186
Test–retest reliability was adequate in general, although evidence was not available
for FHAQ or FSQ. Values for the intraclass correlation coeffi cient (ICC) for the FBI
range from 0.73 to 0.84 whereas for PROMs-FM its range was 0.89–0.96. Reliability
is expected to be measured by using ICC for continuous scales, ICC or weighted kappa
for ordinal scales, and unweighted kappa for nominal scales. However, reliability was
tested by means of correlation in the FIQ (Spearman correlation coeffi cient 0.85 [ 82 ]),
the FIQ-R (Pearson’s r = 0.83 in its Turkish version [ 71 ]), and the ICAF.
Validity
The validity of the original version of the FIQ was not tested—or we were unable to
fi nd any published study—whereas the revised version showed strong correlation
with the FIQ ( r = 0.88 p < 0.001) and SF-36’s physical function and pain subscales
( r = −0.80 and r = −0.60, respectively; note that correlations are negative due to the
fact that higher SF-36 scores relate to being healthier [ 70 ]). Total ICAF score shows
a moderate correlation with the FIQ ( r = 0.69) and HAQ ( r = 0.59). While testing the
construct validity of the FFS, Spearman correlation coeffi cient was computed
between the FFS and the different symptoms of FMS, as measured by VAS scales.
The correlations of the FFS items with pain score and the physical function subscale
of the SF-36 ranged from 0.28 to 0.32 [ 83 ]. There was a positive correlation between
the total FBI with the total King’s Health Questionnaire—a questionnaire that mea-
sures bladder and bowel problems. Individual correlations between the FBI and the
individual King’s Health Questionnaire’s domains ranged from 0.35 to 0.62 [ 73 ].
With regard to the global VASFIQ, its score correlates highly with FIQ scores at
baseline ( r = 0.94). Change in global VASFIQ and FIQ scores correlates similarly to
a Patients’ Global Impression of Change scale ( r = 0.58). Individual VASFIQ scores
correlate with corresponding full-length symptom questionnaire scores at baseline
(VAS fatigue with MAF-GFI, r = 0.64; VAS sleep with SPI, r = 0.50; VAS depression
with HADS-D, r = 0.43; VAS anxiety with HADS-A, r = 0.47). Content construct of
the PROMs-FM scales for functional disability and quality of life revealed correla-
tion with both SF-36 ( r = −0.86) and EuroQoL-5D ( r = 0.88) scores [ 78 ]. The correla-
tion between the FSQ and the clinical severity index was r = 0.53, moderate.
Responsiveness
Regarding responsiveness , it was only tested in the FIQ, the FFS, and the ICAF. The
approach to measure responsiveness in FIQ was rather weak in a clinical trial of
acupuncture [ 84 ]. It showed an area under the curve of 0.77 to discriminate change,
with no clear intervention or anticipated change. The FFS moved signifi cantly (by
Student’s t test) in patients who improved the Clinical Global Impressions scale in
a 24-week trial [ 85 ]. The FSS showed an area under the curve of 0.65 compared to
the Clinical Severity Index [ 84 ]. The ICAF has also proven sensitivity to change
L. Carmona et al.