review has summarized the findings from studies that assessed multiple symptoms in these patients. Therefore, the purposes of this review were to: 1) compare and contrast the characteristics of the three most commonly used instruments to measure multiple symptoms; 2) summarize the prevalence rates for multiple symptoms in studies of oncology patients receiving active treatment; 3) describe the relationships among selected demographic, disease, and treatment characteristics and multiple symptoms; and 4) describe the relationships between the occurrence of multiple symptoms and patient outcomes (i.e., functional status, quality of life). Only 18 studies were found that met the inclusion criteria for this review. The majority of the studies were cross-sectional with sample sizes that ranged from 26 to 527. Approximately 40% of patients experienced more than one symptom. However, little is known about the relationships between demographic and clinical characteristics and the occurrence of multiple symptoms. Findings from this review suggest that the occurrence of multiple symptoms is associated with decreased functional status and quality of life. However, given the large number of oncology patients who undergo active treatment each year, additional research is warranted on the prevalence and impact of multiple symptoms. Only when this descriptive research is completed with homogenous samples of patients in terms of cancer diagnoses and treatments can intervention studies for multiple symptoms be developed and tested.
Patients with cancer can undergo a variety of treatments (e.g., surgery, radiation (RT), chemotherapy (CTX), hormonal therapy), either singly or in combination. While these treatments improve survival, they can produce a variety of symptoms. In fact, findings from several studies suggest that patients receiving active treatment (1,2) experience multiple symptoms simultaneously. For example, in one of the first studies of multiple symptoms (3), women with ovarian cancer reported an average of 10.2 symptoms (range of 0 to 25 concurrent symptoms). More recently, Donovan and colleagues (4) found that 74% of women who received CTX for ovarian cancer reported 13.4 concurrent symptoms.
When these symptoms are not managed effectively, they can cause interruptions or cessation of cancer treatment (5) or decrease patients’ level of adherence with a treatment regimen (6–12). In addition, unrelieved symptoms can have a negative impact on patients’ functional status, mood, and quality of life (QOL) (5,13–16).
Given the negative outcomes associated with multiple symptoms, it seems prudent that clinicians and researchers should evaluate the prevalence and impact of multiple symptoms in oncology patients undergoing active treatment. These types of evaluations could be used to guide the development and testing of interventions for multiple symptoms. However, no systematic review has summarized the findings from studies that evaluated multiple symptoms in oncology patients receiving active treatment. Therefore, the purposes of this review were to: 1) compare and contrast the characteristics of the three most commonly used instruments to measure multiple symptoms; 2) summarize the prevalence rates for multiple symptoms in studies of oncology patients receiving active treatment; 3) describe the relationships among selected demographic, disease, and treatment characteristics and multiple symptoms; and 4) describe the relationships between the occurrence of multiple symptoms and patient outcomes (i.e., functional status, QOL).
For this review, systematic electronic searches of MEDLINE, Cumulative Index to Nursing and Allied Health Literature (CINAHL), and PsycINFO databases were performed. The searches were restricted to adults with cancer and English language articles. The search terms used were symptom, multiple symptoms, cancer, cancer treatment, QOL, and symptom assessment instruments. The searches were limited to the years 1990 through 2007 because no studies of multiple symptoms in oncology patients were published prior to 1990.
Studies were included if they: evaluated the prevalence of multiple (greater than one) symptoms; used one of three valid and reliable instruments (i.e., the Symptom Distress Scale (SDS) (17), the M. D. Anderson Symptom Inventory (MDASI) (18), the Memorial Symptom Assessment Scale (MSAS) (19)) to evaluate multiple symptoms; and included adult oncology patients who were receiving active treatment in inpatient or outpatient settings. Studies were excluded if they: evaluated multiple symptoms in patients who were receiving palliative or hospice care; measured the side effects of treatment; and/or used QOL instruments or symptom specific instruments to measure multiple symptoms.
The retrieved studies were reviewed by the first author (J-EK) initially to determine if they met the inclusion and exclusion criteria. Then the reference lists of selected studies were manually searched to identify any additional studies. Based on the search parameters, 76 abstracts were identified for this review. A total of 69 studies addressed some aspect of multiple symptoms. Fifty-one abstracts were eliminated because they used instruments without established validity and reliability (i.e., Canberra Symptom Score Card, Chemotherapy Symptom Assessment Scale, a Computerized Symptom Assessment Instrument, Pain and Symptom Assessment Record, Symptom Experience Scale, the Symptom Monitor, the Symptom Reporting Tool, the modified Edmonton Symptom Assessment System (ESAS), the MSAS Modified for family caregivers). Therefore, 18 studies of multiple symptoms met the prespecified inclusion/exclusion criteria (see Table 1).
The ideal instrument to measure multiple symptoms should include those symptoms that occur frequently and are most distressing to patients. In addition, it should be relatively short, easy for patients with limited educational backgrounds to understand, and applicable for both clinical practice and research (5,18). Ideally, the instrument should be available in multiple languages.
Several instruments are available to measure multiple symptoms including the ESAS (20,21), the MDASI (18), the MSAS (19), the Oncology Treatment Toxicity Assessment Tool (OTTA) (2), the Rotterdam Symptom Checklist (RSCL) (22), the SDS (17,23), and the Worthing Chemotherapy Questionnaire (24). All of these instruments are comprehensive and have good psychometric properties. For this review, the ESAS, RSCL, OTTA, and the Worthing Chemotherapy Questionnaire were excluded for a number of reasons. The ESAS was designed to assess symptoms in palliative care patients. The RSCL was designed to assess symptoms in cancer patients who participated in a clinical trial. Both the OTTA and the Worthing Chemotherapy Questionnaire assess treatment-related side effects.
It should be noted that several symptom specific instruments and QOL scales can be used to evaluate the presence and severity of a single symptom or pairs of symptoms such as fatigue (e.g., Piper Fatigue Scale (25)), depression and anxiety (e.g., Hospital Anxiety and Depression Scale (26)), and nausea and vomiting (e.g., Rhodes Index of Nausea and Vomiting, (27)). While symptom-specific scales provide valuable information on the multiple dimensions of a single symptom and many QOL instruments contain items that evaluate multiple symptoms often as part of physical and psychological subscales, they capture different aspects of the symptom experience compared to valid and reliable instruments that were designed to capture the occurrence, severity, and/or distress of multiple concurrent symptoms in patients undergoing active cancer treatment. In addition, most of the multidimensional QOL instruments contain only a limited number of common symptoms (19). Therefore, in this review, only those studies that used one of three symptom inventories (i.e., SDS, MDASI, MSAS) were reviewed because they are valid and reliable measures that provide information about a large number of physical and psychological symptoms that are assessed concurrently. The psychometric properties of these three instruments are summarized in Table 2.
The MDASI was developed by the Pain Research Group at the University of Texas M. D. Anderson Cancer Center (18). The original tool included 26 symptoms. However, 13 items (i.e., not able to get things done, weak, worrying, nervous, irritable, sick, constipation, attention, bloated, cough, diarrhea, mouth sores, bleeding) were deleted because they were deemed redundant or had low prevalence rates in oncology patients. The MDASI measures the severity of 10 physical symptoms (i.e., pain, fatigue (tiredness), disturbed sleep, dry mouth, lack of appetite, nausea, vomiting, drowsy, shortness of breath, numbness or tingling), three psychological symptoms (i.e., problem with remembering things, feeling sad, distress), and six interference items (i.e., general activity, mood, work, relations with other people, walking, enjoyment of life).
Each symptom is rated on an 11-point numeric rating scale (NRS) with 0 indicating “not present” and 10 indicating “as bad as you can imagine.” Each symptom on the MDASI is rated at its worst in the past 24 hours. Six interference items that describe how much all of the symptoms interfere with common activities are rated using an 11-point NRS (i.e., 0 “does not interfere to 10 “interferes completely”). Of note, the final 13 symptoms explained 64% of the variance in symptom interference. Validity of the MDASI was determined using factor analysis and internal reliabilities ranged from 0.82 to 0.87 for the symptom items and from 0.91 to 0.94 for the interference items (18).
The MSAS is a self-report instrument that measures, using Likert scales, the severity (1 (mild) to 4 (very severe)), frequency (1 (rarely) to 4 (almost constantly)), and distress (0 (not at all) to 4 (very much)) of 26 physical and 6 psychological symptoms (i.e., difficulty concentrating, feeling sad, worrying, feeling nervous, feeling irritable, and “I don’t look like myself”) in cancer patients during the previous seven days (19). It provides multidimensional information about a large number of symptoms that are experienced by oncology patients. Twenty-four symptoms are evaluated in terms of all three dimensions (i.e., severity, frequency, distress), and eight symptoms (i.e., mouth sores, change in the way food tastes, weight loss, constipation, hair loss, swelling of arms or legs, changes in skin, “I don’t look like myself”) are evaluated for only severity and distress.
The MSAS is scored into physical and psychological subscales as well as a Global Distress Index (GDI). The GDI is made up of four prevalent psychological symptoms (i.e., feeling sad, worrying, feeling irritable, feeling nervous) and six prevalent physical symptoms (i.e., lack of energy, dry mouth, lack of appetite, pain, constipation, feeling drowsy). The GDI provides a measure of global symptom distress.
The physical symptom subscale score (MSAS-PHYS) is the average of the frequency, severity, and distress of the 12 most prevalent physical symptoms. The psychological symptom subscale score (MSAS-PSYCH) is the average of the frequency, severity, and distress of the six most prevalent psychological symptoms. The total MSAS (TMSAS) score is the average of the three symptom scores for all 32 symptoms. The MSAS has demonstrated validity and reliability in patients with cancer (19,21). Concurrent validity of the MSAS was demonstrated through a strong positive correlation with the Functional Living Index for Cancer (FLIC) QOL measure and with the Karnofsky Performance Status (KPS) Score (19). Construct validity was determined through comparisons of MSAS scores among different cancer diagnoses. Discriminant validity was determined by comparing the MSAS scores of inpatients and outpatients (19).
The SDS is a measure of symptom distress defined as “the degree of discomfort from specific symptoms being experienced as reported by the patient” (17). It provides a measure of the severity of symptom distress and was one of the first valid and reliable instruments developed for symptom assessment in oncology patients. This 13-item self-report instrument assesses the level of symptom distress for 13 symptoms (i.e., pain, fatigue, insomnia, lack of appetite, nausea, bowel dysfunction, shortness of breath, coughing, poor activity, difficulty with concentration, mood, altered appearance, poor outlook). In addition, the frequency of occurrence of pain and nausea are reported separately.
Each item is scored on a 5-point Likert scale (i.e., 1= the least amount of distress associated with a symptom to 5 = extreme distress associated with a symptom) that measures the distress associated with each symptom at that moment or for that day. Items rated ≥ 3 indicate serious distress. A total score is obtained by summing the scores for the 13 items and can range from 13 (little distress) to 65 (severe symptom distress). A total score of ≥ 25 indicates moderate distress and a score of ≥ 33 indicates severe distress that requires immediate intervention (23). The SDS has demonstrated acceptable internal consistency (Cronbach α > 0.80) and test-retest reliability in patients with lung cancer (28) as well as content, construct, and criterion validity.
The specific symptoms that are measured by each of these instruments were summarized in Table 3. While these three instruments purport to measure “common” symptoms in oncology patients, the number as well as the specific symptoms that are assessed vary across the three instruments. Only eight symptoms (i.e., pain, fatigue, difficulty sleeping, lack of appetite, nausea, shortness of breath, difficulty with concentration, mood/sad) are measured by all three instruments.
Another difference among these three instruments is the aspect of the symptom experience that is assessed. The SDS focuses on the distress associated with each symptom, which is suggested to be a proxy for symptom severity. While the MDASI measures the severity of each symptom individually, interference is assessed for all of the symptoms collectively. In contrast, the MSAS measures frequency, severity, and distress for each symptom. However, the MDASI and the MSAS measure symptom severity using different scales. In addition, the instructions for the MDASI ask patients to rate symptoms at their worst, while the MSAS asks for ratings of average symptom severity.
Although both the SDS and the MSAS measure symptom distress, it is assessed using different scales (i.e., a 1 to 5 scale on the SDS versus a 0 to 4 scale on the MSAS). Another difference is that the timeframe for symptom assessment varies across the three instruments (i.e., SDS and MDASI = “at that moment or on that particular day;” MSAS = “past week”).
The SDS and the MDASI take 5 to 10 minutes to complete. No information is available on how long it takes to complete the MSAS. The psychometric properties of these three instruments are well established. Factor analysis of the MSAS and MDASI confirmed the factor structure of these instruments (18,19). The Cronbach’s alphas for the three instruments are comparable (see Table 2).
All three instruments were developed in the United States. The SDS and the MDASI have been translated and validated in several languages. Translations of the SDS are available in Dutch, Italian, Spanish, Swedish, Korean, and Taiwanese. The MDASI has been translated into Chinese, Korean, Japanese, Greek, Russian, and Filipino. No information was found on translations of the MSAS.
Table 1 provides a summary of the 18 studies that evaluated multiple symptoms in adult oncology patients receiving active treatment. Of these 18 studies, six (37%) used the SDS (29–34), seven (39%) used the MDASI (4,18,35–39), and five (28%) used the MSAS (3,19,21,40,41).
Sixteen studies (89%) used a cross-sectional design, while only two (11%) were longitudinal. Prospective data were collected in all of the cross-sectional studies. One of the longitudinal studies (33) used a prospective design and assessed multiple symptoms and symptom distress at the initiation of treatment, and one and two months later. The other longitudinal study (34) evaluated multiple symptoms at the start of treatment and again at three and six months. Both of these longitudinal studies used the SDS to describe the patterns of symptom distress in patients with lung cancer.
Sample sizes for the 18 studies varied widely and ranged from 26 (33) to 527 participants (18). Five studies (28%) had sample sizes of less than 100 (29,30,32,33,40). The remaining 13 studies (72%) had sample sizes that ranged from 117 to 527. All of these studies recruited convenience samples from multiple sites.
The mean age of the participants was 59.1 with a range from 47.0 to 66.9 years. About 78% of the studies (n=14) enrolled both genders and overall 52% of the participants were male. Four studies measured symptoms only in women with lung or ovarian cancer (3,4,30,42).
Approximately, 44% of the studies (n=8) were conducted in United States (3,4,18,19,21,29,30,34). Across these eight studies, the majority (73%) of the participants were Caucasian (range 63% to 94%). Of the remaining 10 studies, three were done in Canada (31,32,40), two in China (36,41), one in Sweden (33), one in Japan (35), one in Russia (37), one in the Philippines (38), and one in Taiwan (39).
Fifty percent of the studies (n=9) collected data from heterogeneous samples of patients with a variety of cancer diagnoses (18,19,21,31,35,37,38–40). Regarding the site of cancer, four studies (22%) assessed symptoms only in patients with lung cancer (30,32–34), two (11%) assessed patients with ovarian cancer (3,4), and one (6%) assessed patients with gastrointestinal cancers (41). One study (6%) failed to provide information on cancer diagnosis (29).
Among the six studies that assessed multiple symptoms using the SDS, four studies evaluated patients with lung cancer (30,32–34), one assessed a heterogeneous sample (31), and one failed to describe the patients’ cancer diagnoses (29). In the seven studies that used the MDASI, five recruited patients with a variety of cancer diagnoses (18,35,37–39), one recruited only patients with ovarian cancer (4), and one recruited only patients with lung cancer (36). Of the five studies that used the MSAS, one study assessed patients with prostate, colon, breast, and ovarian cancers (19), one assessed patients with ovarian cancer (3), one with gastrointestinal cancers (41), and two evaluated heterogeneous samples (21,40).
The patients’ stage of disease varied across these studies. Four studies (29,33,40,41) did not provide any information on stage of disease, while three (19,21,35) found that the presence of metastatic disease was associated with an increased number of symptoms. Across the remaining 11 studies, 52% of the patients (range 12% to 87%) had stage III-IV disease. In sixteen studies, patients were receiving active treatment with CTX, RT, biotherapy, surgery, or a combination of treatments. No information on the specific treatments was provided in two studies (19,33).
Table 4 provides a summary of the prevalence rates for the various symptoms in each of the studies as well as a mean prevalence rate across these studies. Symptom prevalence rates ranged from 11% for sore mouth to 62% for fatigue. The ten most prevalent symptoms across the 18 studies were fatigue (62%), worrying (54%), feeling nervous (45%), dry mouth (42%), insomnia (41%), feeling sad/mood (39%), feeling irritable (37%), pain (36%), drowsiness (36%), and distress (34%). The prevalence rates for these 10 symptoms ranged from 34% to 62%. Across the 18 studies, 40% to 61% of patients experienced more than one symptom (30,36) and 22% to 30% of patients experienced more than five concurrent symptoms (18,36,37).
The relationships among a variety of demographic characteristics and the type of symptom experienced were examined in only four studies (19,21,31,41). However, the findings from these studies are inconsistent. Only two studies evaluated for age differences in the severity of symptom distress. In one study (31), age was weakly correlated with symptoms distress (r = −0.11, P< 0.02) and younger patients tended to have higher levels of symptom distress than older patients. In contrast, Yan and Sellick (41) found that patients in their older age group (≥ 70 years) reported higher symptom distress scores than those in their younger age group (< 40 years).
In addition to age, gender differences in symptom distress (31) and symptom prevalence (19) were evaluated in only two studies. In one study that used the SDS (31), women reported higher symptom distress scores than men (P < 0.041). In another study that used the MSAS (19), no gender differences were found in any of the symptom prevalence rates.
The relationships between site and stage of cancer and symptom severity and distress were evaluated in five studies (21,30,31,36,41). In one study (31), patients with lung cancer had higher symptom distress scores than either women with breast cancer or males with genitourinary cancer. In a study of Chinese patients (36), fatigue and sleep disturbance were the most common symptoms in patients with breast and lung cancer, whereas fatigue and lack of appetite were the most common symptoms in patients with gastrointestinal cancer. In another study of symptoms, psychological distress, and QOL in Chinese patients with newly diagnosed gastrointestinal cancer (41), patients with liver cancer had higher symptom frequency, severity, and distress scores than patients with all other gastrointestinal cancer diagnoses. Findings across these three studies suggest that patients with recurrent (30), metastatic (21), or advanced stage of disease (31) reported the most severe and distressing symptoms.
The type of cancer treatment appeared to influence the prevalence and severity of multiple symptoms. However, of the 18 studies, only three evaluated the prevalence of symptoms and symptom severity/distress in relationship to type of cancer treatment (29,30,34). In a study that compared mean SDS scores of patients who received CTX versus RT (29), patients who received CTX reported higher SDS scores especially for tiredness and poor appearance compared to those who received RT. In another study (30), patients who received CTX reported higher symptom distress scores than patients who underwent surgery. More recently, the prevalence of distressing symptoms was evaluated in patients who received a variety of treatments for lung cancer (34). At entry into the study, the three most distressing symptoms for patients with surgery were pain, fatigue, and insomnia; for patients with RT they were fatigue, lack of appetite, and nausea; for patients with CTX they were fatigue, insomnia, and lack of appetite; and for patients with combined treatments they were fatigue, pain, and insomnia. Patients who received only RT reported a significantly higher number of symptoms across time compared to the other three groups.
Symptom prevalence rates appear to differ based on the settings of care. Findings from two studies (3,19) found that inpatients reported a higher number of symptoms than outpatients. The mean number of symptoms for inpatients with ovarian cancer was 11.2 (range of 1 to 25) compared to 7.4 for outpatients (range of 0 to 16, (3)). In another study (19), the mean number of symptoms for inpatients with various cancers was 13.5 compared to 9.7 for outpatients.
The relationships between symptoms and functional status and QOL were examined in only five (28%) of the 18 studies (19,21,30,31,41). Two studies examined the relationships between the number of symptoms, symptom distress, and functional status (19,30). In a study of symptom distress and functional status in women with lung cancer (30), as symptom distress increased, functional status decreased. The other study (19) reported that the higher the number of symptoms, the poorer the patients’ functional status. Patients with KPS scores of ≤ 80 reported 14.8 symptoms while patients with KPS scores of > 80 reported only 9.2 symptoms (P<0.0001).
Four studies found that patients who reported a larger number of symptoms or symptom distress had poorer QOL scores (19,21,30,41). Sarna (30) reported that higher levels of symptom distress in women with lung cancer were significantly correlated with decreases in both the physical and psychological dimensions of QOL. In another study of patients newly diagnosed with gastrointestinal cancers (41), those who reported lower levels of symptoms distress reported higher QOL scores.
Findings from two studies that used the MSAS (19,21) suggest that a higher number of symptoms was strongly correlated with poorer QOL. In one study of 243 adults with various types of cancer (19), significant negative correlations were found between the number of symptoms and patients’ overall QOL (r= −0.67, P< 0.0001). In addition, higher symptom distress scores were associated with increased psychological distress. Another study (21) confirmed that a higher number of symptoms was associated with a poorer QOL. Finally, two studies (31,34) found that symptom distress at diagnosis was a significant predictor of symptom distress over time, as well as decreased functional status, poorer QOL, and decreased survival. Cooley et al. (34) reported that baseline symptom distress predicted nine distressing symptoms at three months and seven distressing symptoms at six months in 117 patients with newly diagnosed lung cancer.
This review is the first to evaluate the prevalence of, as well as the factors associated with, the occurrence of multiple symptoms in adult oncology patients undergoing active treatment. Findings from a limited number of studies suggest that the prevalence rates for multiple symptoms are relatively high. Across 18 studies, more than 50% of oncology patients reported experiencing fatigue and worry. Of note, fatigue was the most prevalent symptom across the 18 studies. In addition, findings from these studies suggest that multiple symptoms are associated with decreases in functional status and QOL. The occurrence of multiple symptoms may be related the disease itself, active treatment, sequelae of treatment, or comorbid conditions. Finally, the experience of multiple symptoms is associated with higher levels of symptom distress.
Several limitations across these studies must be noted. First, of the 18 studies, 89% were descriptive and cross-sectional. Therefore, little is known about how multiple symptoms change across the course of a patient’s treatment trajectory. Longitudinal studies are needed to describe the trajectories of multiple symptoms in oncology patients undergoing active treatment. Without these descriptive, longitudinal studies, it will be difficult to plan intervention studies to manage multiple symptoms.
Second, all of the studies in this review used convenience samples which limit the generalizability of the study findings. In addition, the majority of the patients were Caucasian. Future research should evaluate the prevalence and severity of multiple symptoms in more ethnically diverse samples because some data suggest that differences in symptom severity and distress do occur across ethnic groups (43–45). Third, relationships between various patient and disease characteristics and multiple symptoms warrant additional investigation since only a few studies have examined this aspect.
Perhaps one of the major areas that needs to be addressed in future studies of multiple symptoms is which symptoms should be included in any comprehensive symptom inventory. The number of symptoms in the three instruments included in this review range from 13 to 32 symptoms and only eight of these symptoms are common across instruments. However, it is not clear if any of these instrument’s list of symptoms is comprehensive and appropriate for all cancer diagnoses and treatments. As equally important question that warrants consideration is what symptom dimensions (i.e., severity, frequency, and/or distress) should be assessed to capture the patient’s experience of multiple symptoms.
Clinical experience suggests that cancer and its treatment is marked by the occurrence of multiple symptoms that influence the patient’s ability to continue usual activities and enjoy life. However, a very limited number of studies have attempted to measure the prevalence and impact of multiple symptoms in patients with cancer. The gaps in knowledge identified in this review warrant additional research. That said, within the past five years, the concept of a symptom cluster has emerged as an important area in symptom management research (46–48). However, this concept is still in its infancy and warrants additional concept clarification and refinements in its methodology and approaches (49). Therefore, at the present time studies of multiple symptoms need to continue particularly in samples of patients with homogeneous cancer diagnoses and cancer treatments. These types of studies will guide the development of intervention studies as well as symptom cluster research.
Drs. Miaskowski and Dodd’s research programs are supported by grants from the National Cancer Institute and the National Institute of Nursing Research (NINR). Dr. Aouizerat is funded through the National Institutes of Health Roadmap for Medical Research Grant (8 K12 RR023262). Ms. Kim was funded during her doctoral program through a T32 Grant (NR07088) from the NINR.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.