Advertisement

Development and Validation of the Cancer Fatigue Scale

A Brief, Three-Dimensional, Self-Rating Scale for Assessment of Fatigue in Cancer Patients

      Abstract

      We herein describe the development and validation of the Cancer Fatigue Scale (CFS) for assessment of fatigue in cancer patients. We designed this scale specifically to reflect the nature of fatigue experienced by cancer patients, by using factor analysis; the CFS is a 15-item scale composed of 3 subscales (physical, affective, and cognitive subscales). Three hundred seven cancer patients participated in the validation phase. Construct validity, confirmed by repeating factor analysis, was good. Convergent validity, confirmed by a correlation between CFS and a visual analogue scale for fatigue, was also shown to be good (r = 0.67, P < 0.001). The CFS had good stability (average test–retest reliability r = 0.69, P < 0.001) and good internal consistency (Cronbach's alpha coefficient for all 15 items = 0.88). The present study indicates that the CFS is a brief, valid, and feasible measure of fatigue for use with cancer patients.

      Keywords

      Introduction

      Fatigue is one of the most pervasive symptoms experienced by patients with cancer. Prior studies showed that the prevalence of fatigue is greater than 50% for advanced cancer patients
      • Coyle N.
      • Adelhardt J.
      • Foley K.M.
      • Portenoy R.K.
      Character of terminal illness in the advanced cancer patient pain and other symptoms during the last four weeks of life.

      Vainio A, Auvinen A, with Members of the Symptom Prevalence Group. Prevalence of symptoms among patients with advanced cancer: an international collaborative study. J Pain Symptom Manage 1996;12:3–10.

      • Sureshkumar K.
      • Rajagopal M.R.
      Palliative care in Kerala problems at presentation in 440 patients with advanced cancer in a south Indian state.
      and for cancer patients who have undergone chemotherapy,
      • Nerentz D.R.
      • Leventhal H.
      • Love R.R.
      Factors contributing to emotional distress during cancer chemotherapy.
      • Love R.R.
      • Leventhal H.
      • Easterling D.V.
      • et al.
      Side effects and emotional distress during chemotherapy.
      • Cassileth B.R.
      • Farber J.M.
      • Lusk E.J.
      • et al.
      Chemotherapeutic toxicity the relationship between patients' expectations and post-treatment results.
      or radiotherapy.
      • Greenberg D.B.
      • Sawicha J.
      • Eisenthal S.
      • et al.
      Fatigue syndrome due to localized radiation.
      • Hickok J.T.
      • Morrow G.R.
      • McDonald S.
      • et al.
      Frequency and correlates of fatigue in lung cancer patients receiving radiation therapy.
      • Smets E.M.A.
      • Garssen B.
      • Cull A.
      • et al.
      Application of the Multidimensional Fatigue Inventory (MFI-20) in cancer patients receiving radiotherapy.
      Surprisingly, fatigue is prevalent in long-term survivors who no longer suffer from cancer itself.
      • Fobair P.
      • Hoppe R.T.
      • Bloom J.
      • et al.
      Psychosocial problems among survivors of Hodgkin's disease.
      • Devlen J.
      • Maguire P.
      • Phillips P.
      • et al.
      Psychological problems associated with diagnosis and treatment of lymphomas. 1 retrospective study, 2: prospective study.
      • Berglund G.
      • Bolund C.
      • Fornander T.
      • et al.
      Late effects of adjuvant chemotherapy and postoperative radiotherapy on quality of life among breast cancer patients.
      • Joly F.
      • Henry-Amar M.
      • Arveux P.
      • et al.
      Late psychosocial sequelae in Hodgkin's disease survivors a French population-based case-control study.
      Fatigue has a serious impact on cancer patients' quality of life,
      • Love R.R.
      • Leventhal H.
      • Easterling D.V.
      • et al.
      Side effects and emotional distress during chemotherapy.
      • Rhodes V.A.
      • Watson P.M.
      • Hanson B.M.
      Patients' descriptions of the influence of tiredness and weakness on self-care abilities.
      and may even increase the risk of suicide.
      • Breitbart W.
      Suicide in cancer patients.
      Fatigue is defined as a condition characterized by a subjective feeling of a decrease in energy, and it has both physical and psychological aspects.
      • Holley S.P.
      Fatigue in cancer patients.
      • Akechi T.
      • Kugaya A.
      • Okamura H.
      • et al.
      Fatigue and its associated factors in ambulatory cancer patients a preliminary study.
      Many authors have pointed out the overlap between fatigue and depression.
      • Nerentz D.R.
      • Leventhal H.
      • Love R.R.
      Factors contributing to emotional distress during cancer chemotherapy.
      • Smets E.M.A.
      • Garssen B.
      • Cull A.
      • et al.
      Application of the Multidimensional Fatigue Inventory (MFI-20) in cancer patients receiving radiotherapy.
      • Fobair P.
      • Hoppe R.T.
      • Bloom J.
      • et al.
      Psychosocial problems among survivors of Hodgkin's disease.
      • Bruera E.
      • MacDonald R.N.
      Asthenia in patients with advanced cancer.
      • Smets E.M.A.
      • Garssen B.
      • Schuster-Uitterhoeve A.L.J.
      • et al.
      Fatigue in cancer patients.
      • Smets E.M.A.
      • Garssen B.
      • Bokne B.
      • et al.
      The Multidimensional Fatigue Inventory (MFI) psychometric qualities of an instrument to assess fatigue.
      • Glaus A.
      • Crow R.
      • Hammond S.
      A qualitative study to explore the concept of fatigue/tiredness in cancer patients and in healthy individuals.
      • Hann D.M.
      • Jacobsen P.B.
      • Martin S.C.
      • et al.
      Fatigue in women treated with bone marrow transplantation for breast cancer a comparison with women with no history of cancer.
      Not only somatic illness but also depression often contributes to a patient's fatigue. For patients with cancer in particular, fatigue can be a chronic problem that is caused by a combination of physical, psychological, and situational factors.
      • Holley S.P.
      Fatigue in cancer patients.
      The development of appropriate interventions requires a clear understanding of the etiology of fatigue. Therefore, the comprehensive assessment of fatigue must be multidimensional, covering both the physical and psychological aspects.
      • Smets E.M.A.
      • Garssen B.
      • Bokne B.
      • et al.
      The Multidimensional Fatigue Inventory (MFI) psychometric qualities of an instrument to assess fatigue.
      Instruments for assessing the fatigue experienced by cancer patients must have three qualities: First, the format must be based on self-rating, because fatigue is a subjective symptom, and the questionnaire must be brief and simple enough to be completed by cancer patients suffering from fatigue. Second, a multidimensional approach is needed, because fatigue is thought to be a multidimensional symptom that has aspects of at least physical and affective sensations. Third, the scale must be specifically designed for cancer-related fatigue, and the reliability and validity of the scale for use by cancer patients must be established.
      Existing fatigue scales can be divided into three categories: there are subscales that assess fatigue as a component of cancer-specific quality-of-life measures. For example, fatigue can be measured using the Quality of Life Questionnaire published by the European Organization for Research and Treatment of Cancer Questionnaire.
      • Aaronsen N.
      • Ahmedzai S.
      • Bergman B.
      • et al.
      The European Organization for Research and Treatment of Cancer QLQ-C30 a quality of life instrument for use in international clinical trials in oncology.
      Alternatively, fatigue can be measured with unidimensional scales, such as the Rhoton Fatigue Scale,

      Rhoten D. Fatigue and the postsurgical patient. Concept Clarification in Nursing. Rockville, MD: Aspen Publishers, 1982:277–300.

      or multidimensional scales, such as the Piper Fatigue Scale
      • Piper B.F.
      • Linsey A.M.
      • Dodd M.J.
      • et al.
      The development of an instrument to measure the subjective dimension of fatigue.
      and the Multidimensional Fatigue Inventory.
      • Smets E.M.A.
      • Garssen B.
      • Bokne B.
      • et al.
      The Multidimensional Fatigue Inventory (MFI) psychometric qualities of an instrument to assess fatigue.
      The first type of fatigue scale is not suitable for studies focusing on fatigue because they are too long if the aim is to only assess fatigue. Furthermore, they only allow for a limited number of possible responses. Unidimensional scales also are not suitable for the in-depth study of fatigue because they assess only one aspect of fatigue. They do have some merit in their simplicity, however.
      In recent years, several multidimensional self-assessment scales of fatigue for use by cancer patients have been published. Although some of these scales have been validated using the responses of cancer patients and are promising, there are yet no fatigue scales designed specifically to reflect the nature of fatigue experienced by cancer patients that have been convincingly validated in these populations. After reviewing the existing scales for measuring fatigue in cancer patients, Richardson
      • Richardson A.
      Measuring fatigue in patients with cancer.
      concluded that there remains a dearth of adequate assessment instruments for measuring the different subjective dimensions of fatigue whose properties have been sufficiently psychometrically tested. Furthermore, to our knowledge, no fatigue scales that address the three qualities mentioned above have been developed in Japan. This lack of adequate instruments prevents clinicians from understanding the etiology of their patients' fatigue and from establishing effective therapies.
      The purpose of the present study was to develop a brief self-rating scale for assessing cancer-related fatigue and to validate the new scale using the responses of cancer patients. Factor analysis was used to evaluate the structure and content of the new scale.

      Methods

      Subjects

      The subjects in the present study were outpatients and inpatients with cancer at the National Cancer Center Hospital East, Japan. Eligibility criteria included (a) 18 years or older, (b) informed of their cancer diagnosis, (c) able to complete the questionnaire and participate in a brief interview, and (d) not suffering from a severe mental or cognitive disorder.
      The study was approved by the Institutional Review Board and the Ethics Committee of the National Cancer Center. Written consent was obtained after each patient had been fully informed of the purpose of the study.

      Overview of the Study Procedure

      This study consisted of two phases: scale development and validation. The scale development phase involved the development of a draft scale, completion of the draft questionnaire by cancer patients, and the development of the final Cancer Fatigue Scale (CFS) based on factor analysis. The validation study was conducted in another sample of cancer patients. The feasibility, validity, and reliability of this scale were investigated by evaluating (a) the time required to complete the CFS, (b) construct validity, (c) convergent validity, (d) association with other measures, (e) internal consistency, (f) test–retest reliability, and (g) reliability of telephone use.

      Scale Development Phase

      Our first task in developing the fatigue scale was to collect and generate items that could be used to evaluate the nature of cancer-related fatigue. In-depth interviews about fatigue were conducted with cancer patients in a clinical situation. Discussions were held with 15 medical experts (6 psychiatrists and 9 oncologists, including 4 engaged in palliative care units with at least 3 years of clinical experience) to generate possible items. We also reviewed published papers about fatigue, and some items selected were based on these papers.
      • Glaus A.
      • Crow R.
      • Hammond S.
      A qualitative study to explore the concept of fatigue/tiredness in cancer patients and in healthy individuals.
      • Yellen S.B.
      • Cella D.F.
      • Webster K.
      • et al.
      Measuring fatigue and other anemia-related symptoms with the functional assessment of cancer therapy (FACT) measurement system.
      • Yoshitake H.
      Relations between the symptoms and the feeling of fatigue.
      • McNair D.M.
      • Lorr M.
      • Droppleman L.F.
      • Blesch K.S.
      • Paise J.A.
      • Wickham R.
      • et al.
      Correlates of fatigue in people with breast or lung cancer.
      • Lee K.A.
      • Hicks G.
      • Nino-Murcia G.
      Validity and reliability of a scale to assess fatigue.
      • Chalder T.
      • Berelowitz G.
      • Pawlikowska T.
      • et al.
      Development of fatigue scale.
      • Morant R.
      • Stiefel F.
      • Berchtold W.
      • et al.
      Preliminary results of a study assessing asthenia and related psychological and biological phenomena in patients with advanced cancer.
      We then drafted the scale using ratings of 1 (not at all) to 5 (very much). Cancer patients were asked to complete the draft questionnaire along with a visual analogue scale (VAS) for fatigue. Eighty-nine nurses working at National Cancer Hospital East also were asked to complete the draft questionnaire and to note their impressions and opinions. Based on this survey, items that fell into any of the following categories were eliminated: (1) items that many subjects asked to have explained to them further during completion of the scale; (2) items for which the response distribution was severely skewed; (3) items with weak correlations (Pearson's correlation coefficient < 0.40) with the VAS for fatigue; (4) items related to other physical symptoms, because contamination by the other somatic symptoms should be avoided; and (5) items whose meaning was similar to that of another item.
      Next, the draft questionnaire responses were analyzed by factor analysis followed by varimax rotation. The number of factors was identified by the Scree test. Items with low factor loadings were deleted from each subscale.
      The translation of the complete scale into English was performed using the iterative forward–backward translation sequence. We chose the translation methodology developed by Bonomi et al.
      • Bonomi A.E.
      • Cella D.F.
      • Hahn E.A.
      • et al.
      Multilingual translation of the Functional Assessment of Cancer Therapy (FACT) quality of life measurement system.
      for the Functional Assessment of Cancer Therapy quality-of-life measurement system because they reported excellent feasibility for this translation method.

      Validation Phase

      Measures

      Visual Analogue Scale (VAS) for Fatigue. To investigate the convergent validity of the CFS, we asked patients to express the intensity of the fatigue they experienced at the time on a 100-mm VAS, ranging from not fatigued at all to extremely fatigued, thus expressing a global sensation of fatigue.
      Hospital Anxiety and Depression Scale (HADS). The Hospital Anxiety and Depression Scale (HADS) was used to investigate the association between negative emotions and the CFS. This questionnaire, developed by Zigmond and Snaith,
      • Zigmond A.S.
      • Snaith R.P.
      The Hospital Anxiety and Depression Scale.
      evaluates symptoms of mood disturbance over the preceding week. It is made up of a 7-item anxiety subscale and a 7-item depression subscale. An important characteristic of this scale is that items referring to symptoms that may have a physical cause are not included. We have established the reliability and validity of the Japanese version of this questionnaire for cancer patients.
      • Kugaya A.
      • Akechi T.
      • Okuyama T.
      • et al.
      Screening for psychological distress in Japanese cancer patients.
      Mini-Mental State (MMS). The MMS examination was used to investigate the association between cognitive function and the CFS. This examination, developed by Folstein et al.,
      • Folstein M.F.
      • Folstein S.E.
      • McHugh P.R.
      Mini-Mental State a practical method for grading the cognitive state of patients for the clinician.
      evaluates cognitive performance. It consists of 11 items and is brief enough to be used for cancer patients. The reliability and validity of the Japanese version of this examination have been established.
      • Mori E.
      • Mitani Y.
      • Yamadori A.
      Usefulness of a Japanese version of the Mini-Mental State Test in neurological patients.
      Cancer patients who had not participated in the scale development phase were asked to complete the CFS. Some of the patients were asked to complete the HADS and/or MMS concurrently. HADS and MMS were selected for comparison with the CFS based on the results of our factor analysis of the responses to the draft questionnaire. Medical data, including the specific cancer diagnosis and the patients' Performance Status as defined by the Eastern Cooperative Oncology Group (ECOG), were recorded by physicians. Sociodemographic data for each patient were obtained in an interview.

      Test–Retest Reliability and Telephone Usage

      Ambulatory breast cancer patients who were not receiving active cancer treatment were recruited in a study to assess the test–retest reliability of the new scale. They took this survey by mail with an interval of about 8 days. In addition, ambulatory cancer patients from the Palliative Care Unit were asked to complete the questionnaire on two occasions to investigate the possibility of using the CFS over the telephone. On the first occasion, they completed the questionnaire in the outpatient clinic of the Palliative Care Unit, and on the second occasion, they were asked all of the questions in the CFS over the telephone within about 3 days of the written assessment. On the former, occasion the time required to complete the CFS was measured.

      Statistical Methods

      The CFS was developed by factor analysis of the draft items followed by varimax rotation. This method was used to explore the underlying structure that adequately explains the overall observed variation and to reduce the complexity of the data. The number of factors was identified by the Scree test. Construct validity was evaluated based on whether repeating factor analysis reproduced the factor loading pattern that was seen in the scale development phase. Intersubscale correlations were evaluated by calculating Pearson's correlations. Convergent validity was evaluated by calculating Pearson's correlations between the CFS scores and the VAS scores. The associations between the CFS and the ECOG Performance Status, HADS, and MMS were also assessed by calculating Pearson's correlation coefficients. Reliability was evaluated by calculating Cronbach's alpha coefficient, a measure of the internal consistency of responses to a group of items. Test–retest reliability was assessed by Pearson's correlations comparing the scores reported on two separate occasions.
      All statistical procedures were performed using the SAS statistical software package (SAS Institute Inc., 1997).

      Results

      Scale Development Phase

      We developed a 58-item draft questionnaire. One hundred seven cancer patients completed this draft questionnaire and the VAS for fatigue at the same time. The average age of the participants was 61 years (range, 35–84 years). The most frequent cancer site was the lung (26.2%), followed by the colon (18.7%). Of the participants, 40.2% had physical impairments with a Performance Status of 1 or greater.
      Based on the responses to the draft, 31 items were eliminated. Then, the responses to the remaining 27 items were assessed by factor analysis followed by varimax rotation. Three factors were identified by the Scree test. Twelve items for which the factor loading was less than 0.65 were deleted from each subscale. The first subscale then consisted of 7 items that were general statements or were related to the physical sensations of fatigue; we called it the physical subscale. The second subscale consisted of 4 items that assessed affective activity, and we called it the affective subscale. The third subscale consisted of 4 items that assessed attention and memory, and we called it the cognitive subscale. Cronbach's alpha coefficients for this patient sample were 0.90, 0.78, 0.79, and 0.88 for the physical subscale, affective subscale, cognitive subscale, and total scale, respectively. The CFS has a maximum scores of 28, 16, and 16 for the physical, affective, and cognitive subscales, respectively. The possible response range for the total score is from 0 to 60. Higher scores reveal more severe fatigue.
      An English version of the complete scale is presented in Appendix 1.

      Validation Phase

      Subjects

      Another 307 cancer patients who had not participated in the scale development phase participated in the validation phase (Table 1). There were many more women than men (64.5%). The most frequent cancer site was the breast (50.2%), followed by the lung (31.9%). Forty-six percent had physical impairments with Performance Status scores of 1 or more. Of the 307 patients, 112 randomly selected breast cancer patients (118 recruited, 2 refused, and 4 were ineligible) participated in the test–retest study, and 30 Palliative Care Unit patients (39 recruited, 9 were ineligible) participated in the telephone survey.
      Table 1Demographic and Clinical Characteristics of the Subjects in Validation Phase (N = 307)
      N%
      Gender
      Male10935.5
      Female19864.5
      Age
      Mean58.1 years
      Median58 years (range, 28- 86 years)
      Cancer site
      Breast15450.2
      Lung9831.9
      Other5517.9
      Clinical stage (No prior treatment)
      041.3
      I5016.3
      II8126.4
      III6721.8
      IV8226.7
      Recurrent case227.2
      No staging10.3
      Performance Status (ECOG)
      016654.1
      110032.6
      2299.4
      3103.3
      420.7
      ECOG, Eastern Cooperative Oncology Group.

      Feasibility

      All of the patients easily completed this scale. The average time required by 30 Palliative Care Unit outpatients to complete the questionnaire was 132.9 seconds (SD = 77.2, median = 111.5 seconds).

      Validity

      Construct Validity (Table 2). The Scree test demonstrated the 3-factor solution again. The factor analysis represented the same factor-loading pattern as in the development phase.
      Table 2Factor Loading Pattern (Followed by Varimax Rotation) of the Cancer Fatigue Scale in Validation Phase (N = 307)
      Item number and contentFactor
      Physical
      Physical aspect of fatigue.
      Cognitive
      Cognitive aspect of fatigue.
      Affective
      Affective aspect of fatigue.
      3. exhausted0.860.14−0.11
      6. heavy and tired0.850.18−0.08
      15. fatigue you don't know what to do with yourself0.780.14−0.10
      1. easily tired0.780.17−0.07
      2. having urge to lie down0.770.09−0.16
      12. reluctant0.550.43−0.20
      9. fed-up0.540.36−0.21
      10. forgetful0.100.820.03
      7. errors while speaking0.180.80−0.02
      13. thinking has become slower0.170.78−0.23
      4. careless0.420.55−0.12
      8. interest in something−0.03−0.070.81
      14. encourage yourself to do something−0.07−0.010.79
      5. energetic feeling−0.30−0.030.74
      11. ability to concentrate−0.13−0.180.73
      a Physical aspect of fatigue.
      b Cognitive aspect of fatigue.
      c Affective aspect of fatigue.
      Descriptive Statistics (Table 3). Table 3 shows the mean and standard deviations for each subscale and for the total scale score.
      Table 3Descriptive Data and Reliability of the Cancer Fatigue Scale in Validation Phase
      SubscaleMean (N = 307)s.d.Cronbach's alpha coefficient (N = 307)Test–retest reliability (correlation coefficient) (N = 112)Reliability of telephone use (correlation coefficient) (N = 30)
      Physical6.65.60.890.78
      p < 0.001.
      0.50
      p < 0.01.
      Affective7.63.40.790.52
      p < 0.001.
      0.62
      p < 0.01.
      Cognitive4.03.10.790.77
      p < 0.001.
      0.52
      p < 0.01.
      Total scale18.19.40.880.80
      p < 0.001.
      0.67
      p < 0.01.
      a p < 0.001.
      b p < 0.01.
      Intersubscale Correlation (Table 4). Correlations between the subscales were investigated to evaluate the relationships among the factors. There were significant correlations for all pairs of subscales. The mean value of the intersubscale correlation coefficient was 0.37.
      Table 4Intersubscale Correlations Between the Separate Cancer Fatigue Scale Factors in Validation Phase (N = 307)
      SubscalePhysicalAffectiveCognitive
      Physical
      Affective0.34
      p < 0.001.
      Cognitive0.51
      p < 0.001.
      0.25
      p < 0.001.
      Total scale0.89
      p < 0.001.
      0.65
      p < 0.001.
      0.72
      p < 0.001.
      a p < 0.001.
      Convergent Validity (Table 5). Of 307 patients, 225 were asked to complete the VAS at the same time as the CFS. Convergent validity of the CFS was demonstrated by Pearson's correlations with VAS scores. Each factor significantly correlated with the VAS scores (average r = 0.49, P < 0.001).
      Table 5Correlations between the Cancer Fatigue Scale and VAS, PS, HADS, and MMS
      VAS (N = 225)PS (N = 307)HADS (N
      Thirty-two patients asked to complete both HADS and MMS concurrently.
      = 218)
      SubscaleDepressionAnxietyTotalMMS (Na = 65)
      Physical0.70
      < 0.001.
      0.34
      < 0.001.
      0.59
      < 0.001.
      0.52
      < 0.001.
      0.61
      < 0.001.
      0.12
      Affective0.38
      < 0.001.
      0.16
      < 0.001.
      0.54
      < 0.001.
      0.32
      < 0.001.
      0.47
      < 0.001.
      0.05
      Cognitive0.39
      < 0.001.
      0.030.47
      < 0.001.
      0.44
      < 0.001.
      0.50
      < 0.001.
      −0.05
      Total scale0.67
      < 0.001.
      0.28
      < 0.001.
      0.69
      < 0.001.
      0.69
      < 0.001.
      0.69
      < 0.001.
      0.07
      Abbreviations: VAS, Visual Analogue Scale of fatigue; PS, Performance Status as defined by the Eastern Cooperative Oncology Group; HADS, Hospital Anxiety and Depression Scale; MMS, Mini Mental State.
      a Thirty-two patients asked to complete both HADS and MMS concurrently.
      b < 0.001.
      Association Between the Cancer Fatigue Scale and Performance Status, HADS, and MMS (Table 5). Performance Status scores were obtained for all of the 307 patients; 218 and 65 of the patients were also asked to complete HADS and MMS, respectively. The Performance Status score was significantly associated with the physical and affective subscale scores. Negative emotion (HADS) was significantly associated with each subscale; the correlation coefficient of each subscale was higher for depression than for anxiety. The total score of the MMS was not associated with any of the three subscales.

      Reliability (Table 3)

      Internal Consistency. Cronbach's alpha coefficients, a measure of the internal consistency of the CFS for the entire sample, were 0.89, 0.79, 0.79, and 0.88 for the physical subscale, affective subscale, cognitive subscale, and the total scale, respectively.
      Test–Retest Reliability. In the test–retest study, the mean interval between the first and second sessions was 8.1 days (median = 8 days). The test–retest correlation coefficients of each factor and the total score between the first and second session were all above 0.50 (P < 0.001).
      Reliability of Telephone Use. In the telephone usage study, the mean interval between the written and telephone sessions was 3.1 days (median = 3 days). Correlation coefficients of each factor and total score between the first and second session were above 0.49 (P < 0.01).

      Discussion

      This report describes the development and validation of the Cancer Fatigue Scale, which is composed of 3 factors and 15 items. There are three main advantages of this scale over other fatigue scales. First, this scale can be easily completed in 2 minutes even by advanced cancer patients. Simplicity is an important advantage, because its main purpose is to assess the fatigue experienced by exhausted cancer patients. Second, this scale was specifically designed to reflect the nature of cancer-related fatigue, and to assess the physical, affective, and cognitive aspects of fatigue. Third, this scale showed good validity and reliability in a number of cancer patients sufficient for testing the psychometric properties of this 15-item scale.
      The factor analysis gave the same solution in both the development and validation phases, which confirms the structure of the scale. This analysis also suggested that fatigue consists of three subjective dimensions: the physical, affective, and cognitive aspects of fatigue. This multidimensional concept of fatigue has been described previously by Glaus et al.,
      • Glaus A.
      • Crow R.
      • Hammond S.
      A qualitative study to explore the concept of fatigue/tiredness in cancer patients and in healthy individuals.
      who conducted tape-recorded interviews with cancer patients and healthy individuals, and analyzed the results using content analysis. Although different themes emerged for the two groups, both fit a classification system that categorized fatigue into physical, affective, and cognitive expressions. Similar findings have been reported by Kobashi-Schoot et al.
      • Kobashi-Schoot J.A.M.
      • Hanewald G.J.F.P.
      • Van Dam F.S.A.M.
      • et al.
      Assessment of malaise in cancer patients treated with radiotherapy.
      for cancer patients treated with radiotherapy. The present study supports the growing recognition of the multidimensional nature of fatigue. However, the individuality of the three dimensions involved in the CFS is not completely supported by data for the following three reasons. First, there were significant correlations between the three factors, although ideally subscales should be independent of each other after the orthogonal rotation procedure. Second, some items were loaded for two of the factors, even though factor analysis reproduced the same factor-loading pattern in the development and validation phases. Third, the affective component might have been created by virtue of the positive wording of the question rather than by meanings of the items; responses to negatively versus positively worded questions may draw out patterns that can be misleading.
      Returning to the first point, the strongest correlation between the subscales was that between the physical and cognitive subscales. Thus, the cognitive aspect of fatigue might overlap the physical aspect of fatigue to some extent. However, the mean correlation coefficient of 0.37 for the intersubscale correlation may be acceptable.
      The second point was that some items were loaded for two factors, for example items 12 and 9 might be misplaced because these items were loaded a little heavier on physical subscale than on affective subscale. In particular, item 9, which uses the word “fed-up” (unzari in the Japanese language), covers both affective and physical aspects of fatigue. Although the nuance of this word is delicate (to be exact, it refers to the affective aspect of fatigue rather than physical aspect), this item was loaded not with the items concerned with the affective aspects of fatigue, but with those concerned with the physical aspects of fatigue. We cannot explain this phenomenon, but we have recently confirmed the consistency of this assignment in studies of another two groups of patients, a consecutive set of 158 advanced lung cancer patients and a consecutive set of 139 incurable advanced cancer patients at the Palliative Care Unit (unpublished data; for details contact T.O.). Both sets of data supported the assignment of this factor loading pattern.
      Finally, the third factor, the affective aspect of fatigue, is unlikely to be related to positive wording because several items that had negative wording and were concerned with the affective aspect of fatigue also loaded this factor in the development phase, although these items were deleted because their loadings were less than 0.65. Furthermore, the high Cronbach's alpha coefficient for total scale indicated that the phraseology was not a major problem.
      Convergent validity was shown by the significant correlations between each subscale of the CFS and the VAS of fatigue, which is widely used for assessing fatigue; this result indicated that the three factors we identified, namely physical, affective, and cognitive aspects of fatigue, reflect the patients' perception of fatigue. However, the affective and cognitive subscale scores had slightly weaker correlations with the VAS scores than did the physical subscale score.
      The Performance Status significantly correlated with the physical subscale scores. This means that the physical subscale partially reflects the patient's physical condition and is valid. However, the correlation was lower than we expected. This may have been because the physical performance assessment was objective, whereas physical subscale score was subjective, as discussed by Dimeo et al.
      • Dimeo F.
      • Stieglitz R.D.
      • Novelli-Fischer U.
      • et al.
      Correlation between physical performance and fatigue in cancer patients.
      On the other hand, the Performance Status score did not correlate with the cognitive subscale scores. Thus the cognitive aspect of fatigue may not be associated with the patients' physical condition, but may partly overlap the physical aspect of fatigue.
      The correlations between each subscale of HADS and the CFS physical subscale were about as strong as those between each subscale of HADS and the CFS affective subscale. This result indicates that emotional distress is associated with the physical perception of fatigue by cancer patients. Further research is needed to investigate the discriminative validity of the CFS with regard to affective aspect of fatigue. Anxiety significantly correlated with each CFS subscale and, like depression, may play a role in cancer-related fatigue. Depression and anxiety are two of the most prevalent psychiatric symptoms in cancer patients,
      • Derogatis L.R.
      • Morrow G.R.
      • Fetting J.
      • et al.
      The prevalence of psychiatric disorders among cancer patients.
      and it is often pointed out that they tend to be underestimated by physicians and nursing staff.
      • Massie M.J.
      • Holland J.C.
      Diagnosis and treatment of depression in cancer patient.
      The strong correlation between the CFS total fatigue score and the HADS scores might suggests that the CFS provides clues about psychiatric problems of cancer patients.
      There is no association between the CFS, or even the cognitive subscale, and the MMS score, perhaps because cognitive impairment caused by fatigue is assessed on the CFS as a subjective feeling in everyday experience, whereas the MMS is an objective assessment. Cull et al.
      • Cull A.
      • Stewart M.
      • Altman D.G.
      Assessment and intervention for psychological problems in routine oncology practice.
      reported that 49% of their mixed cancer patients complained of cognitive problems. Furthermore, they suggested that standard neuropsychometric tests evaluating concentration and memory may not show significant differences between patients who complain of fatigue and those who do not complain.
      • Cull A.
      • Hay C.
      • Love S.B.
      • et al.
      What do cancer patients mean when they complain of concentration and memory problems?.
      Our results seem to support their findings. Further research is required to investigate the subjective feeling of the cognitive aspect of fatigue.
      Each of the three subscales and the total scale of the CFS demonstrated high internal consistency, and the strong correlation in the test–retest reliability data suggests that each subscale and total scale provides good reliability. The written and telephone survey responses did not correlate as strongly as we had expected. Two possible explanations for this are researcher bias, which may have changed the results even though the telephone interview was conducted using a structured interview format, and the influence of the interview environment, in the hospital and at home. Although the feasibility of using the telephone interview routinely is not excellent, we think that these results indicated that the telephone use of this scale may facilitate some types of research on fatigue.
      This study had some limitations. First, patient populations were large but heterogeneous. Second, we did not investigate the cross-cultural validity of the CFS, even though we expect that there are differences in the cultural experience of fatigue. Further research is recommended on this topic.
      In conclusion, the findings of the present study indicate that the CFS is a brief, valid, and feasible scale for assessing the three dimensions of fatigue experienced by cancer patients. Although fatigue is recognized as a highly prevalent symptom in cancer patients, adequate palliative strategies for dealing with cancer-related fatigue have not yet been established. The lack of a suitable tool for assessing fatigue has been a barrier to progress in the research and clinical treatment of fatigue. This new scale will contribute to a better conceptual understanding of fatigue and to the establishment of a therapeutic strategy for treating fatigue in cancer patients.2

      Acknowledgements

      This work was supported in part by Grants-in-Aid for Cancer Research (9-31, 9-32) and the Second Term Comprehensive 10-Year Strategy for Cancer Control from the Ministry of Health and Welfare, Japan.
      Toru Okuyama is an Awardee of a Research Resident Fellowship from the Foundation for the Promotion of Cancer Research, Japan.
      We gratefully acknowledge the assistance of Dr. Satoshi Sasaki, MD, PhD, of the Epidemiology and Biostatistics Division, National Cancer Center Research Institute East, Japan for supervising the statistical analysis. We also thank the staff at the National Cancer Center Hospital East, Japan especially Shigeru Imoto, MD, PhD, of the Division of Breast Surgery, as well as Yurie Sugihara, BA, Yuko Kojima, RN, Kumiko Harada, RN, and Ms. Miho Sakai, of the Psycho-Oncology Division, National Cancer Research Institute East, Japan, for their research assistance.

      Appendix 1.

      The Cancer Fatigue Scale

      This questionnaire will ask you about any sense of fatigue you might be experiencing. For each question, please circle only one number you think most aptly describes your current state. Try to answer on the basis of first impressions, without thinking deeply about each question.
      Right now … No A little Somewhat Considerably Very much
      1 Do you become tired easily? 1 2 3 4 5
      2 Do you have the urge to lie down? 1 2 3 4 5
      3 Do you feel exhausted? 1 2 3 4 5
      4 Do you feel you have become careless? 1 2 3 4 5
      5 Do you feel energetic? 1 2 3 4 5
      6 Does your body felt heavy and tired? 1 2 3 4 5
      7 Do you feel that you more often make errors while speaking? 1 2 3 4 5
      8 Do you feel interest in anything? 1 2 3 4 5
      9 Do you feel fed-up? 1 2 3 4 5
      10 Do you feel you have become forgetful? 1 2 3 4 5
      11 Can you concentrate on certain things? 1 2 3 4 5
      12 Do you feel reluctant? 1 2 3 4 5
      13 Do you feel that your thinking has become slower? 1 2 3 4 5
      14 Can you encourage yourself to do anything? 1 2 3 4 5
      15 Do you feel such fatigue that you don't know what to do with yourself? 1 2 3 4 5

      Appendix 2.

      The Calculation Method

      Add the number together in every factor
      Factor 1 = (items 1 + 2 + 3 + 6 + 9 + 12 + 15) − 7 P. (Physical subscale)
      Factor 2 = 20 − (items 5 + 8 + 11 + 14) P. (Affective subscale)
      Factor 3 = (items 4 + 7 + 10 + 13) − 4 P. (Cognitive subscale)
      Add the factors together P. (Total scale score)
      *Subtractions adjust for 0 as a state of no fatigue.

      References

        • Coyle N.
        • Adelhardt J.
        • Foley K.M.
        • Portenoy R.K.
        Character of terminal illness in the advanced cancer patient.
        J Pain Symptom Manage. 1991; 5: 83-93
      1. Vainio A, Auvinen A, with Members of the Symptom Prevalence Group. Prevalence of symptoms among patients with advanced cancer: an international collaborative study. J Pain Symptom Manage 1996;12:3–10.

        • Sureshkumar K.
        • Rajagopal M.R.
        Palliative care in Kerala.
        Palliat Med. 1996; 10: 293-298
        • Nerentz D.R.
        • Leventhal H.
        • Love R.R.
        Factors contributing to emotional distress during cancer chemotherapy.
        Cancer. 1982; 50: 1020-1027
        • Love R.R.
        • Leventhal H.
        • Easterling D.V.
        • et al.
        Side effects and emotional distress during chemotherapy.
        Cancer. 1983; 63: 604-612
        • Cassileth B.R.
        • Farber J.M.
        • Lusk E.J.
        • et al.
        Chemotherapeutic toxicity.
        Am J Clin Oncol. 1985; 8: 419-425
        • Greenberg D.B.
        • Sawicha J.
        • Eisenthal S.
        • et al.
        Fatigue syndrome due to localized radiation.
        J Pain Symptom Manage. 1992; 7: 38-45
        • Hickok J.T.
        • Morrow G.R.
        • McDonald S.
        • et al.
        Frequency and correlates of fatigue in lung cancer patients receiving radiation therapy.
        J Pain Symptom Manage. 1996; 11: 370-377
        • Smets E.M.A.
        • Garssen B.
        • Cull A.
        • et al.
        Application of the Multidimensional Fatigue Inventory (MFI-20) in cancer patients receiving radiotherapy.
        Br J Cancer. 1996; 73: 241-245
        • Fobair P.
        • Hoppe R.T.
        • Bloom J.
        • et al.
        Psychosocial problems among survivors of Hodgkin's disease.
        J Clin Oncol. 1986; 4: 805-814
        • Devlen J.
        • Maguire P.
        • Phillips P.
        • et al.
        Psychological problems associated with diagnosis and treatment of lymphomas. 1.
        BMJ. 1987; 295: 953-957
        • Berglund G.
        • Bolund C.
        • Fornander T.
        • et al.
        Late effects of adjuvant chemotherapy and postoperative radiotherapy on quality of life among breast cancer patients.
        Eur J Cancer. 1991; 27: 1075-1081
        • Joly F.
        • Henry-Amar M.
        • Arveux P.
        • et al.
        Late psychosocial sequelae in Hodgkin's disease survivors.
        J Clin Oncol. 1996; 14: 2444-2453
        • Rhodes V.A.
        • Watson P.M.
        • Hanson B.M.
        Patients' descriptions of the influence of tiredness and weakness on self-care abilities.
        Cancer Nurs. 1988; 11: 186-194
        • Breitbart W.
        Suicide in cancer patients.
        Oncology. 1987; 1: 49-53
        • Holley S.P.
        Fatigue in cancer patients.
        Cancer Nurs. 1991; 14: 13-19
        • Akechi T.
        • Kugaya A.
        • Okamura H.
        • et al.
        Fatigue and its associated factors in ambulatory cancer patients.
        J Pain Symptom Manage. 1999; 17: 42-48
        • Bruera E.
        • MacDonald R.N.
        Asthenia in patients with advanced cancer.
        J Pain Symptom Manage. 1988; 3: 9-14
        • Smets E.M.A.
        • Garssen B.
        • Schuster-Uitterhoeve A.L.J.
        • et al.
        Fatigue in cancer patients.
        Br J Cancer. 1993; 68: 220-224
        • Smets E.M.A.
        • Garssen B.
        • Bokne B.
        • et al.
        The Multidimensional Fatigue Inventory (MFI) psychometric qualities of an instrument to assess fatigue.
        J Psychosom Res. 1995; 39: 315-325
        • Glaus A.
        • Crow R.
        • Hammond S.
        A qualitative study to explore the concept of fatigue/tiredness in cancer patients and in healthy individuals.
        Support Care Cancer. 1996; 4: 82-96
        • Hann D.M.
        • Jacobsen P.B.
        • Martin S.C.
        • et al.
        Fatigue in women treated with bone marrow transplantation for breast cancer.
        Support Care Cancer. 1997; 5: 44-52
        • Aaronsen N.
        • Ahmedzai S.
        • Bergman B.
        • et al.
        The European Organization for Research and Treatment of Cancer QLQ-C30.
        J Natl Cancer Inst. 1993; 85: 365-376
      2. Rhoten D. Fatigue and the postsurgical patient. Concept Clarification in Nursing. Rockville, MD: Aspen Publishers, 1982:277–300.

        • Piper B.F.
        • Linsey A.M.
        • Dodd M.J.
        • et al.
        The development of an instrument to measure the subjective dimension of fatigue.
        in: Funk S.G. Tornquist E.M. Campagne M.T. Key aspects of comfort. Management of pain fatigue and nausea. Springer, New York1989
        • Richardson A.
        Measuring fatigue in patients with cancer.
        Support Care Cancer. 1998; 6: 94-100
        • Yellen S.B.
        • Cella D.F.
        • Webster K.
        • et al.
        Measuring fatigue and other anemia-related symptoms with the functional assessment of cancer therapy (FACT) measurement system.
        J Pain Symptom Manage. 1997; 13: 63-74
        • Yoshitake H.
        Relations between the symptoms and the feeling of fatigue.
        Ergonomics. 1971; 14: 175-186
        • McNair D.M.
        • Lorr M.
        • Droppleman L.F.
        Profile of Mood States. Educational and Industrial Testing Service, San Diego1971
        • Blesch K.S.
        • Paise J.A.
        • Wickham R.
        • et al.
        Correlates of fatigue in people with breast or lung cancer.
        Oncol Nurs Forum. 1991; 18: 81-87
        • Lee K.A.
        • Hicks G.
        • Nino-Murcia G.
        Validity and reliability of a scale to assess fatigue.
        Psychiatry Res. 1991; 36: 291-298
        • Chalder T.
        • Berelowitz G.
        • Pawlikowska T.
        • et al.
        Development of fatigue scale.
        J Psychosom Res. 1993; 37: 147-153
        • Morant R.
        • Stiefel F.
        • Berchtold W.
        • et al.
        Preliminary results of a study assessing asthenia and related psychological and biological phenomena in patients with advanced cancer.
        Support Care Cancer. 1993; 1: 101-117
        • Bonomi A.E.
        • Cella D.F.
        • Hahn E.A.
        • et al.
        Multilingual translation of the Functional Assessment of Cancer Therapy (FACT) quality of life measurement system.
        Quality of Life Res. 1996; 5: 309-320
        • Zigmond A.S.
        • Snaith R.P.
        The Hospital Anxiety and Depression Scale.
        Acta Psychiat Scand. 1983; 67: 361-370
        • Kugaya A.
        • Akechi T.
        • Okuyama T.
        • et al.
        Screening for psychological distress in Japanese cancer patients.
        Jpn J Clin Oncol. 1998; 28: 333-338
        • Folstein M.F.
        • Folstein S.E.
        • McHugh P.R.
        Mini-Mental State.
        J Psychiat Res. 1975; 12: 189-198
        • Mori E.
        • Mitani Y.
        • Yamadori A.
        Usefulness of a Japanese version of the Mini-Mental State Test in neurological patients.
        Jpn J Neuropsychol. 1985; 1: 82-90
        • Kobashi-Schoot J.A.M.
        • Hanewald G.J.F.P.
        • Van Dam F.S.A.M.
        • et al.
        Assessment of malaise in cancer patients treated with radiotherapy.
        Cancer Nurs. 1985; 8: 306-313
        • Dimeo F.
        • Stieglitz R.D.
        • Novelli-Fischer U.
        • et al.
        Correlation between physical performance and fatigue in cancer patients.
        Ann Oncol. 1997; 8: 1251-1255
        • Derogatis L.R.
        • Morrow G.R.
        • Fetting J.
        • et al.
        The prevalence of psychiatric disorders among cancer patients.
        JAMA. 1983; 249: 751-757
        • Massie M.J.
        • Holland J.C.
        Diagnosis and treatment of depression in cancer patient.
        J Clin Psychiat. 1984; 45: 25-28
        • Cull A.
        • Stewart M.
        • Altman D.G.
        Assessment and intervention for psychological problems in routine oncology practice.
        Br J Cancer. 1995; 72: 229-235
        • Cull A.
        • Hay C.
        • Love S.B.
        • et al.
        What do cancer patients mean when they complain of concentration and memory problems?.
        Br J Cancer. 1996; 74: 1674-1679