291
Psychiatry Research, 36:291-298
Elsevier
Validity and Reliability of a Scale to Assess Fatigue Kathryn A. Lee, Gregory Hicks, and German
Nino-Murcia
Received April 20, 1990; revised version received September 1990.
17, 1990; accepted November
2,
Abstract. A visual analogue scale to evaluate fatigue severity (VAS-F) was developed and tested in a sample of 75 healthy individuals and a sample of 57 patients undergoing medical evaluation for sleep disorders. The scale consists of 18 items related to fatigue and energy, has simple instructions, and is completed with minimal time and effort. The VAS-F compares favorably with the Stanford Sleepiness Scale and the Profile of Mood States, and its internal consistency reliabilities are high. Healthy subjects demonstrated significant differences between their evening and morning scores on the VAS-F, while sleep-disordered patients did not. Key Words. Fatigue, energy, visual analogue scales, sleep disorders.
Fatigue is a phenomenon of concern to health care providers because of its implications for safety in the work setting (Yoshitake, 1978) and because of its prevalence among persons with physical or emotional illness (Solberg, 1984; Sugarman and Berg, 1984; Kroenke et al., 1988; Kruesi et al., 1989; Krupp et al., 1989). The prevalence of fatigue in the general U.S. population is estimated to be between 14% and 20% (Chen, 1986). From a physiologist’s perspective, fatigue can be defined as the end result of excessive energy consumption, depleted hormones, or diminished ability of muscle cells to contract. Anemia, infection, impaired oxygenation, and other physiological conditions deplete energy reserves by creating an unrelenting physical demand for energy expenditure. From a psychiatric perspective, fatigue can be defined as a subjective state of weariness related to reduced motivation, prolonged mental activity, or boredom that occurs in situations such as chronic stress, anxiety, or depression. Clinical intervention for patients who complain of fatigue is related to resolution of the physiological or psychiatric problem. With fatigue being an important indicator of illness or responsiveness to medical intervention, valid and reliable measures for perception of fatigue and its severity are important to both researchers and clinicians. Yet, quantitative measures for estimating subjective perceptions are difficult to construct. Instead of obtaining an
Kathryn A. Lee, Ph.D, R.N., is Assistant Professor in the Department of Family Health Care Nursing at the University of California, San Francisco. Gregory Hicks, M.A., was Head Technician at Stanford Sleep Clinic and is now a Lecturer at San Jose State University. German Nino-Murcia, M.D., is now Director of the Sleep Medicine and Neuroscience Institute in Palo Alto, CA. (Reprint requests to Dr. K.A. Lee, N41 lY, Box 0606, Dept. of Family Health Care Nursing, University of California, San Francisco, San Francisco, CA 94143-0606, USA.) 0165-1781/91/$03.50
@ 1991 Elsevier Scientific
Publishers
Ireland
Ltd.
292 estimate of fatigue perception itself, various instruments have been developed to measure consequences of fatigue, impact of fatigue on performance, or fatigue as a component of depression or other symptomatology. Often fatigue is measured using only one item, with individuals asked to rate their level of fatigue using a Likert-type scale. A one-item measure can be a very unstable estimate of any subjective phenomenon. While a scale that measures fatigue should use descriptive adjectives that represent both extremes of a continuum, some Likert-type scales consist of adjectives at each extreme that may not represent the same continuum (i.e., from “fit” or “in excellent shape” to “exhausted”). An instrument to estimate perceived fatigue should be easy to administer, completed in a short period of time without adding to the fatigue already present, and easy to score. Since a visual analogue line contains no numbers and has few words, individuals can read and mark a line in a few seconds. Folstein and Luria (1973) demonstrated the usefulness of visual analogue scales to measure subjective mood states in psychiatric patients. Monk (1989) has demonstrated the advantages of visual analogue scales to measure vigor and affective state in both healthy and depressed individuals. A valid and reliable instrument using visual analogue scales to estimate fatigue should be useful in assessing individual changes in fatigue over time or comparing perceptions of fatigue across patient populations. Methods To assess the therapeutic effect of clinical interventions for individuals with fatigue, an 18-item visual analogue scale was developed and tested in healthy adults with no complaints of chronic fatigue or sleep disorders and in patients being evaluated at a sleep disorders clinic with a major complaint of fatigue. The original instrument was four pages long and contained 37 visual analogue lines measuring 100 mm in length. Each line had bipolar end anchors related to descriptors obtained from the literature and from content analysis of personal interviews with patients complaining of fatigue. These descriptors included aspects of energy and vigor as well as fatigue. After seeing an example of how to complete a visual analogue line, subjects were asked to place a mark along each of the 37 lines to indicate how they currently felt. Since there were very few words to read, subjects were able to complete the instrument in < 2 min. To establish the validity of this visual analogue scale for fatigue (VA!%F), subjects were also asked to complete the Stanford Sleepiness Scale (SSS) and the Profile of Mood States (POMS). The SSS (Hoddes et al., 1973) consists of one question in which subjects rate their degree of sleepiness on ascale of 1 (not sleepy) to 7 (very sleepy). The SSS was included at the end of the VAS-F instrument. The POMS consists of 65 items rated on a 0 (not at all) to 4 (extremely) scale and contains subscales specifically related to fatigue-inertia and vigoractivity as well as depression-dejection, tension-anxiety, anger-hostility, and confusionbewilderment (McNair et al., 1971). The POMS can be completed in < 5 min. To control for effects of fatigue, a coin was tossed to determine the order in which subjects were to complete the POMS and VAS-F. The instruments were then stapled together in the order predetermined by the coin toss. Analysis. Data were collected before and after one night’s sleep from 75 healthy male and female subjects ranging in age from 18 to 55 years, and from 57 male and female patients being evaluated for a sleep disorder. Data from the 75 healthy subjects were used to establish interna; consistency reliabilities, based on Cronbach’s alpha (a) coefficients, for the fatigue and energy subscales of the VAS-F. The final form of the instrument, consisting of 18 reliable items from the normal group analysis, was tested for internal consistency reliability in the
293 patient group. Pearson correlations were used to establish the concurrent validity of the final VAS-F instrument, with its fatigue and energy subscales, in relation to the SSS and the two POMS subscales for vigor-activity and fatigue-inertia. To determine the sensitivity of the 18-item VAS-F in detecting within-subject changes in fatigue and energy before and after a night of sleep, paired t tests were used. Results Reliability. The original 37 items were categorized, a priori, as belonging to the fatigue subscale or the energy subscale. Data from the 75 healthy subjects were used to establish the instrument’s internal consistency reliability. In the determination of reliabilities for the two subscales, items with an interitem correlation coefficient of < 0.5 were deleted from each subscale. Although all items were originally derived from the fatigue literature and from individuals’ descriptions of fatigue, not all items were internally consistent within the two subscales. Nineteen items were deleted from the original instrument. Deleted items consisted of slang terms that subjects asked to have further explained to them during completion of the instrument (i.e., “not at all peppy” to “full of pep”) and descriptors related to aspects of anxiety and relaxation rather than fatigue and energy. The final instrument (see Table l), consisting of 13 items in the fatigue subscale and 5 items in the energy subscale, was then tested for internal consistency reliability in the group of 57 patients. Cronbach’s a for the 13-item fatigue subscale was 0.91 (evening data) and 0.96 (morning data) for the healthy group. Cronbach’s a for the patient sample was 0.96 and 0.95, respectively. Cronbach’s a for the 5-item energy subscale was 0.94 (evening) and 0.95 (morning) for the healthy group, and 0.96 (evening) and 0.94 (morning) for the patient sample. The fatigue and energy subscale evening scores were correlated at -0.54 for the normal group and -0.73 for the patient group. Validity. Pearson correlations were used to establish concurrent validity of the final VAS-F instrument with both the SSS and the POMS. Correlations > 0.30 (p < 0.01) were considered significant for establishing validity. Tables 2 and 3 present these correlations for the group of normal subjects and the group of patients, respectively. Of particular interest are the strong associations between the evening data for the VAS-F fatigue subscale and the morning data for the VAS-F energy subscale. The VAS-F fatigue subscale scores in the evening before sleep onset were more highly correlated with SSS and POMS evening scores than with morning scores. The opposite pattern was evident for the VAS-F energy subscale: morning scores after awakening from sleep were more highly correlated with SSS and POMS morning scores than evening scores. Specificity. To establish the ability of the VAS-F to distinguish fatigue and energy from other aspects of mood state, associations with the other four subscales of the POMS were also tested statistically. The VAS-F fatigue and energy subscales were not significantly correlated with POMS subscales related to depression, tension, or anger (see Table 2) in the normal subjects. There was a significant relationship @ < 0.01) between the POMS confusion-bewilderment subscale and the VAS-F fatigue subscale (r = 0.41) morning scores. There was also a weak relationship
294 Table 1. Visual Analogue Scale for Fatigue Time
a.m.
p.m.
ID #
Date
I am trying to find out about your level of energy before and after your night of sleep. There are 18 items I would like you to respond to. This should onlytake about 1 minute of your time. Thank you. DIRECTIONS: You are asked to place an “X” through these lines to indicate how you are feeling RIGHT NOW For example, suppose you have not eaten since yesterday. Where would you put the “X” on the line below? not atall extremely hungry hungry You would probably put the “X” closer to the “extremely This is where I put it: not at all hungry
hungry” end of the line. exaemely hungry
NOW PLEASE COMPLETE THE FOLLOWING ITEMS. not at all tired
extremely tired
not at all sleepy
extremely sleepy
not at all drowsy
extremely drowsy
not at all fatigued
extremely fatigued
not at all worn out
extremely worn out
not at all energetic
extremely energetic
not at all actlve
extremely active
not at all “lgoro”s
extremely vigorous
not at all efficient
extremely efficient
not at all lovely
extremely lively
not at all bushed
totally bushed
not at all exhausted
totally exhausted
keeping my eyes open 1sno effori at all
keeping my eyes open ISa tremendous chore
moving my body ISno effolt at all
moving my body ISa tremendous chore
concentrating IS no effori at all
concentrating ISa tremendous chore
carrymg on a convarsatlon is no effort at all
carrying on a conversation ISa tremendous chore
I have absolutely no desire to close my eyes
I have a tremendous desire to close my eyes
I have a tremendous absolutely no desire to desire to lie down lie down Items I -5 and 11-18 belong to the fatigue subscale. Items 6-10 belong to the energy subscale. With actual use, the horizontal lines should be exactly 100 mm. This 2-page instrument is available upon request. I have
295
Table 2. Concurrent Validity of the la-item Visual Analogue Scale for Fatigue (VAS-F) with the Stanford Sleepiness Scale (SSS) and the Profile of Mood States (POW) in healthy subjects (n = 75) POMS VAS-I=
SSS
F
V
C
D
T
A
Energy subscale (5 items) morning
-0.76’
-0.63’
0.80’
-0.30
evening
-0.54’
-0.36’
0.45’
-0.06
morning
0.81’
0.78’
-0.73’
evening
0.65’
0.58’
-0.32’
-0.12
-0.19
-0.13
0.02
-0.03
0.08
0.41’
0.22
0.22
0.23
0.29
0.12
0.09
0.07
Fatigue subscale (13-items)
Note. F = fatigue-inertia. V = vigor-activity. anxiety. A = anger-hostility.
C = confusion-bewilderment.
D = depression-dejection.
T = tension-
1. Significant at p < 0.01.
@ < 0.05) between the POMS confusion-bewilderment subscale and the VAS-F energy subscale (r = -0.30) and the SSS (r = 0.30). The confusion-bewilderment subscale of the POMS consists of seven adjectives: confused, unable to concentrate, muddled, bewildered, efficient, forgetful, and uncertain about things. The VAS-F contains a visual analogue line with end anchors ranging from “not at all efficient” to “extremely efficient” in the energy subscale and a line with end anchors ranging from “concentrating is no effort at all” to “concentrating is a tremendous chore” in the fatigue subscale.
Table 3. Concurrent Validity of the la-item Visual Analogue Scale to Measure Fatigue Severity (VAS-F) with the Stanford Sleepiness Scale (SSS) and the Profile of Mood States (POMS) in patients with sleep disorders (n = 57) POMS VAS-F
SSS
F
V
C
D
T
A
-0.83’
-0.73’
0.56’
-0.40’
-0.10
-0.18
-0.23
0.70’
-0.59’
0.56’
0.35
0.10
Energy subscale (5 items] evening Fatigue subscale (13-items) evening
0.73’
Note. F = fatigue-inertia. V = vigor-activity. anxiety. A = anger-hostility.
C = confusion-bewilderment.
D = depression-dejection.
0.06 T = tension-
1. Significant at p < 0.01.
Sensitivity. To determine the sensitivity of the VAS-F in detecting within-subject changes in fatigue and energy before and after sleep, paired t tests were used to test for differences between the mean scores. While the patients had no significant differences between morning and evening, the normal subjects showed significant differences between morning and evening on all measures. Table 4 presents the
296
Table 4. Differences between evening and morning VAS-F, SSS, and POW mean scores (k SD) in healthy subjects (n = 75) Scale
Evening
Morning
VAS-F Fatigue
55.6 (18.74)
31.6 (23.22)’
Energy
34.9 (19.74)
60.3 (23.92)’
4.0
( 1.59)
2.4
(
Fatigue-inertia
12.1 10.7
( 6.63) ( 6.25)
5.5
Vigor-activity
16.8
( 5.72)’ ( 7.43)’
sss
1.36)’
POMS
Note. VAS-F = Visual Analogue Scale to Measure Fatigue Severity. SSS = Stanford Sleepiness
Scale (SSS).
POMS = Profile of Mood States. 1. Paired t tests, r > 3.6, p < 0.001.
means (-+ SD) for the group of healthy subjects, group of patients with sleep disorders.
and Table 5 contains
the data for the
Table 5. Differences between evening and morning VAS-F, SSS, and POMS mean scores (31 SD) in patients evaluated for sleep disorders (n = 21) Scale
Evening
Morning
VAS-F Fatigue
24.6 (45.89)
26.4 (35.40)
Energy
49.9 (23.30)
48.8 (38.90)
sss
2.7
( 1.30)
( 6.26)
5.2
14.5 ( 6.49)
14.8
( 5.27) ( 6.33)
3.0
(
7.4
1.64)
POMS Fatigue-inertia Vigor-activity
Note. VAS-F = Visual Analogue Scale to Measure Fatigue Severity. SSS = Stanford Sleepiness Scale (SSS). POMS = Profile of Mood States.
Discussion This article describes a valid and reliable visual analogue instrument to assess the level of fatigue and energy in both normal and patient populations. There are many advantages and some limitations to the use of this type of instrument (Lee and Kieckhefer, 1989). Visual analogue lines are easily understood by subjects and require very little reading skill. This VAS-F was simple to administer and required little time for completion. Because of the minimal time needed to complete the VAS-F, there were no missing data in this study. Furthermore, a simple, brief instrument is essential when measuring fatigue and energy as the outcome variables of interest. Another advantage to the use of visual analogue lines is the ability to allow subjects to respond with as little bias and as much discrimination as they wish. Unlike Likert-type scales, these lines do not restrict subjects to arbitrary intervals, usually with a high score indicative of a positive response and a low score indicative of a negative response. With the VAS-F, data are obtained as continuous intervals
297
and, therefore, assumptions for parametric statistical analyses, such as a normal distribution of data points, are met (Aitken, 1969). Unlike the POMS, data from the VAS-F do not require any transformation processes to meet statistical assumptions. Unlike the SSS with its single item, the VAS-F subscales, with multiple items, should be more stable measurements of fatigue and energy. One limitation of any self-report instrument is the hesitation some persons have in using the extreme ends of lOO-mm lines in the case of visual analogue lines, or extreme values in a Likert-type scale. While this problem is not easily remedied in Likert-type scales, adjustments are easily made when lOO-mm lines are used. Maxwell (1978) recommends that sensitivity of responses be assured by converting the raw scores to a proportion of the maximum score indicated by each subject. In this study, statistical analyses included both the raw scores and the proportions, and results did not vary. A major limitation of the use of the VAS-F is the time-consuming measurement process for the research team. While minimal training is needed to score the VAS-F lines, interrater reliability must still be established. Because of the objective nature of the measure, reliability of scoring should be very high with minimal training. Caution must also be taken when photocopying the original instrument to assure that the IOO-mm line remains 100 mm in length. Another limitation to the use of VAS-F concerns the motor and visual abilities of the patient or subject. Patients with tremors, paralysis, or visual impairment are unable to complete any type of visual analogue scale reliably. Conclusion In most research, fatigue is measured either qualitatively by self-report or observation, or quantitatively by measurement of muscle cell activity or strength. The VAS-F has been shown to be a valid and reliable instrument for the quantitative assessment of fatigue and energy levels in both healthy subjects and patients who complain of poor sleep. It allows health care providers to assess fatigue in a rapid, quantitative manner. Potential uses include assessments of fatigue before and after clinical interventions as an indication of the effectiveness of therapy, and stable measurement of a subjective variable using continuous interval scales for research purposes. This study was funded Clinical Nurse Scholars Program.
Acknowledgment.
Foundation,
in part
by the
Robert
Wood
Johnson
References Aitken, Society
R.C.B.
Measurement
of Medicine,
62:989-996,
Chen, M.K. The epidemiology 15:74-81,
of feelings
using analogue
scales.
Proceedings
of the Royal
1969.
of self-perceived
fatigue among adults. Preventive
Medicine,
1986.
Folstein, M.F., and Luria, R. Reliability, validity, and clinical application of the visual analogue mood scale. Psychological Medicine. 3:479-486, 1973. Hoddes, E.; Zarcone, V.; Smythe, H.; Phillips, R.; and Dement, W.C. Quantification of
sleepiness: A new approach. Psychophysiology,
10:431-436,
1973.
298 Jerrett, W.A. Lethargy in general practice. Practitioner, 225:731-737, 1981. Kroenke, K.; Wood, D.R.; Mangelsdorff, A.D.; Meier, N.J.; and Powell, J.B. Chronic fatigue in primary care: Prevalence, patient characteristics, and outcome. Journal of the American Medical Association, 260~929-934, 1988. Kruesi, M.J.P.; Dale, J.; and Straus, S.E. Psychiatric diagnosis in patients who have chronic fatigue syndrome. Journal of Clinical Psychiatry, 50:53-56, 1989. Krupp, L.B, LaRocca, N.G.; Muir-Nash, J.; and Steinberg, A.D. The fatigue severity scale: Application to patients with multiple sclerosis and systemic lupus erythematosus. Archives of Neurology, 46:1121-l 123, 1989. Lee, K.A., and Kieckhefer, G.M. Measuring human responses using visual analogue scales. Western Journal of Nursing Research, 11: 128- 132, 1989. Maxwell, C. Sensitivity and accuracy of the visual analogue scale: A psycho-physical classroom experiment. British Journal of Clinical Pharmacology, 6: 15-24, 1978. McNair, D.M.; Lorr, M.; and Droppleman, L.F. Profile of Mood States Manual. San Diego, CA: Educational and Industrial Testing Service, 1971. Monk, T.M. A Visual analogue scale technique to measure global vigor and affect. Psychiatry Research, 27:89-99, 1989. Solberg, L.I. Lassitude, a primary care evaluation. Journal of the American Medical Association, 251:3272-3276, 1984. Sugarman, J.R., and Berg, A.O. Evaluation of fatigue in a family practice. Journnl of Family Practice, 19:643-647, 1984. Yositake, H. Three characteristic patterns of subjective fatigue symptoms. Ergonomics, 21:231-233. 1978.
Psychiatry Research, 36:291-298
Elsevier
Validity and Reliability of a Scale to Assess Fatigue Kathryn A. Lee, Gregory Hicks, and German
Nino-Murcia
Received April 20, 1990; revised version received September 1990.
17, 1990; accepted November
2,
Abstract. A visual analogue scale to evaluate fatigue severity (VAS-F) was developed and tested in a sample of 75 healthy individuals and a sample of 57 patients undergoing medical evaluation for sleep disorders. The scale consists of 18 items related to fatigue and energy, has simple instructions, and is completed with minimal time and effort. The VAS-F compares favorably with the Stanford Sleepiness Scale and the Profile of Mood States, and its internal consistency reliabilities are high. Healthy subjects demonstrated significant differences between their evening and morning scores on the VAS-F, while sleep-disordered patients did not. Key Words. Fatigue, energy, visual analogue scales, sleep disorders.
Fatigue is a phenomenon of concern to health care providers because of its implications for safety in the work setting (Yoshitake, 1978) and because of its prevalence among persons with physical or emotional illness (Solberg, 1984; Sugarman and Berg, 1984; Kroenke et al., 1988; Kruesi et al., 1989; Krupp et al., 1989). The prevalence of fatigue in the general U.S. population is estimated to be between 14% and 20% (Chen, 1986). From a physiologist’s perspective, fatigue can be defined as the end result of excessive energy consumption, depleted hormones, or diminished ability of muscle cells to contract. Anemia, infection, impaired oxygenation, and other physiological conditions deplete energy reserves by creating an unrelenting physical demand for energy expenditure. From a psychiatric perspective, fatigue can be defined as a subjective state of weariness related to reduced motivation, prolonged mental activity, or boredom that occurs in situations such as chronic stress, anxiety, or depression. Clinical intervention for patients who complain of fatigue is related to resolution of the physiological or psychiatric problem. With fatigue being an important indicator of illness or responsiveness to medical intervention, valid and reliable measures for perception of fatigue and its severity are important to both researchers and clinicians. Yet, quantitative measures for estimating subjective perceptions are difficult to construct. Instead of obtaining an
Kathryn A. Lee, Ph.D, R.N., is Assistant Professor in the Department of Family Health Care Nursing at the University of California, San Francisco. Gregory Hicks, M.A., was Head Technician at Stanford Sleep Clinic and is now a Lecturer at San Jose State University. German Nino-Murcia, M.D., is now Director of the Sleep Medicine and Neuroscience Institute in Palo Alto, CA. (Reprint requests to Dr. K.A. Lee, N41 lY, Box 0606, Dept. of Family Health Care Nursing, University of California, San Francisco, San Francisco, CA 94143-0606, USA.) 0165-1781/91/$03.50
@ 1991 Elsevier Scientific
Publishers
Ireland
Ltd.
292 estimate of fatigue perception itself, various instruments have been developed to measure consequences of fatigue, impact of fatigue on performance, or fatigue as a component of depression or other symptomatology. Often fatigue is measured using only one item, with individuals asked to rate their level of fatigue using a Likert-type scale. A one-item measure can be a very unstable estimate of any subjective phenomenon. While a scale that measures fatigue should use descriptive adjectives that represent both extremes of a continuum, some Likert-type scales consist of adjectives at each extreme that may not represent the same continuum (i.e., from “fit” or “in excellent shape” to “exhausted”). An instrument to estimate perceived fatigue should be easy to administer, completed in a short period of time without adding to the fatigue already present, and easy to score. Since a visual analogue line contains no numbers and has few words, individuals can read and mark a line in a few seconds. Folstein and Luria (1973) demonstrated the usefulness of visual analogue scales to measure subjective mood states in psychiatric patients. Monk (1989) has demonstrated the advantages of visual analogue scales to measure vigor and affective state in both healthy and depressed individuals. A valid and reliable instrument using visual analogue scales to estimate fatigue should be useful in assessing individual changes in fatigue over time or comparing perceptions of fatigue across patient populations. Methods To assess the therapeutic effect of clinical interventions for individuals with fatigue, an 18-item visual analogue scale was developed and tested in healthy adults with no complaints of chronic fatigue or sleep disorders and in patients being evaluated at a sleep disorders clinic with a major complaint of fatigue. The original instrument was four pages long and contained 37 visual analogue lines measuring 100 mm in length. Each line had bipolar end anchors related to descriptors obtained from the literature and from content analysis of personal interviews with patients complaining of fatigue. These descriptors included aspects of energy and vigor as well as fatigue. After seeing an example of how to complete a visual analogue line, subjects were asked to place a mark along each of the 37 lines to indicate how they currently felt. Since there were very few words to read, subjects were able to complete the instrument in < 2 min. To establish the validity of this visual analogue scale for fatigue (VA!%F), subjects were also asked to complete the Stanford Sleepiness Scale (SSS) and the Profile of Mood States (POMS). The SSS (Hoddes et al., 1973) consists of one question in which subjects rate their degree of sleepiness on ascale of 1 (not sleepy) to 7 (very sleepy). The SSS was included at the end of the VAS-F instrument. The POMS consists of 65 items rated on a 0 (not at all) to 4 (extremely) scale and contains subscales specifically related to fatigue-inertia and vigoractivity as well as depression-dejection, tension-anxiety, anger-hostility, and confusionbewilderment (McNair et al., 1971). The POMS can be completed in < 5 min. To control for effects of fatigue, a coin was tossed to determine the order in which subjects were to complete the POMS and VAS-F. The instruments were then stapled together in the order predetermined by the coin toss. Analysis. Data were collected before and after one night’s sleep from 75 healthy male and female subjects ranging in age from 18 to 55 years, and from 57 male and female patients being evaluated for a sleep disorder. Data from the 75 healthy subjects were used to establish interna; consistency reliabilities, based on Cronbach’s alpha (a) coefficients, for the fatigue and energy subscales of the VAS-F. The final form of the instrument, consisting of 18 reliable items from the normal group analysis, was tested for internal consistency reliability in the
293 patient group. Pearson correlations were used to establish the concurrent validity of the final VAS-F instrument, with its fatigue and energy subscales, in relation to the SSS and the two POMS subscales for vigor-activity and fatigue-inertia. To determine the sensitivity of the 18-item VAS-F in detecting within-subject changes in fatigue and energy before and after a night of sleep, paired t tests were used. Results Reliability. The original 37 items were categorized, a priori, as belonging to the fatigue subscale or the energy subscale. Data from the 75 healthy subjects were used to establish the instrument’s internal consistency reliability. In the determination of reliabilities for the two subscales, items with an interitem correlation coefficient of < 0.5 were deleted from each subscale. Although all items were originally derived from the fatigue literature and from individuals’ descriptions of fatigue, not all items were internally consistent within the two subscales. Nineteen items were deleted from the original instrument. Deleted items consisted of slang terms that subjects asked to have further explained to them during completion of the instrument (i.e., “not at all peppy” to “full of pep”) and descriptors related to aspects of anxiety and relaxation rather than fatigue and energy. The final instrument (see Table l), consisting of 13 items in the fatigue subscale and 5 items in the energy subscale, was then tested for internal consistency reliability in the group of 57 patients. Cronbach’s a for the 13-item fatigue subscale was 0.91 (evening data) and 0.96 (morning data) for the healthy group. Cronbach’s a for the patient sample was 0.96 and 0.95, respectively. Cronbach’s a for the 5-item energy subscale was 0.94 (evening) and 0.95 (morning) for the healthy group, and 0.96 (evening) and 0.94 (morning) for the patient sample. The fatigue and energy subscale evening scores were correlated at -0.54 for the normal group and -0.73 for the patient group. Validity. Pearson correlations were used to establish concurrent validity of the final VAS-F instrument with both the SSS and the POMS. Correlations > 0.30 (p < 0.01) were considered significant for establishing validity. Tables 2 and 3 present these correlations for the group of normal subjects and the group of patients, respectively. Of particular interest are the strong associations between the evening data for the VAS-F fatigue subscale and the morning data for the VAS-F energy subscale. The VAS-F fatigue subscale scores in the evening before sleep onset were more highly correlated with SSS and POMS evening scores than with morning scores. The opposite pattern was evident for the VAS-F energy subscale: morning scores after awakening from sleep were more highly correlated with SSS and POMS morning scores than evening scores. Specificity. To establish the ability of the VAS-F to distinguish fatigue and energy from other aspects of mood state, associations with the other four subscales of the POMS were also tested statistically. The VAS-F fatigue and energy subscales were not significantly correlated with POMS subscales related to depression, tension, or anger (see Table 2) in the normal subjects. There was a significant relationship @ < 0.01) between the POMS confusion-bewilderment subscale and the VAS-F fatigue subscale (r = 0.41) morning scores. There was also a weak relationship
294 Table 1. Visual Analogue Scale for Fatigue Time
a.m.
p.m.
ID #
Date
I am trying to find out about your level of energy before and after your night of sleep. There are 18 items I would like you to respond to. This should onlytake about 1 minute of your time. Thank you. DIRECTIONS: You are asked to place an “X” through these lines to indicate how you are feeling RIGHT NOW For example, suppose you have not eaten since yesterday. Where would you put the “X” on the line below? not atall extremely hungry hungry You would probably put the “X” closer to the “extremely This is where I put it: not at all hungry
hungry” end of the line. exaemely hungry
NOW PLEASE COMPLETE THE FOLLOWING ITEMS. not at all tired
extremely tired
not at all sleepy
extremely sleepy
not at all drowsy
extremely drowsy
not at all fatigued
extremely fatigued
not at all worn out
extremely worn out
not at all energetic
extremely energetic
not at all actlve
extremely active
not at all “lgoro”s
extremely vigorous
not at all efficient
extremely efficient
not at all lovely
extremely lively
not at all bushed
totally bushed
not at all exhausted
totally exhausted
keeping my eyes open 1sno effori at all
keeping my eyes open ISa tremendous chore
moving my body ISno effolt at all
moving my body ISa tremendous chore
concentrating IS no effori at all
concentrating ISa tremendous chore
carrymg on a convarsatlon is no effort at all
carrying on a conversation ISa tremendous chore
I have absolutely no desire to close my eyes
I have a tremendous desire to close my eyes
I have a tremendous absolutely no desire to desire to lie down lie down Items I -5 and 11-18 belong to the fatigue subscale. Items 6-10 belong to the energy subscale. With actual use, the horizontal lines should be exactly 100 mm. This 2-page instrument is available upon request. I have
295
Table 2. Concurrent Validity of the la-item Visual Analogue Scale for Fatigue (VAS-F) with the Stanford Sleepiness Scale (SSS) and the Profile of Mood States (POW) in healthy subjects (n = 75) POMS VAS-I=
SSS
F
V
C
D
T
A
Energy subscale (5 items) morning
-0.76’
-0.63’
0.80’
-0.30
evening
-0.54’
-0.36’
0.45’
-0.06
morning
0.81’
0.78’
-0.73’
evening
0.65’
0.58’
-0.32’
-0.12
-0.19
-0.13
0.02
-0.03
0.08
0.41’
0.22
0.22
0.23
0.29
0.12
0.09
0.07
Fatigue subscale (13-items)
Note. F = fatigue-inertia. V = vigor-activity. anxiety. A = anger-hostility.
C = confusion-bewilderment.
D = depression-dejection.
T = tension-
1. Significant at p < 0.01.
@ < 0.05) between the POMS confusion-bewilderment subscale and the VAS-F energy subscale (r = -0.30) and the SSS (r = 0.30). The confusion-bewilderment subscale of the POMS consists of seven adjectives: confused, unable to concentrate, muddled, bewildered, efficient, forgetful, and uncertain about things. The VAS-F contains a visual analogue line with end anchors ranging from “not at all efficient” to “extremely efficient” in the energy subscale and a line with end anchors ranging from “concentrating is no effort at all” to “concentrating is a tremendous chore” in the fatigue subscale.
Table 3. Concurrent Validity of the la-item Visual Analogue Scale to Measure Fatigue Severity (VAS-F) with the Stanford Sleepiness Scale (SSS) and the Profile of Mood States (POMS) in patients with sleep disorders (n = 57) POMS VAS-F
SSS
F
V
C
D
T
A
-0.83’
-0.73’
0.56’
-0.40’
-0.10
-0.18
-0.23
0.70’
-0.59’
0.56’
0.35
0.10
Energy subscale (5 items] evening Fatigue subscale (13-items) evening
0.73’
Note. F = fatigue-inertia. V = vigor-activity. anxiety. A = anger-hostility.
C = confusion-bewilderment.
D = depression-dejection.
0.06 T = tension-
1. Significant at p < 0.01.
Sensitivity. To determine the sensitivity of the VAS-F in detecting within-subject changes in fatigue and energy before and after sleep, paired t tests were used to test for differences between the mean scores. While the patients had no significant differences between morning and evening, the normal subjects showed significant differences between morning and evening on all measures. Table 4 presents the
296
Table 4. Differences between evening and morning VAS-F, SSS, and POW mean scores (k SD) in healthy subjects (n = 75) Scale
Evening
Morning
VAS-F Fatigue
55.6 (18.74)
31.6 (23.22)’
Energy
34.9 (19.74)
60.3 (23.92)’
4.0
( 1.59)
2.4
(
Fatigue-inertia
12.1 10.7
( 6.63) ( 6.25)
5.5
Vigor-activity
16.8
( 5.72)’ ( 7.43)’
sss
1.36)’
POMS
Note. VAS-F = Visual Analogue Scale to Measure Fatigue Severity. SSS = Stanford Sleepiness
Scale (SSS).
POMS = Profile of Mood States. 1. Paired t tests, r > 3.6, p < 0.001.
means (-+ SD) for the group of healthy subjects, group of patients with sleep disorders.
and Table 5 contains
the data for the
Table 5. Differences between evening and morning VAS-F, SSS, and POMS mean scores (31 SD) in patients evaluated for sleep disorders (n = 21) Scale
Evening
Morning
VAS-F Fatigue
24.6 (45.89)
26.4 (35.40)
Energy
49.9 (23.30)
48.8 (38.90)
sss
2.7
( 1.30)
( 6.26)
5.2
14.5 ( 6.49)
14.8
( 5.27) ( 6.33)
3.0
(
7.4
1.64)
POMS Fatigue-inertia Vigor-activity
Note. VAS-F = Visual Analogue Scale to Measure Fatigue Severity. SSS = Stanford Sleepiness Scale (SSS). POMS = Profile of Mood States.
Discussion This article describes a valid and reliable visual analogue instrument to assess the level of fatigue and energy in both normal and patient populations. There are many advantages and some limitations to the use of this type of instrument (Lee and Kieckhefer, 1989). Visual analogue lines are easily understood by subjects and require very little reading skill. This VAS-F was simple to administer and required little time for completion. Because of the minimal time needed to complete the VAS-F, there were no missing data in this study. Furthermore, a simple, brief instrument is essential when measuring fatigue and energy as the outcome variables of interest. Another advantage to the use of visual analogue lines is the ability to allow subjects to respond with as little bias and as much discrimination as they wish. Unlike Likert-type scales, these lines do not restrict subjects to arbitrary intervals, usually with a high score indicative of a positive response and a low score indicative of a negative response. With the VAS-F, data are obtained as continuous intervals
297
and, therefore, assumptions for parametric statistical analyses, such as a normal distribution of data points, are met (Aitken, 1969). Unlike the POMS, data from the VAS-F do not require any transformation processes to meet statistical assumptions. Unlike the SSS with its single item, the VAS-F subscales, with multiple items, should be more stable measurements of fatigue and energy. One limitation of any self-report instrument is the hesitation some persons have in using the extreme ends of lOO-mm lines in the case of visual analogue lines, or extreme values in a Likert-type scale. While this problem is not easily remedied in Likert-type scales, adjustments are easily made when lOO-mm lines are used. Maxwell (1978) recommends that sensitivity of responses be assured by converting the raw scores to a proportion of the maximum score indicated by each subject. In this study, statistical analyses included both the raw scores and the proportions, and results did not vary. A major limitation of the use of the VAS-F is the time-consuming measurement process for the research team. While minimal training is needed to score the VAS-F lines, interrater reliability must still be established. Because of the objective nature of the measure, reliability of scoring should be very high with minimal training. Caution must also be taken when photocopying the original instrument to assure that the IOO-mm line remains 100 mm in length. Another limitation to the use of VAS-F concerns the motor and visual abilities of the patient or subject. Patients with tremors, paralysis, or visual impairment are unable to complete any type of visual analogue scale reliably. Conclusion In most research, fatigue is measured either qualitatively by self-report or observation, or quantitatively by measurement of muscle cell activity or strength. The VAS-F has been shown to be a valid and reliable instrument for the quantitative assessment of fatigue and energy levels in both healthy subjects and patients who complain of poor sleep. It allows health care providers to assess fatigue in a rapid, quantitative manner. Potential uses include assessments of fatigue before and after clinical interventions as an indication of the effectiveness of therapy, and stable measurement of a subjective variable using continuous interval scales for research purposes. This study was funded Clinical Nurse Scholars Program.
Acknowledgment.
Foundation,
in part
by the
Robert
Wood
Johnson
References Aitken, Society
R.C.B.
Measurement
of Medicine,
62:989-996,
Chen, M.K. The epidemiology 15:74-81,
of feelings
using analogue
scales.
Proceedings
of the Royal
1969.
of self-perceived
fatigue among adults. Preventive
Medicine,
1986.
Folstein, M.F., and Luria, R. Reliability, validity, and clinical application of the visual analogue mood scale. Psychological Medicine. 3:479-486, 1973. Hoddes, E.; Zarcone, V.; Smythe, H.; Phillips, R.; and Dement, W.C. Quantification of
sleepiness: A new approach. Psychophysiology,
10:431-436,
1973.
298 Jerrett, W.A. Lethargy in general practice. Practitioner, 225:731-737, 1981. Kroenke, K.; Wood, D.R.; Mangelsdorff, A.D.; Meier, N.J.; and Powell, J.B. Chronic fatigue in primary care: Prevalence, patient characteristics, and outcome. Journal of the American Medical Association, 260~929-934, 1988. Kruesi, M.J.P.; Dale, J.; and Straus, S.E. Psychiatric diagnosis in patients who have chronic fatigue syndrome. Journal of Clinical Psychiatry, 50:53-56, 1989. Krupp, L.B, LaRocca, N.G.; Muir-Nash, J.; and Steinberg, A.D. The fatigue severity scale: Application to patients with multiple sclerosis and systemic lupus erythematosus. Archives of Neurology, 46:1121-l 123, 1989. Lee, K.A., and Kieckhefer, G.M. Measuring human responses using visual analogue scales. Western Journal of Nursing Research, 11: 128- 132, 1989. Maxwell, C. Sensitivity and accuracy of the visual analogue scale: A psycho-physical classroom experiment. British Journal of Clinical Pharmacology, 6: 15-24, 1978. McNair, D.M.; Lorr, M.; and Droppleman, L.F. Profile of Mood States Manual. San Diego, CA: Educational and Industrial Testing Service, 1971. Monk, T.M. A Visual analogue scale technique to measure global vigor and affect. Psychiatry Research, 27:89-99, 1989. Solberg, L.I. Lassitude, a primary care evaluation. Journal of the American Medical Association, 251:3272-3276, 1984. Sugarman, J.R., and Berg, A.O. Evaluation of fatigue in a family practice. Journnl of Family Practice, 19:643-647, 1984. Yositake, H. Three characteristic patterns of subjective fatigue symptoms. Ergonomics, 21:231-233. 1978.
