The assessment of fatigue: Psychometric qualities and norms for the Checklist individual strength

[Tom Kindlon]

http://www.sciencedirect.com/science/article/pii/S0022399917301125#MMCvFirst

The assessment of fatigue: Psychometric qualities and norms for the Checklist individual strength
M. Worm-Smeitinka, M. Gielissenb, L. Blootc, H.W.M. van Laarhovend, B.G.M. van Engelene,
P. van Rielf, G. Bleijenbergg, S. Nikolausa, H. Knoopa,b,⁎

a Expert Centre for Chronic Fatigue, Department of Medical Psychology, Amsterdam Public Health Research Institute, VU University Medical Center, Amsterdam, The
Netherlands
b Academic Medical Center (AMC), University of Amsterdam, Department of Medical Psychology, Amsterdam Public Health Research Institute, Amsterdam, The
Netherlands
c Department of Medical Psychology, Isala Klinieken, Zwolle, The Netherlands
d Department of Medical Oncology, Academic Medical Center (AMC), University of Amsterdam, Amsterdam, The Netherlands
e Department of Neurology, Radboud university medical center, Nijmegen, The Netherlands
f IQ healthcare, Radboud University Medical Center, Nijmegen, The Netherlands
g Radboud University Medical Center, The Netherlands

https://doi.org/10.1016/j.jpsychores.2017.05.007

Highlights


•
The Checklist Individual Strength (CIS), an instrument used to assess fatigue, is a reliable and valid measure of fatigue.

•
Cut-off scores for severe fatigue are presented that can be used in clinical practice and research.

•
Norms of CIS scores for the general population and eight different medical condition are presented.



Abstract
Objective
The Checklist Individual Strength (CIS) measures four dimensions of fatigue: Fatigue severity, concentration problems, reduced motivation and activity.

On the fatigue severity subscale, a cut-off score of 35 is used.
This study
1) investigated the psychometric qualities of the CIS;
2) validated the cut-off score for severe fatigue
and
3) provided norms.

Methods

Representatives of the Dutch general population (n = 2288) completed the CIS.

The factor structure was investigated using an exploratory factor analysis. Internal consistency and test-retest reliability were determined.

Concurrent validity was assessed in two additional samples by correlating the CIS with other fatigue scales (Chalder Fatigue Questionnaire, MOS Short form-36 Vitality subscale, EORTC QLQ-C30 fatigue subscale).

To validate the fatigue severity cut-off score, a Receiver Operating Characteristics analysis was performed with patients referred to a chronic fatigue treatment centre (n = 5243) and a healthy group (n = 1906).

Norm scores for CIS subscales were calculated for the general population, patients with chronic fatigue syndrome (CFS; n = 1407) and eight groups with other medical conditions (n = 1411).

Results

The original four-factor structure of the CIS was replicated.

Internal consistency (α = 0.84–0.95) and test-retest reliability (r = 0.74–0.86) of the subscales were high.

Correlations with other fatigue scales were moderate to high.

The 35 points cut-off score for severe fatigue is appropriate, but, given the 17% false positive rate, should be adjusted to 40 for research in CFS.

Conclusion
The CIS is a valid and reliable tool for the assessment of fatigue, with a validated cut-off score for severe fatigue that can be used in clinical practice.

Keywords
Patient-reported outcomes
Fatigue questionnaire
Reliability and validity
Psychometrics
Chronic fatigue
Checklist individual strength

 

[Tom Kindlon]

I am unable to access the supplemental data file. If anyone could get the file and either attach it to a message in this thread, send it to me as a private message or alternatively email it to me at Tom Kindlon at Gmail dot com, I would appreciate it.
Somebody has now kindly sent me the file.

 

[Dolphin]

2.1.2. Cut-off score for severe fatigue

To determine the cut-off score for CIS Fatigue Severity subscale we used the following groups:

1. People referred between 2000 and 2016 to a tertiary treatment center for chronic fatigue at the Radboud University Medical Center for being severely fatigued (n = 5243). All completed the CIS as part of a routine screening at the start of their diagnostic process.

That is a large group of patients for one centre to see.

2. A healthy subgroup of the aforementioned sample from the general population, who reported no sick days in the past month, with people younger than 18 years (n =17) excluded (total n = 1906). The latter was done to match the data with the first group, that only contains adults.

 

[Dolphin]

In CFS patients the correlations between the CIS (subscales) and the CFQ (subscales) were all moderate (Table 2). Correlations with the SF- 36 vitality scale were moderate for the CIS subscale Fatigue and high for the CIS total score. In cancer survivors the CIS subscale Fatigue Severity correlated high to the EORTC QLQ-C30 subscale fatigue.

Correlations between the CIS and other fatigue measures showed a high correlation of the CIS with two of the alternative fatigue measures, which reflects good concurrent validity. With the third fatigue measure, the Chalder fatigue scale, however, the CIS showed only moderate correlations. This reflects the challenge of measuring a complex multifacetted subjective phenomenon such as fatigue. There is no gold standard and no two fatigue questionnaires measure the exact same thing [12]. It cannot be assumed that these alternative questionnaires can be seen as a gold standard. Of the three alternative measures for fatigue the associations were the lowest with the Chalder Fatigue Questionnaire. The correlation of the CIS Fatigue Severity subscale with the Chalder Fatigue Questionnaire total score was remarkably low. These correlations would probably have been higher when they would not have been calculated in a group of CFS patients due to a restriction of range effect. Since this group was selected for being severely fatigued, i.e. patients with a CIS score < 35 were not included. This has restricted the range of scores and could have reduced the correlation. When calculated in screening data (not selected on Fatigue Severity of> 35), the correlations were moderate (0.45) but still relatively low [17]. This suggests that the measured constructs overlap to some extent, but are certainly not identical. For example, fatigue severity may not be the sum of physical and mental fatigue. This is important, as both scales are frequently used as outcome measure in studies investigating the effects of treatment on fatigue. If these questionnaires measure different aspects of fatigue, the treatment effects should probably not be compared directly.

An option would be to use more than one measure of fatigue to adequately assess the effect of interventions on different aspects of fatigue (see Janse et al., submitted, for an example [49]). A disadvantage is that this leads to more items to complete for the patients. A solution may be computer adaptive testing. In a computer adaptive test (CAT), items are automatically selected from a large item pool, on the basis of the answers on previous items. The item pool is scaled using item response theory (IRT). According to IRT each item has a unique place on the continuum of the underlying construct measured, for example fatigue severity. As this method enables the estimation of item characteristics for each item independently, it is possible to select the most informative items for an individual patient. It is even possible to compare scores when patients have not filled in the same items. The advantage of CAT is increased measurement precision with fewer items. The Patient-Reported Outcomes Measurement Information System (PROMIS) initiative developed item banks for a range of health outcomes. The fatigue item bank was developed after an extensive item reviewing process [50] and has been found reliable and valid [51]. A CAT specifically for measuring fatigue in rheumatoid arthritis is already being developed [52].

I thought this was a clever idea. I don't recall seeing it being explicitly mentioned in the CFS literature that I have read before.

 

[Dolphin]

3.3. Population norms

The norms are shown in Appendix B.

For Fatigue Severity the general population scores 23.0 on average.

The means of the patient groups range from 26.7 to 40.4.

The mean of the CFS population is 50.4

It's a pity they don't give more information on this in the main text

 

[Dolphin]

The second objective was the evaluation of the cut-off for severe fatigue using the CIS. The analyses showed that at a cut-off of 35 on the subscale Fatigue Severity, only 2% of people with severe fatigue would not have been identified as severely fatigued, but 17% of the nonfatigued population will be incorrectly identified as severely fatigued.

For the CFS patient group, for the purpose of research, the cut-off of 40 is advised. It is then certain that a subgroup of severely fatigued patients is selected. For the purpose of diagnostics in clinical care, we advise for CIS Fatigue Severity scores between 35 and 40 to use the clinical interview to determine if treatment aimed at fatigue is necessary.

 

[Dolphin]

I thought this was interesting:

Furthermore, measurement invariance was not investigated. It is not known if people from different patient groups answer the questions differently. For example a higher score on activity for one patient group may reflect other problems than for another patient group. Future research could investigate this.

 

[Snow Leopard]

This study kind of suggests that the Chalder Fatigue Scale is rubbish, even if well, they don't say it explicitly.

 

[Sad Dad]

I asked my daughter, who currently has mild CFS, to take the survey. She scored 47 on the Fatigue scale which puts her in the 25th percentile of their CFS patients.

However, I think some of the other scales have a lot of bias depending on what you are used to doing, which is probably a function of how long you have had CFS. For example, she was midway between the 50th and 75th percentiles on the Activity scale and above the 50th percentile on the Concentration scale.

 

[A.B.]

The question how useful a fatigue questionnaire even is remains. They're worthless for detecting or measuring post-exertional malaise I would say because there is no time or exertion dimension.

 

[Tom Kindlon]

 

[Barry53]

Can someone advise me here. Is there a definitive, well accepted definition of what fatigue actually is. Is it a physical phenomenon? A mental perception? Does the medical profession have a commonly agreed understanding of what it actually is?

 

[Invisible Woman]

Can someone advise me here. Is there a definitive, well accepted definition of what fatigue actually is. Is it a physical phenomenon? A mental perception? Does the medical profession have a commonly agreed understanding of what it actually is?

Not as far as I'm aware. When I'm asked about it by a doc I usually ask them to define what they mean by the word "fatigue". That aways seems to take em aback.