Psychometric qualities of a brief self-rated fatigue measure: The Fatigue Assessment Scale

https://doi.org/10.1016/S0022-3999(02)00392-6Get rights and content

Abstract

Objective: The main aim of this study was to examine the dimensionality and psychometric qualities of a new 10-item fatigue measure, the Fatigue Assessment Scale (FAS). Methods: As part of a longitudinal study, the respondents, all workers with at least 20 working hours per week, completed the FAS, four related fatigue measures, a depression questionnaire, and an emotional stability scale. Results: The FAS had a high internal consistency. The pattern of correlations and factor analysis showed good convergent and divergent validity. The FAS correlated strongly with the other fatigue scales. In a factor analysis of the five fatigue questionnaires, the FAS had the highest factor loading on a clear one-factor solution. Moreover, factor analyses revealed that fatigue, on the one hand, and depression and emotional stability, on the other hand, are separate constructs. Finally, it was shown that 8 out of the 10 FAS items were unbiased concerning gender; two had a uniform bias. Conclusions: The FAS represents a potentially valuable assessment instrument with promising internal consistency reliability and validity. Gender bias in the FAS does not have consequences for use of the FAS.

Introduction

The first objective of this study was to examine the psychometric qualities of a new fatigue measure, the Fatigue Assessment Scale (FAS) [1]. The second objective was to analyze possible gender and age differences and to test for the existence of gender bias.

Fatigue is a nonspecific symptom that is highly prevalent among patients in primary health care (e.g., [2], [3], [4]). It is an important component of many physical diseases and psychiatric disorders. For instance, fatigue is one of the most pervasive symptoms experienced by patients suffering from chronic diseases like cancer [5] and multiple sclerosis [6]. Hence, several, often multidimensional, fatigue questionnaires have been developed for specific populations such as cancer patients [5], [7], [8] and multiple sclerosis patients [6]. Fatigue also plays a substantial role in the healthy population. Severe fatigue during a relatively long period can lead to sick leave and work disability. For example, in the Netherlands, over one-third of the recipients of work disability benefit is occupationally disabled on mental grounds [9]. The majority of these individuals suffer from chronic job stress and burn-out. The most characteristic component of burn-out [10] is emotional exhaustion, a fatigue-related concept. Several measures of fatigue are claimed to be useful in patient populations as well as in healthy individuals [11], [12].

Due to the fast growing number of persons suffering from chronic fatigue syndrome in the nineties, interest in fatigue has expanded considerably. This has led to an intense debate about the conceptualization of fatigue, as well as its determinants, manifestations, and direct and indirect consequences. One vehemently debated issue is the dimensionality of fatigue. Nowadays, there is a tendency to claim that fatigue is best conceived of as a multidimensional construct [13], [14]. However, so far, there is no convincing evidence for this view [1]. Statements regarding the multidimensionality of fatigue are based predominantly on the outcomes of factor analyses using the criterion of eigenvalues greater than 1.0 as indicator in order to choose the number of factors (e.g., [11], [15], [16]). However, this particular criterion greatly overestimates the number of factors and often causes factors to split into bloated specifics (e.g., [17], [18]). Other studies have used confirmatory factor analyses to examine the dimensionality of fatigue (e.g., [12], [13]) and claim a good fit for a multidimensional model. Smets et al. [12], however, did not examine whether a one-factor solution would have fit their data equally well. Furthermore, Åhsberg [13] pointed to lack of energy as a general latent factor that represented much of the common variance in items also assessing physical exertion, physical discomfort, lack of motivation, and sleepiness. In line with these investigations, two recent studies examined the dimensionality of fatigue by factor analyzing broad sets of multidimensional fatigue questionnaires [1], [19]. Neither exploratory factor analyses supported the differentiation of fatigue in cognitive, emotional, somatic, and general aspects of fatigue. Instead, clear one-factor solutions were found in a healthy population [1], [19], as well as in a group of chronic pain patients [19].

Consequently, the FAS, a measure of chronic fatigue, was developed [1]. The initial item pool consisted of 40 items taken from four commonly used fatigue questionnaires: the Fatigue Scale (FS) [11]; the Checklist Individual Strength (CIS) [20], the Emotional Exhaustion (EE) subscale of the Dutch version of the Maslach Burnout Inventory (MBI-NL) [10], and the Energy and Fatigue subscale of the World Health Organization Quality of Life assessment instrument (WHOQOL-EF) [21]. A semantical analysis [21] was done in order to guide the selection of items from this item pool. Nine semantical groups were distinguished. One extra group was added in order to have an even number of items representing mental fatigue and physical fatigue. The initial objective was not to develop a one-dimensional scale. At the end of the construction process, the FAS consisted of 10 items (see Appendix A). The first examination of the psychometric qualities of the FAS demonstrated high reliability. Furthermore, factor analysis revealed that the FAS measured one construct.

With regard to age differences in relation to chronic fatigue, the psychological literature is rather equivocal. Some researchers have found a sizeable effect of age on fatigue [22], while others have reported only weak associations or even failed to observe any difference [23], [24], [25]. For instance, David et al. [26] have reported a positive, but low, correlation between age and fatigue, taking duration of fatigue into account.

In a comprehensive review article, Lewis and Wessely [24] have demonstrated convincingly that women report fatigue two to three times more often than men. Similar results were obtained in other studies (e.g., Ref. [27]). In constrast, a sizeable number of studies did not contain such outcomes [26], [28]. However, these differences can be caused by items with gender bias [29]. An item is an unbiased measure of a theoretical construct (e.g., fatigue) if persons from different groups (e.g., males and females), who are equally tired, have the same average score. To date, no systematic research has been done to examine such bias in fatigue items. However, without checking item bias, it remains unclear whether results documenting gender differences in fatigue reflect true mean differences, gender item bias, or a combination of both.

The main aims of this study were to check the dimensionality of the FAS and to examine its reliability and validity. In order to study the validity of the FAS, four additional fatigue questionnaires, a depression scale, and an emotional stability scale were examined in relation to the FAS. The internal consistency of the FAS was expected to be high and the FAS was expected to be unidimensional. With regard to convergent validity, it was anticipated that the FAS would have high associations with related fatigue measures, even when correcting for overlap in items. In addition, it was expected that a factor analysis of the FAS and other fatigue questionnaires would show one factor. Concerning divergent validity, fatigue, depression, and emotional stability were assumed to be different constructs. In addition, gender and age differences were examined. Finally, gender item bias was explored.

Section snippets

Subjects

Randomly selected subjects, after receiving a telephone call, agreed to complete a number of questionnaires as part of a study with five measurement points. This prospective study focused on a population with a minimum employment of 50%. The data presented here were collected at the last measurement time point. Three hundred and fifty-one persons (55%) out of a group of 635 returned a completed test booklet, 183 men (M=45 years; S.D.=8.39) and 166 women (M=43 years; S.D.=9.50). The gender of

Results

In Table 1, the means, standard deviations, and Cronbach's alpha coefficients of the various scales are presented. The internal consistency of the FAS was .90. Exploratory factor analysis of the FAS items showed a unique factor supported by the scree plot. The factor loadings varied from .82 (“I am bothered by fatigue”) to .55 (“When I am doing something, I can concentrate quite well”). The factor explained 53% of the variance. In addition, a factor analysis of the FAS and the eight subscales

Discussion

The FAS has good internal consistency. In addition, factor analysis and Mokken scale analysis provided strong evidence for the unidimensionality of the FAS. Moreover, in a factor analysis of a set of well-established fatigue instruments, the FAS had the highest factor loading on a one-factor solution. In an earlier study [1], the reliability of the FAS appeared to be good for the general Dutch population. This initial evaluation also supported a unidimensional conceptualization. In the present

Acknowledgements

The present study was sponsored by grants from (i) The Netherlands Organization for Scientific Research (NWO), Research Programme “Fatigue at Work” Grant no: 580-02-204 and (ii) WORC, research institute of Tilburg University.

References (50)

  • Michielsen HJ, De Vries J, Van Heck GL, Van de Vijver AJR, Sijtsma K. Examination of the dimensionality of fatigue: the...
  • DW Bates et al.

    Prevalence of fatigue and chronic fatigue syndrome in a primary care practice

    Arch Intern Med

    (1993)
  • J Bensing et al.

    Vermoeidheid: een chronisch probleem

    Med Contact

    (1996)
  • M Foets et al.

    Een nationale studie van ziekten en verrichtingen in de huisartspraktijk. Basisrapport gezondheid en gezondheidsgedrag in de praktijkpopulatie

    (1991)
  • LB Krupp et al.

    The Fatigue Severity Scale: application to patients with multiple sclerosis and systemic lupus erythematosus

    Arch Neurol

    (1989)
  • DM Hann et al.

    Measurement of fatigue in cancer patients: development and validation of the fatigue symptom inventory

    Qual Life Res

    (1998)
  • BF Piper et al.

    The revised Piper Fatigue Scale: psychometric evaluation in women with breast cancer

    Oncol Nurs Forum

    (1998)
  • W Schaufeli et al.

    Burnout, een begrip gemeten. De Nederlandse versie van de Maslach Burnout Inventory (MBI-NL)

    Gedrag Gezond

    (1994)
  • E Åhsberg

    Dimensions of fatigue in different working populations

    Scand J Psychol

    (2000)
  • EMA Smets et al.

    Het vaststellen van vermoeidheid: de Multidimensionele Vermoeidheids Index (MVI)

    Gedrag Gezond

    (1995)
  • H Vertommen et al.

    Vermoeidheid: van onhanteerbaar symptoom tot diagnostisch waardevolle gemoedstoestand

    Tijdschr Klin Psychol

    (1988)
  • RJ Rummel

    Applied factor analysis

    (1970)
  • P Kline

    Factor analysis and personality theory

    Eur J Pers

    (1987)
  • JL Studts et al.

    Symptom structure of fatigue: a multidimensional or unidimensional construct for behavioral medicine?

    Psychosom Med

    (2001)
  • Cited by (518)

    View all citing articles on Scopus
    View full text