ORIGINAL ARTICLE
P ea
Q of
in
Pa , Ju
Ste
Ge
Be
pat
11
com
sho
sin
sho
at
Fe
ver
Sh
a 4
or
FA
sco
ana
(IC
FA
sco
SP
bel
day
app
spe
ret
pai
dic
ma
me
me
Me
T
den
inc
rat
sho
of
cov
sym
on
Wi
and
al7
exp
Th
fre
ou
Le
F
Phy
sity
Fra
Reg
den
(Wh
Dep
Uni
Supported by the American Academy of Orthopaedic Manual Physical Therapists
and Cardon Rehabilitation (grant no. _____).
T
N
sup
zati
C
Sch
Aur
from
0
d
LBP low back pain
1128
Archis study is registered at www.clinicaltrials.gov: Identifier: NCT00835302.
o commercial party having a direct financial interest in the results of the research
porting this article has or will confer a benefit on the authors or on any organi-
on with which the authors are associated.
orrespondence to Paul E. Mintken, PT, DPT, University of Colorado Denver,
ool of Medicine, Physical Therapy Program, 13121 E 17th Ave, PO Box 6508,
ora, CO 80045, e-mail: Paul.mintken@ucdenver.edu. Reprints are not available
the author.
003-9993/10/9107-00897$36.00/0
oi:10.1016/j.apmr.2010.04.009
LR likelihood ratio
PCA principal components analysis
ROC receiver operator characteristic
SPADI Shoulder Pain and Disability Index
TSK Tampa Scale of Kinesiophobia
TSK-11 11-item version of Tampa Scale of
Kinesiophobia
TSK-AA Items 1,2,9,11 of the TSK-11
h Phys Med Rehabil Vol 91, July 2010sychometric Properties of the F
uestionnaire and Tampa Scale
Patients With Shoulder Pain
ul E. Mintken, PT, DPT, Joshua A. Cleland, PT, PhD
ven Z. George, PT, PhD
ABSTRACT. Mintken PE, Cleland JA, Whitman JM,
orge SZ. Psychometric properties of the Fear-Avoidance
liefs Questionnaire and Tampa Scale of Kinesiophobia in
ients with shoulder pain. Arch Phys Med Rehabil 2010;91:
28-36.
Objective: To investigate the reliability and validity of 2
monly used measures of pain related fear in patients with
ulder pain.
Design: A preplanned secondary analysis of a prospective
gle-arm trial involving a repeated-measures design.
Setting: Outpatient physical therapy clinics.
Participants: Patients (N
80) with a primary report of
ulder pain.
Intervention: All patients completed the outcome measures
baseline and at follow-up.
Main Outcome Measures: Patients completed a modified
ar-Avoidance Beliefs Questionnaire (FABQ), the 11-item
sion of Tampa Scale of Kinesiophobia (TSK-11), and the
oulder Pain and Disability Index (SPADI) at baseline and at
8-hour follow-up. Patients were dichotomized as improved
stable at follow-up based on the Global Rating of Change.
Results: Factor analysis indicated 3 stable factors for the
BQ and 1 stable factor for the TSK-11. Shoulder specific
ring for the FABQ and TSK-11 were used in subsequent
lyses. Test-retest reliability intraclass correlation coefficient
C) was substantial for the FABQ and the TSK-11. The
BQ correlated significantly with SPADI pain and disability
res, while the TSK-11 correlated significantly only with
ADI pain scores. The shoulder-specific FABQ–W (work
iefs subscale) was a better than chance predictor of missing
s of work during the 48-hour study period.
Conclusions: The modified FABQ and TSK-11 may be
ropriate for use in patients with shoulder pain. Shoulder-
rom the University of Colorado Denver, School of Medicine, Department of
sical Therapy, Aurora, CO (Mintken); Wardenburg Health Center at the Univer-
of Colorado Boulder, Boulder, CO (Mintken); Department of Physical Therapy,
nklin Pierce University (Cleland), and Concord Hospital (Cleland), Concord, NH;
is University Manual Therapy Fellowship Program, Denver, CO (Cleland), Evi-
ce in Motion’s Orthopedic Manual Physical Therapy Program, Louisville, KY
itman), School of Physical Therapy, Regis University, Denver, CO (Whitman),
artment of Physical Therapy, Center for Pain Research and Behavioral Health,
versity of Florida, Gainesville, FL (George).r-Avoidance Beliefs
Kinesiophobia
lie M. Whitman, PT, DSc,
cific scoring of these measures resulted in substantial test-
est reliability, and the FABQ correlated with the SPADI for
n and disability. The FABQ also showed potential for pre-
tion of short-term work loss in this sample. Pain-related fear
y be an important variable in patients with shoulder pain and
rits future consideration in longitudinal studies.
Key Words: Avoidance learning; Fear; Psychology; Move-
nt; Rehabilitation.
© 2010 by the American Congress of Rehabilitation
dicine
HE POINT PREVALENCE of shoulder symptoms has
been reported to range from 20% to 33%,1 and the inci-
ce of shoulder complaints in the general population is
reasing.2 Furthermore, several authors have reported low
es of perceived recovery for patients with a new episode of
ulder or neck pain.3-6 According to Bot et al,3 less than 25%
patients with a first episode of shoulder pain reported re-
ery after 3 months, and only 32% stated they no longer had
ptoms after 1 year. Croft et al4 reported recovery rates of
ly 21% after 6 months and 49% after 18 months. Van der
ndt et al5 and Winters et al6 reported recovery rates of 51%
59% after 12 to 18 months, respectively. Finally, Rekola et
reported that 25% of patients with shoulder or neck pain
erienced at least 1 episode of recurrence within 12 months.
ese findings suggest that shoulder pain can be recurrent and
quently progresses to the chronic stage.
Behavior that is guided by fear has the potential to impact
tcomes negatively for patients with musculoskeletal pain.8-13
them et al8 proposed the Fear-Avoidance Model of Exag-
List of Abbreviations
AUC area under the curve
CI confidence interval
EFA exploratory factor analysis
FABQ Fear-Avoidance Beliefs Questionnaire
FABQ-PA Fear-Avoidance Beliefs Questionnaire Physical
Activity Subscale
FABQ-W Fear-Avoidance Beliefs Questionnaire Work
Subscale
GRC global rating of change

gerated Pain Perception that described the influence fear had on
outcomes. They hypothesized that the sensory and emotional
component of pain may lead to a fear of pain that may con-
trib
syn
fac
rea
the
des
psy
by
chr
hav
abi
fea
in
sho
tio
fea
mu
TS
wh
(FA
me
abi
TS
and
Lik
dis
fro
int
sup
ver
TS
pro
Th
TS
LB
sho
fou
sho
Ge
wh
po
dis
and
rul
pre
ou
pat
int
int
inv
TS
rep
the
sho
sco
rel
wh
det
ute
METHODS
Su
con
der
ph
tiv
wa
on
tig
tio
bet
pai
sio
no
reg
the
bil
ger
ner
ski
ner
the
sub
the
Co
De
We
cor
Me
tig
At
ma
low
Se
the
to
of
con
sub
me
ses
siv
the
Ro
cal
val
tha
wi
im
a m
fun
20
str
ert
the
pat
4-i
ite
gre
po
1129SHOULDER PAIN AND FEAR AVOIDANCE BELIEFS, Mintkenute to the development of chronic musculoskeletal pain
dromes. The complex interaction between the physiologic
tors (sensation of pain) and psychologic factors (emotional
ction to pain) may lead to a maladaptive response in which
perception of pain becomes exaggerated.14 Patients with a
ynchronous relationship between the physiologic and the
chologic components may attempt to reduce pain symptoms
avoiding physical function, which potentially leads to
onic disability over time.14 Several cross-sectional studies
e shown an association between pain-related fear and dis-
lity,15-17 and further research has shown that pain-related
r may lead to poor clinical outcomes.9-13,18-22
Currently, the evidence supports assessing pain-related fear
patients with LBP,15,16,18,23-28 neck pain,29-35 knee pain,36,37
ulder pain,13,30,34,38-40 and other musculoskeletal condi-
ns.11,41-48 Two questionnaires commonly used to measure
r of movement and fear-avoidance beliefs in patients with
sculoskeletal conditions include the FABQ and the
K.16,29,49 Waddell et al17 originally developed the FABQ,
ich consists of a work (FABQ-W) and a physical activity
BQ-PA) scale. The FABQ has been shown to be a valid
asure of fear-avoidance behavior with high test-retest reli-
lity in patients with neck and LBP.23,29,49,50 The original
K was a 17-item measure that assessed fear of movement
injury or reinjury.49 Patients rate each item on a 4-point
ert scale with scoring alternatives ranging from “strongly
agree” to “strongly agree.” Woby et al51 shortened the TSK
m 17 items to 11 items. The TSK-11 had marginally better
ernal consistency (

.79) than the TSK, which partially
ports previous findings that showed removing the 4 in-
sely scored items increased the internal consistency of the
K. Therefore, the TSK-11 possesses similar measurement
perties to the original and offers the advantage of brevity.51
e psychometric properties of the FABQ, TSK-11, and the
K have been well reported in patients with neck pain and
P.17-19,22,29,49,51-55
Pain-related fear has been shown to have a correlation with
ulder disability in several studies.30,34,38-40,56 Lentz et al40
nd that pain-related fear contributed additional variance to
ulder disability in subjects with shoulder-related pathology.
orge et al56 investigated pain-related fear in healthy subjects
o were subjected to a shoulder fatigue protocol. They re-
rted that higher fear was predictive of higher levels of
ability, and fear of pain was associated with pain, disability,
kinesiophobia. Kuijpers et al57 derived a clinical prediction
e for shoulder pain-related sick leave and found that the
sence of psychologic variables was prognostic for poor
tcome in work-related shoulder pain.
If pain-related fear is correlated to pain and disability in
ients with shoulder pain, this would give clinicians insight
o potential modifiable factors that may help guide behavioral
ervention strategies.40 The purpose of this study was to
estigate specific psychometric properties of the FABQ and
K-11 in a group of patients with shoulder pain. First, we
orted the factor structure of the FABQ and TSK-11 because
se data have not been previously reported for patients with
ulder pain. In this process, we developed shoulder-specific
ring for these measures. Second, we investigated test-retest
iability of the shoulder-specific fear measures for patients
o remained stable after receiving intervention. Last, we
ermined which specific pain-related fear measure contrib-
d to reports of pain, disability, and work loss.bjects
Subjects included in this preplanned secondary analysis were
secutive patients with a primary report of mechanical shoul-
pain referred to physical therapy at 1 of 7 outpatient
ysical therapy clinics who agreed to participate in a prospec-
e single-group cohort trial.58 The specific aim of this trial
s to investigate the effects of cervicothoracic manipulation
patients with shoulder pain, with a secondary aim of inves-
ating the presence of fear-avoidance beliefs in this popula-
n. Inclusion criteria for the cohort trial included patient age
ween 18 and 65 years with a primary complaint of shoulder
n and a baseline SPADI59 score of 20% or greater. Exclu-
n criteria included any medical red flags suggestive of
nmusculoskeletal disorders, acute fractures in the shoulder
ion, acute severe trauma in the cervical or thoracic region in
last 6 weeks, a diagnosis of cervical spinal stenosis or
ateral upper-extremity symptoms, osteoporosis, prior sur-
y to the cervical or thoracic region, evidence of central
vous system involvement, insufficient English language
lls to complete the questionnaires, or signs consistent with
ve root compression (defined as impairment in at least 2 of
following: reflexes, myotomal strength, or sensation). All
jects reviewed and signed a consent form approved by 1 of
following institutional review boards: the University of
lorado at Boulder, Boulder, CO; the University of Colorado
nver, Aurora, CO; Regis University, Denver, CO; Newton
llesley Hospital, Newton, MA; or Concord Hospital, Con-
d, NH.
asurement and Procedures
A single-group repeated measures design was used to inves-
ate the psychometric properties of the FABQ and TSK-11.
the initial session, all patients provided demographic infor-
tion and completed a variety of self-report measures, fol-
ed by a standardized history and physical examination.
lf-report measures included the numerical pain rating scale,60
SPADI,59 the FABQ, and the TSK-11. The SPADI was used
capture the patient’s perceived level of disability as a result
shoulder pain.59 The SPADI is a 13-item questionnaire
sisting of a pain subscale with 5 questions and a disability
scale with 8. The SPADI is a well investigated outcome
asure and has demonstrated reliability and validity in as-
sing shoulder pain and disability. The validity and respon-
eness to change of SPADI have been described in physical
rapy as well as primary and secondary care settings.61
ach et al61 determined the reliability of the SPADI through
culation of internal consistency values, with Cronbach alpha
ues ranging from .86 to .95. Beaton and Richards62 reported
t the individual-level reliability of measurements obtained
th the SPADI had an ICC of .91. The minimal clinically
portant difference is 10 points.63 Williams et al63 also found
oderate correlation between the SPADI and the physical
ction and pain components of the Medical Outcomes Study
-Item Short-Form Health Survey, thus supporting the con-
uct validity of the SPADI. Based on the psychometric prop-
ies of the SPADI, Roddey et al64 have supported the use of
SPADI as a superior region-specific outcomes tool for
ient with shoulder disorders.
In patients with LBP, the FABQ consists of 2 scales: the
tem FABQ-PA and the 7-item FABQ-W subscale.17 Each
m is scored from 0 to 6, with higher scores representing
ater levels of fear-avoidance behavior. In patients with LBP,
ssible scores range from 0 to 24 and 0 to 42 for the
Arch Phys Med Rehabil Vol 91, July 2010