European Journal of Oncology Nursing 19 (2015) 427e432
Contents lists available at ScienceDirect
European Journal of Oncology Nursing journal homepage: www.elsevier.com/locate/ejon
Development of the Hirai Cancer Fatigue Scale: Testing its reliability and validity Kazue Hirai a, *, Kiyoko Kanda b, Junko Takagai c, Mai Hosokawa d a
Faculty of Medicine, School of Nursing, Tokyo Medical University, 6-1-1 Shinjuku, Shinjuku-ku, Tokyo, 160-8402, Japan Department of Nursing, Gunma University Graduate School of Health Science, 3-29-21 Showa-machi, Maebashi, Gunma, 371-8514, Japan c Faculty of Medicine, School of Health Science, Akita University, 1-1-1 Hondou, Akita, Akita, 010-8543, Japan d The Jikei University School of Nursing, 8-3-1, Kokuryocho, Chofu, Tokyo, 182-8570, Japan b
a b s t r a c t Keywords: Fatigue Cancer Measurement Symptom assessment Quality of life
Purpose: The purpose of this study was to develop the Hirai Cancer Fatigue Scale (HCFS) to assess the fatigue experienced by cancer patients, and to verify its reliability and validity. Methods: Based on qualitative research about the perception of fatigue by Japanese cancer patients, we developed a questionnaire. The content validity was confirmed by 5 expert oncology nurses and 5 oncologists. 281 Japanese cancer patients participated in this study. Construct validity was analyzed using factor analysis, and internal consistency was analyzed using Cronbach's a coefficient. Results: A 15-item scale with 3 dimensions, “physical/mental sensation”, “activity-related sensation” and “cognitive sensation,” was developed by factor analysis. This scale had an overall Cronbach's a coefficient of .943 and a test-retest reliability coefficient of r ¼ .820 (p < 0.01), confirming the high reliability of the scale. The correlation coefficient was r ¼ .759 (p < 0.01) between HCFS and abridged Profile of Mood States-Fatigue (POMS-F), and r ¼ .763 (p < 0.01) between HCFS and Cancer Fatigue Scale (CFS), both showing high correlations and confirming criterion-related validity. Conclusion: HCFS enables reliable and valid evaluation of Japanese cancer patients' fatigue. Use of the HCFS would assist in convenient self-evaluation of fatigue, and would allow information to be effectively provided to healthcare professionals. It could also be used for outcome evaluation in an intervention study. © 2014 Elsevier Ltd. All rights reserved.
Introduction Fatigue is the most common symptom among cancer patients (Lawrence et al., 2004; Prue et al., 2006), and is known to be one of the main symptoms of pain. It exists in all stages of cancer survivorship, from diagnosis to end of life; some people have already experienced it at the time of diagnosis of cancer (Vogelzang et al., 1997), and it is an extremely common adverse effect during treatments such as chemotherapy (Butt et al., 2008; Siefert, 2010) and radiation therapy (Stricker et al., 2004). It persists long after completing a standard course of treatment (Kutner et al., 2001),
* Corresponding author. Tel.: þ81 3 5357 7304; fax: þ81 3 3351 3693. E-mail addresses: [email protected] (K. Hirai), [email protected] (K. Kanda), [email protected] (J. Takagai), [email protected] (M. Hosokawa). http://dx.doi.org/10.1016/j.ejon.2014.12.004 1462-3889/© 2014 Elsevier Ltd. All rights reserved.
with almost all patients experiencing fatigue in the terminal stage (Fu et al., 2005). Therefore, properly managing fatigue is a significant issue that must be addressed in order for cancer survivors to maintain an adequate quality of life (QOL). The number one cause of death in Japan, the world's leader in longevity, is cancer, and 1 in 2 Japanese people suffer from cancer at some point in life (Cancer Statistics in Japan, 2013). Until about 15 years ago, it was usual for Japanese cancer patients to be hospitalized to complete chemotherapy or radiotherapy. Recently, outpatient cancer treatments have become common, the duration of hospitalization for all kinds of cancer treatment have been shortened, and services that allow cancer patients to spend their terminal stages at home have been promoted. However, fatigue in cancer patients is still rarely a focus in either clinical settings or in research fields in Japan. Outpatient nurses can play an important role in the development of a plan to properly manage fatigue that occurs frequently and diminishes the QOL of cancer patients. To
428
K. Hirai et al. / European Journal of Oncology Nursing 19 (2015) 427e432
accomplish this, a scale with high reliability and high validity to properly evaluate fatigue, and that is easy enough for use at an outpatient level, is necessary. Despite the various definitions of fatigue that have been tested to date, a universal definition does not exist. As the consensus is that fatigue is a subjective and multidimensional symptom (Rhodes et al., 1995; McDaniel and Rhodes, 1998; Fu et al., 2002), fatigue scales require the capability of multidimensional evaluation. Currently, the fatigue scales for cancer patients that can be used in Japanese are: revised Piper Fatigue Scale (revised PFS) (Piper et al., 1998), Cancer Fatigue Scale (CFS) (Okuyama et al., 2000), Brief Fatigue Inventory (BFI) (Mendoza et al., 1999), Profile of Mood StatesFatigue (POMS-F) (McNair et al., 1992), and Functional Assessment of Cancer Therapy (FACT) (Yellen, 1997). However, of these scales, only the revised PFS and CFS are multidimensional scales. The revised PFS, the first multidimensional fatigue scale, is a well-designed scale based on a theoretical model that makes 4dimensional measurements of fatigue: severity/behavioral, sensory, affective/meaning, and cognitive/mood. It is a significantly improved version of the original 82-item version with complex wording (Piper et al., 1989). However, as Schwartz (1998) pointed out, there are too many question items, 22 in total, which require a long time to answer, thus possibly becoming a burden on the respondents. In Japan, Kamizato (1999) announced a Japanese version of Piper's 22-item scale and used it in a study to elucidate fatigue in cancer patients undergoing radiation therapy; however, since then it has been little used in either research or clinical settings. CFS is the only multidimensional fatigue scale developed by Japanese researchers that measures fatigue in 3-dimensions: physical, affective and cognitive. This scale is often used in nursing research in Japan because it has 15 items, making it considerably shorter than the revised PFS, and thus less of a burden on patients. However, it is far from convenient, as complicated calculations for each dimension are required for its evaluation. For this reason, this scale has not been used very often in clinical practice to quickly assess outpatient fatigue. In addition, it includes items such as “Do you feel that you make errors more often while speaking?”, which is an expression that is unfamiliar and not normally used by Japanese people when they talk about fatigue. Given the limitations of the previous fatigue scales, we believe that development of a multidimensional fatigue scale based on how Japanese cancer patients express their perception of fatigue, and one that is easier to use, is much required. Purpose The purpose of this study was to develop the Hirai Cancer Fatigue Scale (HCFS) that assesses fatigue experienced by cancer patients, and to verify its reliability and validity.
symptoms' frequency, intensity, distress, and meaning, and that there are individual differences in perception and expression. The same model also states that the expression or consequences of the symptoms are described as changes in functional and cognitive activities, including adjustment to illness and quality of life. Based on these models, we set up the conceptual framework of this study, as described below. Fatigue experienced by cancer patients is a series of dynamic processes, where “perception of fatigue”, “coping”, and “outcome” influence each other. It is also influenced by demographic characteristics (sex and age), disease characteristics (diagnosis, stage of disease, treatment and health status), and personal characteristics (physical, psychological, social and cultural factors) (Fig.1). Based on this conceptual framework, we developed a scale that measures the perception of fatigue in this study. Development stage Preliminary investigation We first elucidated how Japanese cancer patients express their perception of fatigue, and, with the objective of conceptualizing this, we conducted an open-ended questionnaire survey that asked, “How would you describe your sensation of fatigue?” among 400 inpatient or outpatient Japanese cancer patients at 2 national hospitals in eastern Japan (Hirai et al., 2014). The survey was reviewed and approved by the respective ethical review boards of each hospital. Content analysis showed that 237 codes of fatigue were obtained and 35 subcategories/15 categories/4 corecategories were extracted. The 4 core-categories were “physical sensation”, “mental sensation”, “cognitive sensation”, and “sensation that cannot be described with words”. Physical sensations consisted of “My body feels heavy”, “I am tired/feel listless or sluggish/don't feel quite right”, “I want to lie down/sit down”, “I am exhausted/can't get comfortable”, “My body doesn't do what I want it to do”, “I am sleepy”, and “I lack endurance”. Mental sensations consisted of “I have no willpower/motivation”, “Everything feels bothersome”, “I don't want to do anything/I don't want to move”, “I am depressed/anxious”, and “I am frustrated”. Cognitive sensations consisted of “reduction in ability to think” and “reduction in concentration”. In addition, indescribable sensations were indicated by “cannot describe in words”. Furthermore, we performed a literature review of fatigue and fatigue scales for cancer patients, which aimed to establish the definition and the conceptual framework and to understand the characteristics of each scale. The review revealed that the multidimensionality of fatigue was mostly described in physical, mental and cognitive aspects. It also revealed that existing scales have not necessarily been developed based on the definition of fatigue. The
Methods Conceptual framework According to the Revised Symptom Management Conceptual Model (Dodd et al., 2001), symptom management consists of three interrelated concepts: “symptom experience”, “symptom management strategy”, and “outcomes”. According to this model, symptom management is a dynamic process that is modified by individual outcomes and the influences of the nursing domains of person, health/illness, or environment. Additionally, according to the Symptoms Experience Model (Armstrong, 2003), various elements of individuals become antecedents of symptom experience and act on symptom production. This model states that the perceived components of the symptom experience include the
Fig. 1. Conceptual framework.
K. Hirai et al. / European Journal of Oncology Nursing 19 (2015) 427e432
results of the literature review contributed to establishment of the basis of the present study and to develop a draft questionnaire. Theoretical establishment of construct and subscale From the results of the questionnaire survey and literature review, we defined the perception of fatigue by cancer patients as “a sensation characterized by the decline of functional status and the discomfort associated with lack of energy”. We also theoretically established the construct as: 1) Physical sensation: Unpleasant sensation perceived physically, reduction in physical function and a sense of loss of physical control; 2) Mental sensation: Sensation characterized by depression and decline in motivation and willpower; and 3) Cognitive sensation: Sensation characterized by diminished ability to think and pay attention, and by reduction in memory. Creation of the first draft of the questionnaire Each of the three constructs included codes obtained through preliminary investigation of the item pool. Questionnaire items were also added by brainstorming among the authors, and by searching the reports identified in the literature review for the keywords “fatigue scale/measurement” and “fatigue”. Assessments and modifications were repeated among the researchers to avoid semantic duplication and to improve clarity of expression for each item, and a 49-item questionnaire was created. The questions focused on the “current condition”, with the instructions to participants stating, “This is a questionnaire about your fatigue. Please circle the number that most accurately describes your current condition”. The evaluation method used a 5-point Likert scale (1: Does not apply at all/Not at all, 2: Slightly applicable, 3: Somewhat applicable, 4: Very applicable, 5: Extremely applicable). This became the first draft of the HCFS. Assessment of content validity We assessed content validity by requesting advice related to the suitability of the HCFS first draft from 10 experts (5 oncologists and 5 expert oncology nurses). We asked these experts to evaluate each of the 49 items on the following three points: 1) whether the question expresses perception of fatigue, 2) whether the question expresses different aspects of said perception, including physical, mental and cognitive perceptions, and 3) whether the question is easy to understand in Japanese and easy to answer. There were 4 items (“I am distracted”, “I am frustrated”, “I am overwhelmed”, and “I am fed up”) that were removed either because the experts' opinions were divided on these points or because the item was determined unsuitable, resulting in a final total of 45 items in the HCFS. This was the preliminary version of the HCFS. Validation stage Sample Participants included outpatients and inpatients with various types of cancer, who were receiving chemotherapy and/or radiotherapy and/or hormone therapy or were under observation after completion of their treatment, at 4 large-scale public hospitals in eastern Japan. Patients eligible for participation in this study included (a) those who were aware of their diagnosis of cancer, (b) were in their 20e70's, and (c) were not suffering from mental or cognitive disorders. The nursing administrators of each outpatient department and ward selected the patients that met the eligibility criteria (except for immediately after surgery). Thereafter, the researchers were introduced to the patients. In this study, data were collected between January 2008 and October 2009.
429
Data collection method In addition to the preliminary version of the HCFS, a questionnaire that asks basic information, such as sex and age, and a 24-h version of the SF-8TM Health Survey that evaluates health-related QOL (HRQOL), was distributed to all participants. Of the participants, i) 126 participants also received an abridged version of POMS-Fatigue and ii) 122 participants also received CFS. In addition, a retest of the preliminary version of HCFS was given to 111 participants who participated in i) or ii) and who indicated their willingness to participate in a retest at the time of the initial explanation of the survey. The retest was conducted 1e4 weeks after the first test only if there were no significant changes in circumstances (such as regular outpatient chemotherapy days) from the first survey. Contents of the different Questionnaire Surveys used. Contents of the questionnaire surveys used other than the preliminary version of HCFS are as follows. All of the scales were chosen to verify the criterion-related validity. a) CFS: This is a cancer patient fatigue scale developed in Japan, and is a questionnaire consisting of 15 items from 3 subscales (physical, affective and cognitive). The questionnaire is evaluated on 5 levels. Cronbach's a coefficient for CFS is .88. This scale was chosen because it is a multidimensional fatigue scale with subscales similar to HCFS. b) Abridged POMS: POMS is a questionnaire developed in the US that consists of a total of 65 items that measure temporary mood and emotions. It evaluates the changes according to the conditions in which the subjects are placed. Five items from one of the subscales, “fatigue”, in this abridged version (30 items) are evaluated on 5 levels. It has verified reliability and validity in Japanese (Yokoyama, 2006). This scale was chosen because it is one of the key measures commonly used internationally. c) SF-8: This is a comprehensive HRQOL scale that measures the following 8 health concepts: physical functioning, role-physical, bodily pain, general health, vitality, social functioning, roleemotional and mental health. The survey is evaluated on 6 levels. The 24-h version used in this study is a survey where the recall period is 24 h. Japanese version is available (Fukuhara and Suzukamo, 2004). The significant correlation between HCFS and SF8 lends support to the conceptual framework of this study. SF8 was chosen because it is less burdensome to complete, requiring only about 1e2 min.
Data analysis SPSS 21.0 for Windows was utilized and data were analyzed by the following methods. Item analysis: For each of the 45 items, the mean values and SD (standard deviation) were calculated. Since all the HCFS questions were evaluated with scores of 1e5, it was conducted with a ceiling effect of 6 and floor effect of 1 for mean ± SD values, items with these values being excluded from analysis. Assessment of validity Construct validity: Factor analysis using principal factor analysis and promax rotation were conducted on items remaining after item analysis. Conditions in which factor loading was .4 and where multiple factors did not have high factor loadings were considered the standard for selection. Criterion-related validity: Pearson's correlation coefficients were determined between total HCFS scores and the respective total scores of CFS and abridged POMS-F. Both CFS and abridged POMS-F were predicted to have a theoretical association with HCFS.
430
K. Hirai et al. / European Journal of Oncology Nursing 19 (2015) 427e432
Discriminant validity: The mean value and SD (standard deviation) of CFS were calculated, subjects were divided into mean þ 1 SD and mean 1 SD groups, and the differences in mean values of the total HCFS scores were analyzed between these two groups. Assessment of reliability To confirm internal consistency of the survey, Cronbach's a coefficients of the overall scale and each factor were calculated. To confirm stability, a retest was conducted 1e4 weeks after the survey, Pearson's correlation coefficient between the total HCFS scores of the first test and retest was determined, and the reliability coefficient was calculated. Ethical considerations This research plan was reviewed and approved by the ethical review board of each of the 4 hospitals where data collection took place. The study was verbally explained to the participants using documents that explained the purpose of the study, the required task and time, and the measures taken to protect the participants' human rights. All participants signed a consent form before participating in the study. Results From the 322 participants who participated in this study, 281 valid responses (response rate 87.3%) were analyzed. Participant attributes The analysis included 147 men (52.3%) and 134 women (47.7%), with a mean age of 58.1 years (±12.9). The most commonly diagnosed cancer site was the breast (30.2%), followed by hematological and lymphatic (26.4%) and gastrointestinal (22.4%) cancers. Chemotherapy was the most frequently used treatment during the survey (70.8%) (Table 1). Item analysis There were no items in the preliminary version of the HCFS that showed a ceiling effect, but there were 25 items that showed floor effects; hence, these items were excluded. Table 1 Sample characteristics.
n Gender Age
Cancer site
Treatment
Place of service
Male Female Mean Median Range Breast Hematological/Lymphatic Head and neck Gastrointestinal Urinary Lung Other Chemotherapy Radiotherapy Chemoradiation Hormonal therapy Other Outpatient Inpatient
n ¼ 281 %
147 52.3 134 47.7 58.1 (±12.9) 59 20e79 85 30.2 74 26.4 5 1.8 63 22.4 37 13.2 14 5.0 3 1.2 199 70.8 24 8.5 7 2.5 35 12.5 16 5.7 164 58.4 117 41.6
Assessment of validity Exploratory factor analysis and designation of factors: Factor analysis was conducted on the 20 items extracted following item analysis. Using the principal factor method, the number of factors was set to 3 and exploratory factor analysis was conducted with promax rotation. There were no items with factor loadings of .4. However, items (3, 28, 36) with high factor loadings in multiple factors were excluded. Next, the same analysis was conducted again for the remaining 17 items, and 1 item (23) with factor loading of .4 was excluded. The same analysis was conducted additionally on these 16 items, with which the analysis converged to 3 factors with 16 items since there were no items that deviated from the standards. However, differences in meaning between “1. I got tired” and “16. I get tired” in Factor I were not clear. Therefore, the same analysis was conducted twice with 15 items, excluding one item or the other. The results indicated that no items deviated from the standard in either analysis, and that there were no changes in the factor structure of 3 factors with 15 items. For this reason, “1. I got tired”, which has a lower reliability coefficient, was excluded to match the verb tenses of other items. This was established as the Hirai Cancer Fatigue Scale (HCFS) (Table 2). In addition, the cumulative contribution ratio before rotation was 65.46% and Spearman's correlation coefficients between the 3 factors ranged between .61 and .74, demonstrating significant positive correlations. The interpretation and designation of each factor is listed below. Factor I, a sensation perceived as one's physical and mental state, was designated as ‘physical/mental sensation’. Factor II, a sensation perceived through carrying out daily activities, was designated as ‘activity-related sensation’. Factor III, a sensation perceived as changes in the ability to think, pay attention and memorize information, was designated as ‘cognitive sensation’. Criterion-related validity: The relationships between HCFS and existing scales were assessed. There was a high correlation between HCFS and CFS with r ¼ .763 (p < 0.01). There was a high correlation between HCFS and abridged POMS-F with r ¼ .759 (p < 0.01). In particular, there was a high correlation between HCFS and Factor I (physical/mental sensation), r ¼ .847 (p < 0.01). The criterionrelated validity of HCFS was confirmed based on these relationships with CFS and abridged POMS-F. We then assessed the relationship between HCFS and SF-8. The total score of HCFS had
Table 2 Factor loading.
H16 H10 H27 H40 H17 H25 H22 H21 H26 H18 H11 H29 H20 H15 H19
I get tired. I feel lethargic. I get tired easily. I have lingering fatigue. I don't feel quite right. I feel down. I feel like I have no energy. It takes much more effort to do things now than before. I can't do anything without taking frequent breaks. I can't do things quickly or efficiently. My body doesn't do what I want it to do. I am unable to finish things right away. I have trouble remembering little things. I feel my thinking ability has declined. I feel my abilities have declined.
I
II
III
.899 .870 .734 .713 .713 .619 .167 .097
.116 .014 .100 .035 .041 .062 1.002 .775
.090 .208 .061 .038 .003 .054 .073 .175
.199
.750
.061
.213 .300 .162 .186 .152 .075
.625 .610 .569 .009 .083 .123
.068 .081 .179 .855 .786 .717
Factor extraction method: Principal factor analysis. Rotation method: Promax rotation with Kaiser Normalization.
K. Hirai et al. / European Journal of Oncology Nursing 19 (2015) 427e432
negative correlations with both RP (role-physical) and PCS (physical component summary), r ¼ .526 (p < 0.01) and r ¼ .465 (p < 0.01), respectively. A negative correlation of r ¼ .443 (p < 0.01) was observed with PF (physical functioning). Discriminant validity: Using CFS, the subjects were divided into mean þ 1 SD and mean 1 SD groups, and their total HCFS scores were compared. The results showed that the mean þ 1 SD group (n ¼ 20) had a significantly greater score of 49.30 points compared to the mean 1 SD group (n ¼ 17) with 22.35 points (p < 0.001), confirming the discriminant validity of HCFS. Assessment of reliability Cronbach's a coefficient was .943 for the overall HCFS, .897 for Factor I, .926 for Factor II, and .843 for Factor III, confirming internal consistency. Reliability coefficient by the test-retest method showed that, overall, the scale had r ¼ .820 (p < 0.01), Factor I had r ¼ .781 (p < 0.01), Factor II had r ¼ .796 (p < 0.01), and Factor III had r ¼ .773 (p < 0.01), confirming the reliability of the scale. Discussion We confirmed that HCFS is a multidimensional fatigue scale with high reliability and high validity. One of the key features of this scale is that it was developed based on the expressions of fatigue used by Japanese cancer patients. Thus, each of the questionnaire items was written in a language that is easy for Japanese people to understand. Items such as “My body doesn't do what I want it to do” in the HCFS were perceptions of fatigue expressed by many patients at the questionnaire development stage, which is not included in the revised PFS or CFS. Currently, in Japan, the available fatigue scales are translated versions of scales developed in other countries or developed by item generation based on literature from other countries. Consequently, the expressions used in these scales, such as “I am fed up” and “I feel that I make errors more often while speaking” which are unfamiliar to many Japanese people in explaining fatigue, may have confused the respondents. HCFS has overcome this issue by creating questions that are easy to understand and easy to answer by the subjects for whom it is intended. The second key feature of HCFS is that it enables multidimensional measurement of fatigue with 15 items. Some of the representative multidimensional fatigue scales available in Japan include the 4-dimensional 22-item revised PFS and 3-dimensional 15-item CFS. Similar to CFS, HCFS can measure fatigue on a 15-item scale in the following 3 dimensions: physical/mental, activity-related, and cognitive sensations. Since the fatigue scale is used by patients who are predicted to have fatigue, having fewer items is extremely important in preventing it from being a burden on the respondents. This in return reduces missing values and the possibility of dropping out of the survey. The third key feature of HCFS is that it is easy to calculate the total score of each dimension and to determine its overall evaluation. For example, in CFS, it is necessary to pick appropriate items in each of the 3 dimensions and perform complicated additions and subtractions. This has been a major obstacle in using this scale in the clinical setting, because with this questionnaire it is impossible to evaluate fatigue with merely a quick glance through the answered questionnaire. HCFS questionnaire items do not include inverted questions, and higher numbers represent greater fatigue in all items; thus, the patient's fatigue can be roughly assessed with just a glance through the answered questionnaire, and the overall fatigue can be evaluated by simple addition. In
431
future, this feature of HCFS will most likely result in its being routinely used to evaluate fatigue as the 6th vital sign, or play an important role in evidence creation of effective fatigue management and appropriate assessment of patient's fatigue, which will significantly impact appropriate evaluation of the effect of intervention in Japan. Furthermore, a negative correlation was confirmed between HCFS and HRQOL measured by SF8. This signifies its concurrence with the conceptual model of this study. In the scale development stage, we set up 3 theoretical subscales (physical, mental, and cognitive fatigue), but after factor analysis, 3 factors (physical/mental, activity-related, and cognitive sensations) were included. As encapsulated by the phrase “I don't feel quite right”, this suggests that fatigue is a complex sensation, within which it is difficult to separate physical and mental sensations. In addition, HCFS includes sensations that are activity-related, such as “I can't do anything without taking frequent breaks” or “I can't do things quickly or efficiently.” With regard to items related to activity, revised PFS evaluates the impact of fatigue on work and daily life, as well as social and sex life, while HCFS directly assesses the extent of sensation perceived in association with daily activities. This is an attribute that does not exist in other existing multi-dimensional fatigue scales, and is considered to be a characteristic of HCFS that enables evaluation of the “current state of fatigue.” Limitations and implications for nursing research and practice This study has two limitations. First, this study was conducted on Japanese people based on phrasing used by Japanese people; thus, it is necessary in the future to assess whether it is applicable for use in non-Japanese people. Second, the data was collected in 2008e2009 and is now outdated. However, fatigue still remains the most common symptom for cancer patients. Hence, the findings are still relevant. Use of HCFS enables convenient, less burdensome selfevaluation of fatigue by Japanese cancer patients. This enables patients to effectively provide information regarding their fatigue to medical professionals, and would provide useful information in the self-management of fatigue. Additionally, when an intervention program for fatigue management is developed, it can be used for outcome evaluation. If HCFS becomes available in multiple languages, it is expected to become beneficial in the management of non-Japanese cancer patients as well. Widespread adoption of this fatigue scale will enable international comparisons about cancer-related fatigue. Further investigation in cooperation with non-Japanese researchers is therefore required. In addition, to further validate the HCFS, future research should assess the responsiveness of this scale, specifically its ability to detect changes that may follow modifications in clinical status, and determine cut-off values for distinguishing severe fatigue. Conflict of interest There is no conflict of interest. Acknowledgments This study was supported by MEXT JSPS KAKENHI Grant Number 0819592507. We thank all the patients and staff at medical institutions that participated in this study for their cooperation.
432
K. Hirai et al. / European Journal of Oncology Nursing 19 (2015) 427e432
Appendix 1. Hirai Cancer Fatigue Scale This is a questionnaire about your fatigue. Please circle the number that most accurately describes your current condition. Please base your response on your first impression, without thinking too much.
Not Slightly Somewhat Very Extremely at all 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
I get tired. I feel lethargic. I get tired easily. I have lingering fatigue. I don't feel quite right. I feel down. I feel like I have no energy. It takes much more effort to do things now than before. I can't do anything without taking frequent breaks. I can't do things quickly or efficiently. My body doesn't do what I want it to do. I am unable to finish things right away. I have trouble remembering little things. I feel my thinking ability has declined. I feel my abilities have declined.
1 1 1 1 1 1 1 1
2 2 2 2 2 2 2 2
3 3 3 3 3 3 3 3
4 4 4 4 4 4 4 4
5 5 5 5 5 5 5 5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
Physical/mental sensation Total score for items 1e6 (A). Activity-related sensation Total score for items 7e12 (B). Cognitive sensation Total score for items 13e15 (C). Overall ðAÞ þ ðBÞ þ ðCÞ
* HCFS has been developed in Japanese and has been translated for the purposed of this manuscript.
References Armstrong, T.S., 2003. Symptoms experience: a concept analysis. Oncology Nursing Forum 30 (4), 601e606. Butt, Z., Rosenbloom, S.K., Abernethy, A.P., Beaumont, J.L., Paul, D., Hampton, D., et al., 2008. Fatigue is the most important symptom for advanced cancer patients who have had chemotherapy. Journal of the National Comprehensive Cancer Network 6 (5), 448e455. Dodd, M., Janson, S., Facione, N., Faucett, J., Froelicher, E.S., Humphreys, J., et al., 2001. Advancing the science of symptom management. Journal of Advanced Nursing 33 (5), 668e676. Foundation for Promotion of Cancer Research, 2013. Cancer Statistics in Japan 2013. Available from: http://ganjoho.jp/data/professional/statistics/backnumber/
2013/cancer_statistics_2013.pdf#search¼'cancerþstatisticsþinþJApan' (accessed 30.06.14.). Fu, M.R., Anderson, C.M., McDaniel, R., Armer, J., 2002. Patients' perception of fatigue in response to biochemotherapy as a treatment for metastatic melanoma. Oncology Nursing Forum 29, 961e966. Fu, M.R., Roxanne, W., McDaniel, R., Rhodes, V.A., 2005. Fatigue. In: Yarbro, C.H., Frogge, M.H., Goodman, M. (Eds.), Cancer Nursing: Principles and Practice, sixth ed. Jones and Bartlett, Massachusetts, pp. 741e760. Fukuhara, S., Suzukamo, Y., 2004. Manual of the SF-8 Japanese Version. Institute for Health Outcomes and Process Evaluation Research, Kyoto. Hirai, K., Kanda, K., Hosokawa, M., Takagai, J., 2014. Perception of fatigue expressed in Japanese cancer patients. Kitakanto Medical Journal 64, 43e49. Kamizato, M., 1999. Fatigue in cancer patients receiving treatment with radiotherapy. Journal of Japanese Society of Cancer Nursing 13 (2), 48e59. Kutner, J.S., Kassner, C.T., Nowels, D.E., 2001. Symptom burden at the end of life: hospice providers' perceptions. Journal of Pain and Symptom Management 21 (6), 473e480. Lawrence, D.P., Kupelnick, B., Miller, K., Devine, D., Lau, J., 2004. Evidence report on the occurrence, assessment, and treatment of fatigue in cancer patients. Journal of the National Cancer Institute Monographs 32, 40e50. McDaniel, R.W., Rhodes, V.A., 1998. Development of a preparatory sensory information videotape for women receiving chemotherapy for breast cancer. Cancer Nursing 21, 143e148. McNair, D., Lorr, M., Droppleman, L., 1992. Profile of Mood States. Education and Industrial Testing Service, San Diego. Mendoza, T., Wang, X., Cleeland, C., Morrissey, M., Johnson, B., Wendt, J., et al., 1999. The rapid assessment of fatigue severity in cancer patients: use of the brief fatigue inventory. Cancer 85, 1186e1196. Okuyama, T., Akechi, T., Kugaya, A., Okamura, H., Shima, Y., Maruguchi, M., et al., 2000. Development and validation of the Cancer Fatigue Scale: a brief, threedimensional, self-rating scale for assessment of fatigue in cancer patients. Journal of Pain and Symptom Management 19 (1), 5e14. Piper, B.F., Lindsey, A.M., Dodd, M.J., Ferketich, S., Paul, S.M., Weller, S., 1989. The development of an instrument to measure the subjective dimension of fatigue. In: Funk, S.G., Tornquist, E.M., Champagne, M.T., Copp, L.A., Wiese, R.A. (Eds.), Key Aspects of Comfort: Management of Pain, Fatigue, and Nausea. Springer, New York, pp. 199e208. Piper, B.F., Dibble, S.L., Dodd, M.J., Weiss, M.C., Slaughter, R.E., Paul, S.M., 1998. The revised Piper Fatigue Scale: psychometric evaluation in women with breast cancer. Oncology Nursing Forum 25, 677e684. Prue, G., Rankin, J., Allen, J., Gracey, J., Cramp, F., 2006. Cancer-related fatigue: a critical appraisal. European Journal of Cancer 42, 846e863. Rhodes, V.A., Watson, P.M., McDaniel, R.W., Hanson, B.M., Johnson, M.H., 1995. Expectation and occurrence of postchemotherapy side effects. Cancer Practice 3, 247e253. Schwartz, A.L., 1998. The Schwartz Cancer Fatigue Scale: testing reliability and validity. Oncology Nursing Forum 25, 711e717. Siefert, M.L., 2010. Fatigue, pain, and functional status during outpatient chemotherapy. Oncology Nursing Forum 37 (2), 114e123. Stricker, C., Drake, D., Hoyer, K., Mock, V., 2004. Evidence-based practice for CRF management in adults with cancer: exercises as an intervention. Oncology Nursing Forum 31 (5), 963e973. Vogelzang, N.J., Breitbart, E., Cella, D., Curt, G.A., Groopman, J.E., Horning, S.J., et al., 1997. Patient, caregiver, and oncologist perceptions of cancer-related fatigue: results of a tripart assessment survey. The fatigue coalition. Seminars in Hematology 34 (Suppl. 2), 4e12. Yellan, E., 1997. Measuring fatigue and other anemia-related symptoms with the functional assessment of cancer therapy (FACT) measurement system. Journal of Pain and Symptom Management 13 (2), 63e74. Yokoyama, K., 2006. Abridged POMS-f: Guidance and Case Commentary. Kanekosyoho, Tokyo.
Contents lists available at ScienceDirect
European Journal of Oncology Nursing journal homepage: www.elsevier.com/locate/ejon
Development of the Hirai Cancer Fatigue Scale: Testing its reliability and validity Kazue Hirai a, *, Kiyoko Kanda b, Junko Takagai c, Mai Hosokawa d a
Faculty of Medicine, School of Nursing, Tokyo Medical University, 6-1-1 Shinjuku, Shinjuku-ku, Tokyo, 160-8402, Japan Department of Nursing, Gunma University Graduate School of Health Science, 3-29-21 Showa-machi, Maebashi, Gunma, 371-8514, Japan c Faculty of Medicine, School of Health Science, Akita University, 1-1-1 Hondou, Akita, Akita, 010-8543, Japan d The Jikei University School of Nursing, 8-3-1, Kokuryocho, Chofu, Tokyo, 182-8570, Japan b
a b s t r a c t Keywords: Fatigue Cancer Measurement Symptom assessment Quality of life
Purpose: The purpose of this study was to develop the Hirai Cancer Fatigue Scale (HCFS) to assess the fatigue experienced by cancer patients, and to verify its reliability and validity. Methods: Based on qualitative research about the perception of fatigue by Japanese cancer patients, we developed a questionnaire. The content validity was confirmed by 5 expert oncology nurses and 5 oncologists. 281 Japanese cancer patients participated in this study. Construct validity was analyzed using factor analysis, and internal consistency was analyzed using Cronbach's a coefficient. Results: A 15-item scale with 3 dimensions, “physical/mental sensation”, “activity-related sensation” and “cognitive sensation,” was developed by factor analysis. This scale had an overall Cronbach's a coefficient of .943 and a test-retest reliability coefficient of r ¼ .820 (p < 0.01), confirming the high reliability of the scale. The correlation coefficient was r ¼ .759 (p < 0.01) between HCFS and abridged Profile of Mood States-Fatigue (POMS-F), and r ¼ .763 (p < 0.01) between HCFS and Cancer Fatigue Scale (CFS), both showing high correlations and confirming criterion-related validity. Conclusion: HCFS enables reliable and valid evaluation of Japanese cancer patients' fatigue. Use of the HCFS would assist in convenient self-evaluation of fatigue, and would allow information to be effectively provided to healthcare professionals. It could also be used for outcome evaluation in an intervention study. © 2014 Elsevier Ltd. All rights reserved.
Introduction Fatigue is the most common symptom among cancer patients (Lawrence et al., 2004; Prue et al., 2006), and is known to be one of the main symptoms of pain. It exists in all stages of cancer survivorship, from diagnosis to end of life; some people have already experienced it at the time of diagnosis of cancer (Vogelzang et al., 1997), and it is an extremely common adverse effect during treatments such as chemotherapy (Butt et al., 2008; Siefert, 2010) and radiation therapy (Stricker et al., 2004). It persists long after completing a standard course of treatment (Kutner et al., 2001),
* Corresponding author. Tel.: þ81 3 5357 7304; fax: þ81 3 3351 3693. E-mail addresses: [email protected] (K. Hirai), [email protected] (K. Kanda), [email protected] (J. Takagai), [email protected] (M. Hosokawa). http://dx.doi.org/10.1016/j.ejon.2014.12.004 1462-3889/© 2014 Elsevier Ltd. All rights reserved.
with almost all patients experiencing fatigue in the terminal stage (Fu et al., 2005). Therefore, properly managing fatigue is a significant issue that must be addressed in order for cancer survivors to maintain an adequate quality of life (QOL). The number one cause of death in Japan, the world's leader in longevity, is cancer, and 1 in 2 Japanese people suffer from cancer at some point in life (Cancer Statistics in Japan, 2013). Until about 15 years ago, it was usual for Japanese cancer patients to be hospitalized to complete chemotherapy or radiotherapy. Recently, outpatient cancer treatments have become common, the duration of hospitalization for all kinds of cancer treatment have been shortened, and services that allow cancer patients to spend their terminal stages at home have been promoted. However, fatigue in cancer patients is still rarely a focus in either clinical settings or in research fields in Japan. Outpatient nurses can play an important role in the development of a plan to properly manage fatigue that occurs frequently and diminishes the QOL of cancer patients. To
428
K. Hirai et al. / European Journal of Oncology Nursing 19 (2015) 427e432
accomplish this, a scale with high reliability and high validity to properly evaluate fatigue, and that is easy enough for use at an outpatient level, is necessary. Despite the various definitions of fatigue that have been tested to date, a universal definition does not exist. As the consensus is that fatigue is a subjective and multidimensional symptom (Rhodes et al., 1995; McDaniel and Rhodes, 1998; Fu et al., 2002), fatigue scales require the capability of multidimensional evaluation. Currently, the fatigue scales for cancer patients that can be used in Japanese are: revised Piper Fatigue Scale (revised PFS) (Piper et al., 1998), Cancer Fatigue Scale (CFS) (Okuyama et al., 2000), Brief Fatigue Inventory (BFI) (Mendoza et al., 1999), Profile of Mood StatesFatigue (POMS-F) (McNair et al., 1992), and Functional Assessment of Cancer Therapy (FACT) (Yellen, 1997). However, of these scales, only the revised PFS and CFS are multidimensional scales. The revised PFS, the first multidimensional fatigue scale, is a well-designed scale based on a theoretical model that makes 4dimensional measurements of fatigue: severity/behavioral, sensory, affective/meaning, and cognitive/mood. It is a significantly improved version of the original 82-item version with complex wording (Piper et al., 1989). However, as Schwartz (1998) pointed out, there are too many question items, 22 in total, which require a long time to answer, thus possibly becoming a burden on the respondents. In Japan, Kamizato (1999) announced a Japanese version of Piper's 22-item scale and used it in a study to elucidate fatigue in cancer patients undergoing radiation therapy; however, since then it has been little used in either research or clinical settings. CFS is the only multidimensional fatigue scale developed by Japanese researchers that measures fatigue in 3-dimensions: physical, affective and cognitive. This scale is often used in nursing research in Japan because it has 15 items, making it considerably shorter than the revised PFS, and thus less of a burden on patients. However, it is far from convenient, as complicated calculations for each dimension are required for its evaluation. For this reason, this scale has not been used very often in clinical practice to quickly assess outpatient fatigue. In addition, it includes items such as “Do you feel that you make errors more often while speaking?”, which is an expression that is unfamiliar and not normally used by Japanese people when they talk about fatigue. Given the limitations of the previous fatigue scales, we believe that development of a multidimensional fatigue scale based on how Japanese cancer patients express their perception of fatigue, and one that is easier to use, is much required. Purpose The purpose of this study was to develop the Hirai Cancer Fatigue Scale (HCFS) that assesses fatigue experienced by cancer patients, and to verify its reliability and validity.
symptoms' frequency, intensity, distress, and meaning, and that there are individual differences in perception and expression. The same model also states that the expression or consequences of the symptoms are described as changes in functional and cognitive activities, including adjustment to illness and quality of life. Based on these models, we set up the conceptual framework of this study, as described below. Fatigue experienced by cancer patients is a series of dynamic processes, where “perception of fatigue”, “coping”, and “outcome” influence each other. It is also influenced by demographic characteristics (sex and age), disease characteristics (diagnosis, stage of disease, treatment and health status), and personal characteristics (physical, psychological, social and cultural factors) (Fig.1). Based on this conceptual framework, we developed a scale that measures the perception of fatigue in this study. Development stage Preliminary investigation We first elucidated how Japanese cancer patients express their perception of fatigue, and, with the objective of conceptualizing this, we conducted an open-ended questionnaire survey that asked, “How would you describe your sensation of fatigue?” among 400 inpatient or outpatient Japanese cancer patients at 2 national hospitals in eastern Japan (Hirai et al., 2014). The survey was reviewed and approved by the respective ethical review boards of each hospital. Content analysis showed that 237 codes of fatigue were obtained and 35 subcategories/15 categories/4 corecategories were extracted. The 4 core-categories were “physical sensation”, “mental sensation”, “cognitive sensation”, and “sensation that cannot be described with words”. Physical sensations consisted of “My body feels heavy”, “I am tired/feel listless or sluggish/don't feel quite right”, “I want to lie down/sit down”, “I am exhausted/can't get comfortable”, “My body doesn't do what I want it to do”, “I am sleepy”, and “I lack endurance”. Mental sensations consisted of “I have no willpower/motivation”, “Everything feels bothersome”, “I don't want to do anything/I don't want to move”, “I am depressed/anxious”, and “I am frustrated”. Cognitive sensations consisted of “reduction in ability to think” and “reduction in concentration”. In addition, indescribable sensations were indicated by “cannot describe in words”. Furthermore, we performed a literature review of fatigue and fatigue scales for cancer patients, which aimed to establish the definition and the conceptual framework and to understand the characteristics of each scale. The review revealed that the multidimensionality of fatigue was mostly described in physical, mental and cognitive aspects. It also revealed that existing scales have not necessarily been developed based on the definition of fatigue. The
Methods Conceptual framework According to the Revised Symptom Management Conceptual Model (Dodd et al., 2001), symptom management consists of three interrelated concepts: “symptom experience”, “symptom management strategy”, and “outcomes”. According to this model, symptom management is a dynamic process that is modified by individual outcomes and the influences of the nursing domains of person, health/illness, or environment. Additionally, according to the Symptoms Experience Model (Armstrong, 2003), various elements of individuals become antecedents of symptom experience and act on symptom production. This model states that the perceived components of the symptom experience include the
Fig. 1. Conceptual framework.
K. Hirai et al. / European Journal of Oncology Nursing 19 (2015) 427e432
results of the literature review contributed to establishment of the basis of the present study and to develop a draft questionnaire. Theoretical establishment of construct and subscale From the results of the questionnaire survey and literature review, we defined the perception of fatigue by cancer patients as “a sensation characterized by the decline of functional status and the discomfort associated with lack of energy”. We also theoretically established the construct as: 1) Physical sensation: Unpleasant sensation perceived physically, reduction in physical function and a sense of loss of physical control; 2) Mental sensation: Sensation characterized by depression and decline in motivation and willpower; and 3) Cognitive sensation: Sensation characterized by diminished ability to think and pay attention, and by reduction in memory. Creation of the first draft of the questionnaire Each of the three constructs included codes obtained through preliminary investigation of the item pool. Questionnaire items were also added by brainstorming among the authors, and by searching the reports identified in the literature review for the keywords “fatigue scale/measurement” and “fatigue”. Assessments and modifications were repeated among the researchers to avoid semantic duplication and to improve clarity of expression for each item, and a 49-item questionnaire was created. The questions focused on the “current condition”, with the instructions to participants stating, “This is a questionnaire about your fatigue. Please circle the number that most accurately describes your current condition”. The evaluation method used a 5-point Likert scale (1: Does not apply at all/Not at all, 2: Slightly applicable, 3: Somewhat applicable, 4: Very applicable, 5: Extremely applicable). This became the first draft of the HCFS. Assessment of content validity We assessed content validity by requesting advice related to the suitability of the HCFS first draft from 10 experts (5 oncologists and 5 expert oncology nurses). We asked these experts to evaluate each of the 49 items on the following three points: 1) whether the question expresses perception of fatigue, 2) whether the question expresses different aspects of said perception, including physical, mental and cognitive perceptions, and 3) whether the question is easy to understand in Japanese and easy to answer. There were 4 items (“I am distracted”, “I am frustrated”, “I am overwhelmed”, and “I am fed up”) that were removed either because the experts' opinions were divided on these points or because the item was determined unsuitable, resulting in a final total of 45 items in the HCFS. This was the preliminary version of the HCFS. Validation stage Sample Participants included outpatients and inpatients with various types of cancer, who were receiving chemotherapy and/or radiotherapy and/or hormone therapy or were under observation after completion of their treatment, at 4 large-scale public hospitals in eastern Japan. Patients eligible for participation in this study included (a) those who were aware of their diagnosis of cancer, (b) were in their 20e70's, and (c) were not suffering from mental or cognitive disorders. The nursing administrators of each outpatient department and ward selected the patients that met the eligibility criteria (except for immediately after surgery). Thereafter, the researchers were introduced to the patients. In this study, data were collected between January 2008 and October 2009.
429
Data collection method In addition to the preliminary version of the HCFS, a questionnaire that asks basic information, such as sex and age, and a 24-h version of the SF-8TM Health Survey that evaluates health-related QOL (HRQOL), was distributed to all participants. Of the participants, i) 126 participants also received an abridged version of POMS-Fatigue and ii) 122 participants also received CFS. In addition, a retest of the preliminary version of HCFS was given to 111 participants who participated in i) or ii) and who indicated their willingness to participate in a retest at the time of the initial explanation of the survey. The retest was conducted 1e4 weeks after the first test only if there were no significant changes in circumstances (such as regular outpatient chemotherapy days) from the first survey. Contents of the different Questionnaire Surveys used. Contents of the questionnaire surveys used other than the preliminary version of HCFS are as follows. All of the scales were chosen to verify the criterion-related validity. a) CFS: This is a cancer patient fatigue scale developed in Japan, and is a questionnaire consisting of 15 items from 3 subscales (physical, affective and cognitive). The questionnaire is evaluated on 5 levels. Cronbach's a coefficient for CFS is .88. This scale was chosen because it is a multidimensional fatigue scale with subscales similar to HCFS. b) Abridged POMS: POMS is a questionnaire developed in the US that consists of a total of 65 items that measure temporary mood and emotions. It evaluates the changes according to the conditions in which the subjects are placed. Five items from one of the subscales, “fatigue”, in this abridged version (30 items) are evaluated on 5 levels. It has verified reliability and validity in Japanese (Yokoyama, 2006). This scale was chosen because it is one of the key measures commonly used internationally. c) SF-8: This is a comprehensive HRQOL scale that measures the following 8 health concepts: physical functioning, role-physical, bodily pain, general health, vitality, social functioning, roleemotional and mental health. The survey is evaluated on 6 levels. The 24-h version used in this study is a survey where the recall period is 24 h. Japanese version is available (Fukuhara and Suzukamo, 2004). The significant correlation between HCFS and SF8 lends support to the conceptual framework of this study. SF8 was chosen because it is less burdensome to complete, requiring only about 1e2 min.
Data analysis SPSS 21.0 for Windows was utilized and data were analyzed by the following methods. Item analysis: For each of the 45 items, the mean values and SD (standard deviation) were calculated. Since all the HCFS questions were evaluated with scores of 1e5, it was conducted with a ceiling effect of 6 and floor effect of 1 for mean ± SD values, items with these values being excluded from analysis. Assessment of validity Construct validity: Factor analysis using principal factor analysis and promax rotation were conducted on items remaining after item analysis. Conditions in which factor loading was .4 and where multiple factors did not have high factor loadings were considered the standard for selection. Criterion-related validity: Pearson's correlation coefficients were determined between total HCFS scores and the respective total scores of CFS and abridged POMS-F. Both CFS and abridged POMS-F were predicted to have a theoretical association with HCFS.
430
K. Hirai et al. / European Journal of Oncology Nursing 19 (2015) 427e432
Discriminant validity: The mean value and SD (standard deviation) of CFS were calculated, subjects were divided into mean þ 1 SD and mean 1 SD groups, and the differences in mean values of the total HCFS scores were analyzed between these two groups. Assessment of reliability To confirm internal consistency of the survey, Cronbach's a coefficients of the overall scale and each factor were calculated. To confirm stability, a retest was conducted 1e4 weeks after the survey, Pearson's correlation coefficient between the total HCFS scores of the first test and retest was determined, and the reliability coefficient was calculated. Ethical considerations This research plan was reviewed and approved by the ethical review board of each of the 4 hospitals where data collection took place. The study was verbally explained to the participants using documents that explained the purpose of the study, the required task and time, and the measures taken to protect the participants' human rights. All participants signed a consent form before participating in the study. Results From the 322 participants who participated in this study, 281 valid responses (response rate 87.3%) were analyzed. Participant attributes The analysis included 147 men (52.3%) and 134 women (47.7%), with a mean age of 58.1 years (±12.9). The most commonly diagnosed cancer site was the breast (30.2%), followed by hematological and lymphatic (26.4%) and gastrointestinal (22.4%) cancers. Chemotherapy was the most frequently used treatment during the survey (70.8%) (Table 1). Item analysis There were no items in the preliminary version of the HCFS that showed a ceiling effect, but there were 25 items that showed floor effects; hence, these items were excluded. Table 1 Sample characteristics.
n Gender Age
Cancer site
Treatment
Place of service
Male Female Mean Median Range Breast Hematological/Lymphatic Head and neck Gastrointestinal Urinary Lung Other Chemotherapy Radiotherapy Chemoradiation Hormonal therapy Other Outpatient Inpatient
n ¼ 281 %
147 52.3 134 47.7 58.1 (±12.9) 59 20e79 85 30.2 74 26.4 5 1.8 63 22.4 37 13.2 14 5.0 3 1.2 199 70.8 24 8.5 7 2.5 35 12.5 16 5.7 164 58.4 117 41.6
Assessment of validity Exploratory factor analysis and designation of factors: Factor analysis was conducted on the 20 items extracted following item analysis. Using the principal factor method, the number of factors was set to 3 and exploratory factor analysis was conducted with promax rotation. There were no items with factor loadings of .4. However, items (3, 28, 36) with high factor loadings in multiple factors were excluded. Next, the same analysis was conducted again for the remaining 17 items, and 1 item (23) with factor loading of .4 was excluded. The same analysis was conducted additionally on these 16 items, with which the analysis converged to 3 factors with 16 items since there were no items that deviated from the standards. However, differences in meaning between “1. I got tired” and “16. I get tired” in Factor I were not clear. Therefore, the same analysis was conducted twice with 15 items, excluding one item or the other. The results indicated that no items deviated from the standard in either analysis, and that there were no changes in the factor structure of 3 factors with 15 items. For this reason, “1. I got tired”, which has a lower reliability coefficient, was excluded to match the verb tenses of other items. This was established as the Hirai Cancer Fatigue Scale (HCFS) (Table 2). In addition, the cumulative contribution ratio before rotation was 65.46% and Spearman's correlation coefficients between the 3 factors ranged between .61 and .74, demonstrating significant positive correlations. The interpretation and designation of each factor is listed below. Factor I, a sensation perceived as one's physical and mental state, was designated as ‘physical/mental sensation’. Factor II, a sensation perceived through carrying out daily activities, was designated as ‘activity-related sensation’. Factor III, a sensation perceived as changes in the ability to think, pay attention and memorize information, was designated as ‘cognitive sensation’. Criterion-related validity: The relationships between HCFS and existing scales were assessed. There was a high correlation between HCFS and CFS with r ¼ .763 (p < 0.01). There was a high correlation between HCFS and abridged POMS-F with r ¼ .759 (p < 0.01). In particular, there was a high correlation between HCFS and Factor I (physical/mental sensation), r ¼ .847 (p < 0.01). The criterionrelated validity of HCFS was confirmed based on these relationships with CFS and abridged POMS-F. We then assessed the relationship between HCFS and SF-8. The total score of HCFS had
Table 2 Factor loading.
H16 H10 H27 H40 H17 H25 H22 H21 H26 H18 H11 H29 H20 H15 H19
I get tired. I feel lethargic. I get tired easily. I have lingering fatigue. I don't feel quite right. I feel down. I feel like I have no energy. It takes much more effort to do things now than before. I can't do anything without taking frequent breaks. I can't do things quickly or efficiently. My body doesn't do what I want it to do. I am unable to finish things right away. I have trouble remembering little things. I feel my thinking ability has declined. I feel my abilities have declined.
I
II
III
.899 .870 .734 .713 .713 .619 .167 .097
.116 .014 .100 .035 .041 .062 1.002 .775
.090 .208 .061 .038 .003 .054 .073 .175
.199
.750
.061
.213 .300 .162 .186 .152 .075
.625 .610 .569 .009 .083 .123
.068 .081 .179 .855 .786 .717
Factor extraction method: Principal factor analysis. Rotation method: Promax rotation with Kaiser Normalization.
K. Hirai et al. / European Journal of Oncology Nursing 19 (2015) 427e432
negative correlations with both RP (role-physical) and PCS (physical component summary), r ¼ .526 (p < 0.01) and r ¼ .465 (p < 0.01), respectively. A negative correlation of r ¼ .443 (p < 0.01) was observed with PF (physical functioning). Discriminant validity: Using CFS, the subjects were divided into mean þ 1 SD and mean 1 SD groups, and their total HCFS scores were compared. The results showed that the mean þ 1 SD group (n ¼ 20) had a significantly greater score of 49.30 points compared to the mean 1 SD group (n ¼ 17) with 22.35 points (p < 0.001), confirming the discriminant validity of HCFS. Assessment of reliability Cronbach's a coefficient was .943 for the overall HCFS, .897 for Factor I, .926 for Factor II, and .843 for Factor III, confirming internal consistency. Reliability coefficient by the test-retest method showed that, overall, the scale had r ¼ .820 (p < 0.01), Factor I had r ¼ .781 (p < 0.01), Factor II had r ¼ .796 (p < 0.01), and Factor III had r ¼ .773 (p < 0.01), confirming the reliability of the scale. Discussion We confirmed that HCFS is a multidimensional fatigue scale with high reliability and high validity. One of the key features of this scale is that it was developed based on the expressions of fatigue used by Japanese cancer patients. Thus, each of the questionnaire items was written in a language that is easy for Japanese people to understand. Items such as “My body doesn't do what I want it to do” in the HCFS were perceptions of fatigue expressed by many patients at the questionnaire development stage, which is not included in the revised PFS or CFS. Currently, in Japan, the available fatigue scales are translated versions of scales developed in other countries or developed by item generation based on literature from other countries. Consequently, the expressions used in these scales, such as “I am fed up” and “I feel that I make errors more often while speaking” which are unfamiliar to many Japanese people in explaining fatigue, may have confused the respondents. HCFS has overcome this issue by creating questions that are easy to understand and easy to answer by the subjects for whom it is intended. The second key feature of HCFS is that it enables multidimensional measurement of fatigue with 15 items. Some of the representative multidimensional fatigue scales available in Japan include the 4-dimensional 22-item revised PFS and 3-dimensional 15-item CFS. Similar to CFS, HCFS can measure fatigue on a 15-item scale in the following 3 dimensions: physical/mental, activity-related, and cognitive sensations. Since the fatigue scale is used by patients who are predicted to have fatigue, having fewer items is extremely important in preventing it from being a burden on the respondents. This in return reduces missing values and the possibility of dropping out of the survey. The third key feature of HCFS is that it is easy to calculate the total score of each dimension and to determine its overall evaluation. For example, in CFS, it is necessary to pick appropriate items in each of the 3 dimensions and perform complicated additions and subtractions. This has been a major obstacle in using this scale in the clinical setting, because with this questionnaire it is impossible to evaluate fatigue with merely a quick glance through the answered questionnaire. HCFS questionnaire items do not include inverted questions, and higher numbers represent greater fatigue in all items; thus, the patient's fatigue can be roughly assessed with just a glance through the answered questionnaire, and the overall fatigue can be evaluated by simple addition. In
431
future, this feature of HCFS will most likely result in its being routinely used to evaluate fatigue as the 6th vital sign, or play an important role in evidence creation of effective fatigue management and appropriate assessment of patient's fatigue, which will significantly impact appropriate evaluation of the effect of intervention in Japan. Furthermore, a negative correlation was confirmed between HCFS and HRQOL measured by SF8. This signifies its concurrence with the conceptual model of this study. In the scale development stage, we set up 3 theoretical subscales (physical, mental, and cognitive fatigue), but after factor analysis, 3 factors (physical/mental, activity-related, and cognitive sensations) were included. As encapsulated by the phrase “I don't feel quite right”, this suggests that fatigue is a complex sensation, within which it is difficult to separate physical and mental sensations. In addition, HCFS includes sensations that are activity-related, such as “I can't do anything without taking frequent breaks” or “I can't do things quickly or efficiently.” With regard to items related to activity, revised PFS evaluates the impact of fatigue on work and daily life, as well as social and sex life, while HCFS directly assesses the extent of sensation perceived in association with daily activities. This is an attribute that does not exist in other existing multi-dimensional fatigue scales, and is considered to be a characteristic of HCFS that enables evaluation of the “current state of fatigue.” Limitations and implications for nursing research and practice This study has two limitations. First, this study was conducted on Japanese people based on phrasing used by Japanese people; thus, it is necessary in the future to assess whether it is applicable for use in non-Japanese people. Second, the data was collected in 2008e2009 and is now outdated. However, fatigue still remains the most common symptom for cancer patients. Hence, the findings are still relevant. Use of HCFS enables convenient, less burdensome selfevaluation of fatigue by Japanese cancer patients. This enables patients to effectively provide information regarding their fatigue to medical professionals, and would provide useful information in the self-management of fatigue. Additionally, when an intervention program for fatigue management is developed, it can be used for outcome evaluation. If HCFS becomes available in multiple languages, it is expected to become beneficial in the management of non-Japanese cancer patients as well. Widespread adoption of this fatigue scale will enable international comparisons about cancer-related fatigue. Further investigation in cooperation with non-Japanese researchers is therefore required. In addition, to further validate the HCFS, future research should assess the responsiveness of this scale, specifically its ability to detect changes that may follow modifications in clinical status, and determine cut-off values for distinguishing severe fatigue. Conflict of interest There is no conflict of interest. Acknowledgments This study was supported by MEXT JSPS KAKENHI Grant Number 0819592507. We thank all the patients and staff at medical institutions that participated in this study for their cooperation.
432
K. Hirai et al. / European Journal of Oncology Nursing 19 (2015) 427e432
Appendix 1. Hirai Cancer Fatigue Scale This is a questionnaire about your fatigue. Please circle the number that most accurately describes your current condition. Please base your response on your first impression, without thinking too much.
Not Slightly Somewhat Very Extremely at all 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
I get tired. I feel lethargic. I get tired easily. I have lingering fatigue. I don't feel quite right. I feel down. I feel like I have no energy. It takes much more effort to do things now than before. I can't do anything without taking frequent breaks. I can't do things quickly or efficiently. My body doesn't do what I want it to do. I am unable to finish things right away. I have trouble remembering little things. I feel my thinking ability has declined. I feel my abilities have declined.
1 1 1 1 1 1 1 1
2 2 2 2 2 2 2 2
3 3 3 3 3 3 3 3
4 4 4 4 4 4 4 4
5 5 5 5 5 5 5 5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
Physical/mental sensation Total score for items 1e6 (A). Activity-related sensation Total score for items 7e12 (B). Cognitive sensation Total score for items 13e15 (C). Overall ðAÞ þ ðBÞ þ ðCÞ
* HCFS has been developed in Japanese and has been translated for the purposed of this manuscript.
References Armstrong, T.S., 2003. Symptoms experience: a concept analysis. Oncology Nursing Forum 30 (4), 601e606. Butt, Z., Rosenbloom, S.K., Abernethy, A.P., Beaumont, J.L., Paul, D., Hampton, D., et al., 2008. Fatigue is the most important symptom for advanced cancer patients who have had chemotherapy. Journal of the National Comprehensive Cancer Network 6 (5), 448e455. Dodd, M., Janson, S., Facione, N., Faucett, J., Froelicher, E.S., Humphreys, J., et al., 2001. Advancing the science of symptom management. Journal of Advanced Nursing 33 (5), 668e676. Foundation for Promotion of Cancer Research, 2013. Cancer Statistics in Japan 2013. Available from: http://ganjoho.jp/data/professional/statistics/backnumber/
2013/cancer_statistics_2013.pdf#search¼'cancerþstatisticsþinþJApan' (accessed 30.06.14.). Fu, M.R., Anderson, C.M., McDaniel, R., Armer, J., 2002. Patients' perception of fatigue in response to biochemotherapy as a treatment for metastatic melanoma. Oncology Nursing Forum 29, 961e966. Fu, M.R., Roxanne, W., McDaniel, R., Rhodes, V.A., 2005. Fatigue. In: Yarbro, C.H., Frogge, M.H., Goodman, M. (Eds.), Cancer Nursing: Principles and Practice, sixth ed. Jones and Bartlett, Massachusetts, pp. 741e760. Fukuhara, S., Suzukamo, Y., 2004. Manual of the SF-8 Japanese Version. Institute for Health Outcomes and Process Evaluation Research, Kyoto. Hirai, K., Kanda, K., Hosokawa, M., Takagai, J., 2014. Perception of fatigue expressed in Japanese cancer patients. Kitakanto Medical Journal 64, 43e49. Kamizato, M., 1999. Fatigue in cancer patients receiving treatment with radiotherapy. Journal of Japanese Society of Cancer Nursing 13 (2), 48e59. Kutner, J.S., Kassner, C.T., Nowels, D.E., 2001. Symptom burden at the end of life: hospice providers' perceptions. Journal of Pain and Symptom Management 21 (6), 473e480. Lawrence, D.P., Kupelnick, B., Miller, K., Devine, D., Lau, J., 2004. Evidence report on the occurrence, assessment, and treatment of fatigue in cancer patients. Journal of the National Cancer Institute Monographs 32, 40e50. McDaniel, R.W., Rhodes, V.A., 1998. Development of a preparatory sensory information videotape for women receiving chemotherapy for breast cancer. Cancer Nursing 21, 143e148. McNair, D., Lorr, M., Droppleman, L., 1992. Profile of Mood States. Education and Industrial Testing Service, San Diego. Mendoza, T., Wang, X., Cleeland, C., Morrissey, M., Johnson, B., Wendt, J., et al., 1999. The rapid assessment of fatigue severity in cancer patients: use of the brief fatigue inventory. Cancer 85, 1186e1196. Okuyama, T., Akechi, T., Kugaya, A., Okamura, H., Shima, Y., Maruguchi, M., et al., 2000. Development and validation of the Cancer Fatigue Scale: a brief, threedimensional, self-rating scale for assessment of fatigue in cancer patients. Journal of Pain and Symptom Management 19 (1), 5e14. Piper, B.F., Lindsey, A.M., Dodd, M.J., Ferketich, S., Paul, S.M., Weller, S., 1989. The development of an instrument to measure the subjective dimension of fatigue. In: Funk, S.G., Tornquist, E.M., Champagne, M.T., Copp, L.A., Wiese, R.A. (Eds.), Key Aspects of Comfort: Management of Pain, Fatigue, and Nausea. Springer, New York, pp. 199e208. Piper, B.F., Dibble, S.L., Dodd, M.J., Weiss, M.C., Slaughter, R.E., Paul, S.M., 1998. The revised Piper Fatigue Scale: psychometric evaluation in women with breast cancer. Oncology Nursing Forum 25, 677e684. Prue, G., Rankin, J., Allen, J., Gracey, J., Cramp, F., 2006. Cancer-related fatigue: a critical appraisal. European Journal of Cancer 42, 846e863. Rhodes, V.A., Watson, P.M., McDaniel, R.W., Hanson, B.M., Johnson, M.H., 1995. Expectation and occurrence of postchemotherapy side effects. Cancer Practice 3, 247e253. Schwartz, A.L., 1998. The Schwartz Cancer Fatigue Scale: testing reliability and validity. Oncology Nursing Forum 25, 711e717. Siefert, M.L., 2010. Fatigue, pain, and functional status during outpatient chemotherapy. Oncology Nursing Forum 37 (2), 114e123. Stricker, C., Drake, D., Hoyer, K., Mock, V., 2004. Evidence-based practice for CRF management in adults with cancer: exercises as an intervention. Oncology Nursing Forum 31 (5), 963e973. Vogelzang, N.J., Breitbart, E., Cella, D., Curt, G.A., Groopman, J.E., Horning, S.J., et al., 1997. Patient, caregiver, and oncologist perceptions of cancer-related fatigue: results of a tripart assessment survey. The fatigue coalition. Seminars in Hematology 34 (Suppl. 2), 4e12. Yellan, E., 1997. Measuring fatigue and other anemia-related symptoms with the functional assessment of cancer therapy (FACT) measurement system. Journal of Pain and Symptom Management 13 (2), 63e74. Yokoyama, K., 2006. Abridged POMS-f: Guidance and Case Commentary. Kanekosyoho, Tokyo.
