Pain. 1993 (Feb); 52 (2): 157–168 ~ FULL TEXT
Gordon Waddell, Mary Newton, Iain Henderson, Douglas Somerville and Chris J. Main
Glasgow, Scotland UK.
Pilot studies and a literature review suggested that fear-avoidance beliefs about physical activity and work might form specific cognitions intervening between low back pain and disability. A Fear-Avoidance Beliefs Questionnaire (FABQ) was developed, based on theories of fear and avoidance behaviour and focussed specifically on patients' beliefs about how physical activity and work affected their low back pain. Test-retest reproducibility in 26 patients was high. Principal-components analysis of the questionnaire in 210 patients identified 2 factors: fear-avoidance beliefs about work and fear-avoidance beliefs about physical activity with internal consistency (alpha) of 0.88 and 0.77 and accounting for 43.7% and 16.5% of the total variance, respectively.
Regression analysis in 184 patients showed that fear-avoidance beliefs about work accounted for 23% of the variance of disability in activities of daily living and 26% of the variance of work loss, even after allowing for severity of pain; fear-avoidance beliefs about physical activity explained an additional 9% of the variance of disability. These results confirm the importance of fear-avoidance beliefs and demonstrate that specific fear-avoidance beliefs about work are strongly related to work loss due to low back pain. These findings are incorporated into a biopsychosocial model of the cognitive, affective and behavioural influences in low back pain and disability. It is recommended that fear-avoidance beliefs should be considered in the medical management of low back pain and disability.
Key words: Low back pain; Disability; Cognitive factors; Fear-avoidance beliefs; Psychometric properties;
Cognitive-behavioural analysis; Biopsychosocial model; Medical management; Rehabilitationn
. The 15 items
in PAIRS cover a broad range of functional limitations
which the patient attributes to pain but only I item
From the FULL TEXT Article:
Low back pain and disability are frequently treated
by patients and clinicians as if they were synonymous
but should be distinguished both conceptually and clinically.
Pain is the most common presenting symptom in
medical practice [Engel 1959; Penfield 1969]. It is now
widely recognised that pain has both sensory and affective
dimensions; the current neurophysiological view is
that pain activates affective change directly through the
reticular system and that this affective disturbance is
an integral part of pain rather than a secondary effect
[Panksepp et al. ‘1991]. This is reflected in the IASP
definition of pain which includes its emotional and
aversive qualities [Merskey 1979]. A more comprehensive
concept of pain, derived from the gate-control
theory of pain [Melzack and Wall 1965; Melzack and
Casey 1968] and tested empirically by the McGill Pain
Questionnaire [Melzack 1975; Turk et al. 1985; Lowe
et al. 1991], includes sensory, affective and cognitive
dimensions. From a different background, clinical psychology
has further developed and focussed the cognitive-
behavioural dimensions of chronic pain [Fordyce
1976; Turk et al. 1983; Turner and Roman0 1990]. In
lay usage and clinical practice, however, pain is still
regarded primarily as a sensory symptom. Patients and
physicians generally think about pain, communicate
about pain and treat pain in terms of tissue damage,
injury and nociception, although in chronic low back
pain there may no longer be any demonstrable structural
impairment [Waddell et al. 1992]. Finally, since
pain is subjective, clinical assessment of pain is dependent
on the patient’s communication of pain whether
by verbal report [Melzack 197.5, 1987] or overt pain
behaviour [Keefe and Block 1982; Waddell and
Disability is restricted function and can be assessed
reliably by clinical interview [Waddell and Main 1984],
questionnaire [Fairbank et al. 1980; Roland and Morris
1983] or work loss. Disability is always defined as
restriction resulting from an impairment [Waddell and
Main 1984]. The World Health Organisation 
definition of impairment in general is “any loss or
abnormality of psychological, physiological or anatomical
structure or function”. Physical impairment is
“pathological, anatomical or physiological abnormality
of structure or function leading to loss of normal bodily
ability” [Waddell et al. 1992]. In practice the emphasis
in clinical evaluation of physical impairment is again on
tissue damage. In chronic low back pain, however,
functional restriction due to pain, which may be regarded
as physiological impairment, may be more important
than any anatomical or structural impairment
[Waddell et al. 1992]. In chronic pain the major problem
lies in the interpretation of ‘psychological impairment’.
Psychiatric illness meeting DSM-III criteria
would constitute such an impairment [Mendelsson
1988]. More commonly, however, the affective dimension
of chronic pain presents clinically as psychological
distress in the form of anxiety, increased somatic
awareness and depressive symptoms [Waddell et al.
1984; Main et al. 1992a]. This level of distress is not
normally diagnostic of psychiatric illness. The extent to
which distress can be said to constitute a psychological
impairment, in the sense of the definition, then becomes
a matter of debate.
In the absence of proportionate physical, physiological
or psychological impairment, disability due to
chronic low back pain requires further exploration of
the patient’s cognitions and pain behaviour.
Pilot studies on the relationship between low back pain and disability
A historical review [Allan and Waddell 1989] suggested
that low back pain has affected man throughout
recorded history but that chronic disability due to
simple backache is a relatively recent and peculiarly
Western epidemic [Waddell 1987b]. The increase in
low back disability appears to depend more on society’s
and medicine’s understanding and management of low
back pain than on any change in the biological disorder.
Re-analysis of data from a previous series [Waddell
et al. 1992] showed that self-reported disability in
activities of daily living was indeed related to severity
of pain on a visual analogue scale and also to objective
clinical evaluation of current physical impairment (Fig.
11, but the relationship was weak. Severity of pain only
accounted for some 10% of the variance of physical
impairment and disability. These results have been
confirmed by Linton , Waddell [1987a], Riley et
al.  and Slater et al. . Low back disability
must therefore also depend on other factors than solely
the severity of pain or objective physical impairment.
Regression analysis of a second series [Main and
Waddell 1991] showed that a larger proportion of the
variance of disability in activities of daily living could
be explained by a combination of severity of pain,
psychological distress (particularly depressive symptoms)
and illness behaviour (Table Ia). The relative
importance of pain, depressive symptoms and illness
behaviour varied in different patient groups [Waddell
et al. 1984] but similar results have also been reported
by Reesor  and Reesor and Craig . This
confirms that low back disability also depends on the
affective component of pain and on the pattern of
illness behaviour which develops - both of which may
become just as disabling as pain itself. Work loss due
to low back pain is harder to explain and in this series
severity of pain, depressive symptoms and illness behaviour
explained a smaller proportion of the variance
of work loss (Table Ib).
In a third unpublished study loss of time from work
depended more on social and work-related factors such
as the physical demands of work. In a separate
prospective study [Waddell et al. 1986] return to work
after spinal surgery was again influenced most by occupational
factors. This is supported by other reports that
socio-economic and work-related factors may be better
determinants of low back disability than either biological
or medical factors [Bigos et al. 1991; Volinn et al.
Jensen et al.  have recently reviewed the relationships
among and distinctions between beliefs, coping
and adjustment to chronic pain. In our second
series quoted above [Main and Waddell 1991] cognitive
measures [Wallston et al. 1978; Rosensteil and Keefe
1983; Flor and Turk 1988; Main et al. 1992b] and
particularly catastrophising [Rosensteil and Keefe 1983]
explained 35% of the variance of the depressive symptoms
associated with chronic low back pain, even after
allowing for severity of pain. This is consistent with
other observations that illness behaviour may depend
just as much on cognitive factors as on severity of pain
or any physical impairment [Mechanic et al. 1982; Turk
et al. 1983; Smith et al. 1986; Philips 1987; Waddell et
al. 1989]. It also supports the hypothesis that coping
strategies may be an important mediator between pain
and depression [Rudy et al. 1988] and hence also low
back disability. In this second series [Main and Waddell
1991], however, these measures of coping strategies
only accounted for 9% of the variance of work loss
after allowing for severity of pain.
A fourth exploratory study of 120 patients found
that 3 self-reports about low back pain were significantly
related to work loss: pain aggravated by physical
activity in general, by walking and by physiotherapy.
Similar observations were made by Troup .
From these pilot studies it was postulated that current
cognitive measures are too general to explain
adequately low back disability [Main and Waddell
1991]. The present study therefore attempted to explore
further the clinical, experimental and theoretical
nature of cognitions about pain.
Pain is one of the most powerful aversive drives in
animals and humans and is closely allied to fear. The
present neurophysiological view is that the systems of
fear and pain may interact in the reticular system
[Panksepp et al. 1991]. Fear theory is now based on a
powerful body of animal experimental work [Denny
1991]. For more than 40 years this has emphasised
classical conditioning and the learned nature of fear
and avoidance behaviour, though it is important to
recognise also the innate, unconditioned nature not
only of pain but also of fear and avoidance behaviour
[Panksepp et al. 1991]. Fordyce [Fordyce et al. 1968;
Fordyce 1976] applied learning theory to behavioural
medicine, relating pain behaviour to operant conditioning.
The initial emphasis was on positive re-inforcement
of pain behaviour. Subsequently, however,
he considered ‘avoidance-learning’, i.e., reduction of
pain by avoidance behaviour resulting in negative re-inforcement
[Fordyce et al. 1982]. He noted that avoidance
behaviour was based on anticipated, consequences,
so little re-inforcement was required to maintain
the behaviour. Philips  further explored the
clinical role of avoidance behaviour. She found little
evidence that avoidance behaviour reduces chronic pain
either on a short- or long-term basis. From a limited
mechanical view of pain, avoidance behaviour may at
first appear to be adaptive but a broader cognitive-behavioural
view shows that in chronic pain it is generally
maladaptive. Philips  also emphasised the importance of beliefs and cognitions in avoidance behaviour,
unlike the predominantly behavioural emphasis of
Fordyce . Lethem et al.  and Troup et al.
 outlined the most specific “fear-avoidance model
of exaggerated pain perception” in chronic low back
pain. This emphasised the central role of fear of pain
and consequent pain-avoidance behaviour as the most
important cognitive-behavioural dimension. Troup
 later noted that fear avoidance might be of
specific relevance to incapacity for work.
It might be argued that fear-avoidance beliefs simply
reflect the mechanical characteristics of low back
pain and its relationship to physical activity, and this
may indeed be the origin of such beliefs. However,
other evidence suggests that patients’ perception of
physical activity and its relation to pain and also their
perception of their physical capabilities are often quite
erroneous. Patients with low back pain do generally
show lower physical performance levels than normal
asymptomatic subjects but their perception of theit
physical capacity is reduced more than actual performance
[Fordyce et al. 1981; Linton 1985; Schmidt
1985]. Dolce et al. [1986~] showed that pain tolerance,
exercise performance [Dolce et al. 1986b] and treatment
outcome [Dolce et al. 1986a] all correlate with
self-efficacy expectancies, i.e., with patients’ own prediction
of their ability to cope. This shows that patients
can estimate to some extent their physical performance
and it is suggested that this is based on their previous
experience [Dolce 1987]. But this does not address the
cognitive factors underlying that estimate. Council et
al.  found that chronic low back pain patients’
expectations of the pain associated with certain physical
activities correlated 0.40-0.74 with their subsequent
performance. They concluded that actual performance
was best predicted by self-efficacy ratings, which in
turn appeared to be determined by pain response
expectancies. However, Rachman and Lopatka 
showed that 25% of patients with chronic arthritic pain
tend to over-predict the amount of pain which they
subsequently report on physical exercise. Schmidt
 found that the treadmill endurance of chronic
low back pain patients was 73% of that of healthy
matched controls when they had low external information
However, with this particular form
of exercise patients did not report any increase in pain.
Although post-exercise ratings of exertion were similar
in the 2 groups, the low back pain patients actually
showed lower levels of physiological demand and
stopped because of over-estimate of exertion rather
than increased pain. Pope et al.  showed that
restricted spinal mobility, reduced flexion-extension
torque ratio and reduced straight leg raising due to low
back pain were all associated with lower pain threshold
and tolerance. Exercise to the limit of pain tolerance is
also strongly influenced by conscious feedback of how
much exercise has been done [Cairns and Pasino 1977;
Fordyce 1988]. In the absence of such feedback, chronic
pain patients increase their performance on an incremented
exercise programme at the same rate as normal
pain-free subjects [Fordyce et al. 1982]. Exercise
quotas produce systematic increases in both exercise
levels and expectancies of exercise capabilities while
reducing worry and concern about exercising [Dolce et
al. 1986b]. Finally, patients with low back pain perceive
their work as being physically more demanding than do
their healthy co-workers [Magora 1973; Dehlin and
Berg 1977; Troup et al. 1987]. From this evidence it is
clear that while fear-avoidance beliefs may originate
from the patient’s own experience of how physical
activity affects their pain, fear-avoidance beliefs can
then be modified greatly by cognitive and affective
Turk and Flor  pointed to the lack of empirical
evidence for cognitive-behavioural theories of
chronic low back pain. Caldwell and Chase 
made one of the earliest clinical observations of the
importance of “over-learning of pain-fear” in chronic
low back pain but no data was presented. Slade et al.
 related low back disability in students to previous
general coping strategies for other types of pain
but they did not look specifically at cognitions related
to current low back pain. Smith et al.  showed
that cognitive errors, particularly over-generalisation
about chronic low back pain-related activities, were
related to low back disability though this study was
limited by the nature of the Cognitive Errors Questionnaire
[Lefebvre 1981]. Sandstrom and Esbjornsson
 showed that the low back pain patient himself
or herself gave the best prediction of return to work
after a vocational rehabilitation programme. Two specific
questions were used. “I am afraid to start working
again, because I don’t think I will be able to manage.”
“My closest relatives worry that my condition will
deteriorate if I start working.” They concluded that the
patient’s attitude towards work should be carefully
evaluated before a rehabilitation programme but they
did not develop the measurement of beliefs further.
The Pain and Impairment Relationship Scale
(PAIRS) Beliefs Questionnaire developed by Riley et
al.  specifically attempted to measure patients’
beliefs about the relationship between pain and functional
impairment. They developed the instrument in
56 chronic pain patients referred to a multi-disciplinary
pain treatment programme. Slater et al.  confirmed
its psychometric properties in a small series of
31 male patients attending a general orthopaedic clinic
with chronic low back pain. Even in such small series,
both studies showed a strong relationship between
PAIRS beliefs and functional restriction, even after
allowing for severity of pain [Riley et al. 1988] or
affective disturbance Water et al. 1991]
The Survey of Pain Attitudes (SOPA) was designed
by Jensen et al.  to assess attitudes considered
important in the long term adjustment of chronic pain
patients. The original instrument was constructed with
5 a priori scales directed to beliefs about medical cure,
pain control, solicitude, disability and medication. They
subsequently added a 6th emotional scale [Jensen and
Karoiy 1987] and the factor structure of this version
was essentially rephcated by Strong et al. . A 7th
harm scale has since been added to assess the extent to
which patients believe that pain means they are damaging
themselves and that they should avoid exercise
[Jensen 1991]. Conceptually, this is very similar to the
Fear-Avoidance Beliefs Questionnaire (FABQ) but it
does not consider work. There is no pubIished data on
the psychometric properties or chnical validity of the
SOPA harm scale.
This theoretical analysis, literature review and pilot
studies all support the hypothesis that fear-avoidance
beliefs may be a specific and powerful cognitive factor
in low back pain. The aims of the present study were
a-fold. Firstly, to develop a questionnaire to measure
fear-avoidance beliefs about physical activity and work
suitable for routine clinica use in patients with Iow
back pain. Secondly, to use that questionnaire to investigate
the relationship between low back pain, fearavoidance
beliefs and chronic disability in activities of
daily living and work loss.
Material and methods
The present FABQ was developed in 2 pilot groups totalling 30
patients attending an orthopaedic out-patient clinic. The FABQ
focussed on patients’ beliefs about how physical activity and work
affected their current low back pain. It was based mainly on fear
theory and fear-avoidance cognitions but also drew on the concept of
Disease Conviction [Pilowsky and Spence 1975, 1983] which includes
beliefs about the seriousness of the illness and its effect on the
patient’s life and on the concepts of somatic focussing and increased
somatic awareness [Main 1983]. The final format was a self-report
questionnaire of 16 items presented on a single page (Appendix).
The items were original though the wording of many of the items was
derived from Fordyce’s teaching aphorisms about pain behaviour
[Fordyce 1988] and the 2 questions from Sandstrom and Esbjornsson
. Only I item was similar to 1 of the items in PAIRS [Riley et
al. 1988] and 3 items to the SOPA harm scale [Jensen 1991]. Each
item was answered on a ‘?-point Likert scale from strongly disagree
to strongly agree.
Short-term test-retest reproducibility was estimated on 26 out-patients
referred to a hospital physiotherapy department with low back
pain. The questionnaire was administered on the first attendance for
assessment and repeated on the first treatment visit 48 h later. No
active treatment was given between these 2 tests. The time interval
was chosen to minimise clinical or cognitive change, but also makes it
unlikely that patients would remember their previous answers.
The main study was of 184 patients with low back pain and/or
sciatica referred to various hospital out-patients departments: 22%
were direct GP referrals to a physiotherapy department, 43% were
primary GP referrals to 2 orthopaedic departments in a teaching
hospital and a district general hospital. 28% were secondary referrals
from other consultants to a regional problem back clinic and 6r%
were attending a rehabilitation unit. All patients were Caucasian
with English as their native language, 55.7% were male, 44.3%
female. All patients were aged 18-60 years, mean age 39.7* 11.7
years. Patients with serious spinal pathology such as tumour, infection
or inflammatory disease, spinal fractures, structural spinal deformity,
major neurology, history of primary psychiatric disease or
alcohol abuse, or inability to read and write were excluded.
The mean time from the first ever attack of low back symptoms
was 7.4 ± 8.4 years and mean duration of the present attack 13.7 ± 19.8
months. Work status was 34% still working, 24% off work because of
back pain but still had a job, 15% had lost their job because of back
pain and 27% not relevant. The mean duration of time off work due
to back pain at the time of assessment was 4.0 ± 4.3 months and total
work loss over the past year was 6.0 ± 10.8 months.
In the main study clinical assessment of all patients was carried
out by one of the authors. A standard medical history and subsequent
review of the hospital records checked for any of the exclusion
conditions listed above. Clinical assessment of pain included the
anatomical pattern, time pattern and severity. The anatomi~ai pattern
was low back pain alone in 32%, low back pain plus referred
thigh pain in 36% and nerve root pain in 32% [Waddell l982]. The
time pattern was acute in 5%, recurrent in 20% and chronic in 75’%
[Waddell 1982]. Severity was assessed by a visual analogue pain scale
[Waddell 1987a]. The self-report battery included the FABQ, disability
in activities of daily living [Roland and Morris 1983] and 2
measures of current psychological distress - the Modified Somatic
Perception Questionnaire [Main 1983] and modified Zung Depressive
Inventory [Zung 1965; Main et al. 1992a].
Reliability statistics were calculated on the 26 patients of the
test-retest study. Simple percent agreement on retest was checked by
K statistics to allow for the level of chance agreement and distribution
effects [Cohen 1960; Fleiss et al. 1969].
The scale structure of the questionnaire was analysed on the
combined total of 210 patients from the main study and the reproducibility
study and confirmed in 2 random halves. The factor
structure was explored using principal-components analysis with varimax
rotation [Gorsuch 1974]. An orthogonal rotation was used as
this was an exploratory analysis. This maximises the association of
each item with its ‘best’ factor to produce the most parsimonious and
conceptually clearest scale structure. A series of alternative factor
structures was explored and the final structure chosen on the basis of
scree, mineigen, correlation matrix and item analysis. Items were
accepted on the final factors if they had a loading of at least 0.45 on
that factor and less than 0.30 on any other factor. The possibility of
higher order factors was checked by non-orthogonal rotation.
The relationship of fear-avoidance beliefs to other clinical variables
was analysed in the main study of I84 patients. Significant
relationships were identified by correlation statistics and the relationship
between disability and fear-avoidance beliefs was then explored
further with hierarchical multiple linear regression analysis.
The fiied order of entry in the regression analysis was based on
theory. Biomedical measures of pain were entered first, accepting
that low back pain itself is the primary problem and should be
controlled for. Fear-avoidance beliefs were inserted second and
depressive symptoms third, on the basis that cognitions develop
secondarily to pain and then mediate between pain and depressive
symptoms [Rudy et al. 1988].
All 16 individual items reached acceptable levels of
test-retest reproducibility. 71% of individual answers
were identical on retest which is high for 7-point scales.
K statistics confirmed that all 16 items had high level of
reproducibility. Two items had moderate concordance
of 0.41-0.60, 8 had substantial concordance of 0.61-
0.80 and 6 had close to complete concordance of
greater than 0.80 [Landis and Koch 1977]. The average
level of K for all 16 items was 0.74 and all reached the
0.001 level of significance. Pearson product-moment
correlation coefficients for the 2 scales (see below)
were 0.95 and 0.88.
Distribution of individual items
In the complete series of 210 patients none had to
be excluded because of lack of understanding or compliance
on the FABQ itself, although 21 patients were
unable to complete items 6-16 because they were not
employed. Otherwise, there were only 1.3% missing
answers and no individual item had greater than 2.1%
missing answers. The answers to all 16 individual items
were distributed across at least 4 categories and no
item had to be excluded from further analysis because
of excessively skewed distribution [Maxwell 1971].
Principal-components analysis indicated a 2-factor
structure (Table II, Appendix). Consideration of the item content suggested that Factor 1 (FABQl) concerns
fear-avoidance beliefs about the relationship between
low back pain and work while Factor 2 (FABQ2)
concerns fear-avoidance beliefs about physical activity
in general. Item analysis showed that items 13, 14 and
16 were redundant while item 1 was characterised by
low communality and inconsistent factor loadings. Item
8 did not fit on either factor. Non-orthogonal analysis
produced an identical factor structure.
An alternative 3-factor solution was unstable with
the third factor just reaching mineigen criteria, largely
associated with item 8 (compensation) and accounting
for a small percentage of the total variance. Alternatively,
a single general factor of 11 items yielded an (Y
of 0.82 and certainly the 2 factors do inter-correlate
0.39. However, scree, mineigen and the correlation
matrix all suggested that a 2-factor structure is stable
and statistically reliable. Moreover, treating the 11
items as 1 scale would lose the conceptual differentiation
between the 2 dimensions of fear-avoidance beliefs
about work and physical activity.
Relationship of FABQ to clinical cariables
FABQl was weakly related to sex (female 19.3 ± 12.7, male 24.0 ± 12.8, t value 2.39, P = 0.02). FABQ2
was not significantly related to sex. Neither factor
correlated significantly with age.
The 2 factors were largely independent of any of the
biomedical measures of pain (Table III). Only FABQl
was related to the anatomical pattern of pain though
not in the order of pathological severity. The mean
score of FABQl in patients with low back pain alone
was 21.9 ±_ 14.3, in low back pain plus referred thigh
pain 25.2 ± 12.1 and in nerve root pain 17.8 + 12.0
(analysis of variance: F ratio 4.61, P = 0.01). Only
FABQl was significantly related to severity of pain
(Table III) but even then the correlation was low with
only 5% of variance in common. Neither of the factors
showed any significant correlation with total duration,
duration of the present episode or time pattern of
Fear-avoidance beliefs correlated strongly with both
self-reported disability in activities of daily living and
work loss (Table III>. Regression analysis confirmed
this relationship (Table IV). Both disability and work
loss were poorly explained by any of the biomedical
characteristics of pain. Fourteen per cent of the variance
of disability was explained by severity of pain but
only 5% of work loss in the past year was explained by
the time pattern of pain, which is to some extent a
statistical artefact. Fear-avoidance beliefs explained a
much higher proportion of the variance. Fear-avoidance
beliefs about work (FABQl) was consistently the
stronger of the 2 scales and explained a highly significant
portion of the variance of disability and work loss
even after allowance for pain. Fear-avoidance beliefs
about physical activity explained an additional portion
of disability but not of work loss. Analysis of present
work loss gave similar results to those presented for
total work loss in the past year, although with a generally
lower amount of variance explained. Fear-avoidance
beliefs about work were consistently more powerful
in males than in females accounting for 27% and
35% of the variance of disability and work loss in the
past year, respectively, in males but only 17% and 18%
in females. When a gender X beliefs interaction was
tested, however, it did not enter significantly into the
When the rejected items were reconsidered on an
individual basis, neither causal beliefs nor compensation added anything to the analysis of either disability
in activities of daily living or work loss, though in this
series only 18% of patients had or were considering a
claim for compensation. Finally, depressive symptoms
still explained a small but significant additional proportion
of the variance of disability in activities of daily
living in males but added little to the analysis of work
loss after allowing for fear-avoidance beliefs. This is
consistent with findings that depression associated with
pain is very largely mediated by and secondary to
cognitive factors [Rudy et al. 1988].
There are 2 main findings in this study. The first is
that there was little direct relationship between pain
and disability. This study employed a correlational
design and causation could only be confirmed by longitudinal
experiment or a true experimental design. Nevertheless,
in this analysis severity of pain only explained
14% of the variance of disability in activities of
daily living. All of the available biomedical measures
combined could only explain 5% of the variance of
work loss. Although this may at first seem surprising to
many clinicians, it is consistent with the findings of
Pilot Studies 1–4.
The second main finding is the strength of the
relationship between fear-avoidance beliefs about work
and both work loss and disability in activities of daily
living. Fear-avoidance beliefs about physical activity
were weaker but did explain an additional proportion
of the variance of disability in activities of daily living.
Compensation also appeared to be less important in
this analysis. Disability is generally more protracted in
patients with work-related low back injuries though
many studies have failed to show any difference in the
amount of psychological distress of patients with
work-related or non-work-related low back injuries
[Mendelsson 1991]. However, fear-avoidance beliefs
about work are likely to be profoundly affected by a
work-related injury and may more directly and powerfully
influence work loss than any secondary gain from
In this study fear-avoidance beliefs were able to
predict a substantial portion of the variance of disability,
both in work loss and activities of daily living, both
stronger than biomedical measures of pain and over
and above these biomedical measures. Specific fearavoidance
beliefs about work were stronger than more
general fear-avoidance beliefs about physical activity.
Fear-avoidance beliefs about work appeared to be better
able to predict work loss in this study than more
general cognitive measures such as the CSQ or PLC in
Pilot Study 5. There are certain similarities in the
theoretical bases and some of the physical activity
items of the FABQ and both PAIRS [Riley et al. 1988]
and SOPA [Jensen et al. 1987]. However, the FABQ
specifically focuses on work which the present results
suggest is most important whereas neither PAIRS nor
SOPA assess beliefs about work. To confirm that fearavoidance
beliefs about work are the most specific and
strongest cognitive factor associated with low back pain
further research should compare the FABQ directly
with PAIRS [Riley et al. 1988] and the latest version of
SOPA including the harm scale [Jensen 1991] and also
with more general cognitive measures such as the CSQ
and PLC [Main and Waddell 1991].
These findings also confirm the conclusion of the
literature review that fear-avoidance beliefs are not
simply a reflection of the mechanical characteristics of
the low back pain. There was little direct relationship
between fear-avoidance beliefs and any of the biomedical
characteristics of pain measured in this study (Table
III). They were not related to any clinically significant
extent to the severity of pain. Indeed, when considering
the different anatomical patterns of pain,
fear-avoidance beliefs did not increase with pathological
severity but rather with increasing uncertainty of
diagnosis. In the regression analysis of disability (Table
IV) fear-avoidance beliefs added a completely separate
and additional dimension to the biomedical characteristics
of pain. In their final expression it is the patient’s
beliefs rather than the underlying physical reality which
govern behaviour. Fear of pain and what we do about
pain may be more disabling than pain itself.
There are several limitations to the present study.
The analysis was based entirely on self-report measures
and should ideally, as far as possible, be checked or
compared with externally validated measures. These
findings could be tested with alternative and more
comprehensive measures of pain such as the McGill
Pain Questionnaire [Melzack 1975, 1987] and with
behavioural observations [Waddell and Richardson
1991]. It only considered chronic pain and further
study is required of acute pain. The dominance of work
beliefs means that the FABQ is at present only validated
for patients who are or have recently been
employed. It provides a more limited single scale
(FABQ2) in those who are unemployed or off work for
other health or non-health reasons in addition to low
back pain. It is unclear how long a person has to be
unemployed before they cannot use items 6-16 or
whether they tend to answer with reference to their
previous work, which is likely to be biased, or to work
which they hope to do in the future, which would bc
It is possible that the FABQ could
be used for housewives with modified instructions to
regard household duties as their work, although in this
study a number of housewives spontaneously completed
the questionnaire on this basis. At present, only
scale FABQ2 should be used in these groups. The
lower proportion of variance accounted for by fearavoidance
beliefs in females reflects the generally
weaker relationships between any biomedical measures
and work loss associated with low back pain in females.
There appear to be sex differences not only in work
circumstances and attitudes to work but also in interacting
family and social factors.
The FABQ provides strong empirical evidence for
the fear-avoidance theory. Based on these results it is
also possible to form a series of hypotheses for a
cognitive-behavioural model of low back disability (Figs.
2-4). Figure 2 is a further development of Loeser’s conceptual
model of pain [Loeser 1982] and the Glasgow Illness Model [Waddell et al. 1984]. It offers a biopsychosocial
cross-section of the clinical presentation and
assessment of chronic pain and disability at one point
in time. Fig. 3 compares the different elements of pain
and illness behaviour in acute and chronic pain. Fig. 4
considers the major causal pathways postulated between
low back pain and disability. Several points may
Throughout, the original physical basis of low back
pain is accepted, even if the anatomical site or pathological
nature of nociception cannot be identified clinically.
From the original definitions, physical impairment
may lead directly to disability. In chronic low
back pain, however, there may be no evidence of any
permanent structural impairment: instead, there is limitation
of physical function due to pain [Waddell et al.
1992]. This may be associated with muscle guarding,
deranged movement and a disuse syndrome. Chronic
pain is sometimes described as persisting beyond normal
healing time: if there is no longer any evidence of
tissue damage it is sometimes implied that there is no
remaining nociception. This would incorrectly imply
that there is no longer any sensory component to the
pain. This is neither theoretically nor clinically acceptable. Instead, it may be hypothesised that the physiological
impairment described above can give rise to
musculoskeletal nociception related to physical activity,
though this continued nociception now has a very different
pathological basis and different implications for
It should again be emphasised that this study employed
a correlational design and the effects of fearavoidance
beliefs could only be tested in a longitudinal
study or true experimental design. Nevertheless, the
model illustrates the various possible roles of cognitive,
affective and behavioural factors linking pain, illness
behaviour and disability. These may become increasingly
important in chronic pain with self-sustaining
feed-back. Figure 4 shows the main cognitive pathways
postulated between pain and disability, though obviously
other cross-links could be postulated and tested.
The alternative pathways would allow for the completely
different clinical populations seen in highly
selected Pain Clinic patients and in medico-legal practice.
It should, however, be emphasised that such a
biopsychosocial model of low back pain and disability
is best regarded as a hypothesis for further empirical
The strength of fear-avoidance beliefs and their
powerful relationship to disability has implications for
medical management. It may be postulated that current
medical advice and treatment for low back pain,
and particularly unjustified restriction of activity, the
prescription of rest and sick certification by rote [Waddell
1987b], would appear likely to cause or re-inforce
fear-avoidance beliefs and hence iatrogenic disability.
Changed medical management in this way could even
be one of the mechanisms leading to our current
epidemic of low back disability [Allan and Waddell
1989]. Physicians and therapists should be aware constantly
of the possible central role of fear-avoidance
beliefs in the development of chronic incapacity and
work loss. Although this study only concerned chronic
pain, fear-avoidance beliefs could develop much earlier.
To prevent chronicity, such inappropriate fearavoidance
beliefs would need to be recognised from
the acute stage, tackled directly and changed early
before they become fixed. Indeed, it is possible that the
first step to successful rehabilitation may be to overcome
mistaken fear-avoidance beliefs. Further consideration
should be given to the interaction between
work-related injury, fear-avoidance beliefs and compensation.
Many of these hypotheses should be
amenable to empirical testing using the FABQ.
Pain and disability, beliefs and behaviour, medical
management: the relationship between them may be
the key to understanding the present epidemic of low
back disability. Fear-avoidance beliefs merit further
study as one of the possible links in a biopsychosocial
model of low back pain and disability.
Fear-Avoidance Beliefs Questionnaire (FABQ)
Here are some of the things which other patients have told us about their pain. For each statement please circle any
number from 0 to 6 to say how much physical activities such as bending, lifting, walking or driving affect or would
affect your back pain.
We are grateful to the Chief Scientist Office of the
Scottish Home and Health Department and The Mactaggart
Trust for financial support. Duncan Troup first
stimulated our study of fear avoidance. Bill Fordyce,
John Hunter, John Loeser, Tom Rudy and Dennis
Turk provided helpful comments on an earlier draft.
Allan, D.B. and Waddell, G., An historical perspective on low back
pain and disability, Acta Orthop. Stand., 60, Suppl. 234 (1989)
Bigos, S.J., Battie, M.C., Spengler, D.M., Fisher, L.D., Fordyce,
W.E., Hansson, T.H., Nachemson, A.L. and Wortley, M.D., A
prospective study of work perceptions and psychosocial factors
affecting the report of back injury, Spine, 16 (1991) 1-6.
Cairns, D. and Pasino, J., Comparison of verbal reinforcement and
feedback in the operant treatment of disability due to chronic low
back pain, Behav. Ther., 8 (1977) 621-630.
Caldwell, A.B. and Chase, C., Diagnosis and treatment of personality
factors in chronic low back pain, Clin. Orthop., 129 (1977) 141-
Cohen, J., A coefficient of agreement for nominal scales. Educ.
Psychol. Meas., 20 (19601 37-46.
Council, J.R., Ahern, D.K., Follick, M.J. and Kline, CL., Expectancies
and functional impairment in chronic low back pain, Pain, 33
Dehlin, 0. and Berg, S., Back symptoms and psychological perception
of work: a study among nursing aides in a geriatric hospital,
Stand. J. Rehab. Med., 9 (1977) 61-65.
Denny, M.R. (Ed), Fear, avoidance and phobias: a fundamental
analysis, Lawrence Erlbaum, Hillsdale, NJ, 1991.
Dolce, J.J.. Self-efficacy and disability beliefs in behavioural treatment
of pain, Behav. Res. Ther., 25 (1987) 289-299.
Dolce, J.J., Cracker, M.F. and Doleys. D.M., Prediction of outcome
among chronic pain patients, Behav. Res. Ther., 24 (1986a)
Dolce, J.J., Cracker, M.F., Moletteire, C. and Doleys, D.M.. Exercise
quotas, anticipatory concern and self-efficacy expectancies in
chronic pain: a preliminary report, Pain, 24 (1986bl 3655372.
Dolce, J.J., Doleys, D.M., Raczynski, J.M., Lossie, J., Poole, L. and
Smith, M., The role of self-efficacy expectancies in the prediction
of pain tolerance, Pain. 27 (1986~1 261-272.
Engel, G.L., Psychogenic pain and the pain prone patient, Am. J.
Med., 26 (1959) 899-918.
Fairbank, J.C.T., Mbaot, J.C., Davies, J.B. and O’Brien, J.P., The
Oswestry Low Back Pain Disability Questionnaire, Physiotherapy,
66 (1980) 271-273.
Fleiss, J.L., Cohen, J. and Everitt, B.S., Large sample standard errors
of kappa and weighted kappa, Psychol. Bull., 72 (1969) 323-327.
Flor, H. and Turk, D.C., Chronic back pain and rheumatoid arthritis:
predicting pain and disability from cognitive variables, J. Behav.
Med., 11 (1988) 251-265.
Fordyce, W.E., Behavioural methods for chronic pain and illness,
Mosby, St Louis, MO, 1976.
Fordyce, W.E., Pain and suffering: a reappraisal, Am. J. Psycho]., 43
Fordyce, W.E., Fowler, R.S., Lehmann, J.F. and Delateur, B.J.,
Some implications of learning in problems of chronic pain, J.
Chron. Dis., 21 (1968) 179-190.
Fordyce, W.E., McMahon, R., Rainwater, G., Jackins, S., Questad.
K., Murphy, T. and DeLateur, B., Pain complaint-exercise performance
relationship in chronic pain, Pain, 10 (1981) 311-321.
Fordyce, W.E., Shelton, J.L. and Dundore, DE., The modification
of avoidance learning pain behaviours, J. Behav. Med., 5 (1982)
Gorsuch, R.L., Factor Analysis. Saunders, Philadelphia, PA, 1974.
Jensen, M.P., The Survey of Pain Attitudes, Arizona State University,
Tempe, AZ, 1991.
Jensen, M.P. and Karoly, P., Notes on the Survey of Pain Attitudes
(SOPA): original (24-item) and revised (35-item) versions, Arizona
State University, Tempe, AZ, 1987.
Jensen, M.P., Karoly, P. and Huger, R., The development and
preliminary validation of an instrument to assess patient’s attitudes
towards pain, J. Psychosom. Res., 31 (1987) 393-400.
Jensen, M.P., Turner, J.A., Romano, J.M. and Karoly, P., Coping
with chronic pain: a critical review of the literature, Pain, 47
Keefe, F.J. and Block, A.R., Development of an observation method
for assessing pain behavior in chronic low back pain patients,
Behav. Ther., 13 (1982) 363-375.
Landis, J.R. and Koch, G.G., The measurement of observer agreement
for categorical data, Biometrics, 33 (1977) 159-174.
Lefebvre, M.F., Cognitive distortion and cognitive errors in depressed
psychiatric and low back pain patients, J. Consult. Clin.
Psycho]., 49 (1981) 517-525.
Lethem, J., Slade, P.D., Troup, J.D.G. and Bentley, G., Outline of a
fear-avoidance model of exaggerated pain perception, Behav.
Res. Ther., 21 (1983) 401-408.
Linton, S.J., The relationship between activity and chronic pain,
Pain, 21 (1985) 289-294.
Loeser, J.D., Concepts of pain. In: M. Stanton-Hicks and R. Boas
(Eds.), Chronic Low Back Pain, Raven Press, New York, 1982,
Lowe, N.K., Walker, S.N. and MacCallum, R.C., Confirming the
theoretical structure of the McGill Pain Questionnaire in acute
clinical pain, Pain, 46 (1991) 53-60.
Magora, A., Investigation of the relation between low back pain and
occupation, Indust. Med., 39 (1970) 504-510.
Main, C.J., The modified somatic perception questionnaire, J. Psychosom.
Res., 27 (1983) 503-514.
Main, C.J. and Waddell, G., A comparison of cognitive measures in
low back pain: statistical structure and clinical validity at initial
assessment, Pain, 46 (1991) 287-298.
Main, C.J., Wood, P.L.R., Hollis, S., Spanswick, C.C. and Waddell,
G., The distress and risk assessment method (DRAM): a simple
patient classification to identify distress and evaluate the risk of a
poor outcome, Spine, 17 (1992a) 42-52.
Main, C.J., Wood, P.L.R., Spanswick, C.C., Roberts, A.P. and Robson,
J., The pain locus of control questionnaire, Pain, submitted
for publication, 1992b.
Maxwell, A.E., Multivariate statistical methods and classification
problems, Br. J. Psychiat., 119 (1971) 121-127.
Mechanic, D., Cleary, P.D. and Greenely, J.R., Distress syndrome,
illness behaviour, access to care and medical utilisation in a
defined population, Med. Care, 20 (1982) 361-372.
Melzack, R., The McGill pain questionnaire: major properties and
scoring methods, Pain, 1 (1975) 277-299.
Melzack, R., The short-form McGill Pain Questionnaire, Pain, 30
Melzack, R. and Casey, K.L., Sensory, motivational and central
control determinants of pain. In: D.R. Kenshalo (Ed.), The Skin
Senses, Thomas, Springfield, IL, 1968, 423-443.
Melzack, R. and Wall, P.D., Pain mechanisms: a new theory, Science,
150 (1965) 971-979.
Mendelson, G., Psychiatric aspects of personal injury claims, Thomas,
Springfield, IL, 1988.
Mendelson, G., Compensation in the maintenance of disability, Eur.
J. Pain, 12 (1991) l-11.
Merskey, H., Pain terms: a list with definitions and notes on usage.
Recommended by the IASP Subcomittee on Taxonomy, Pain, 6
Panksepp, J., Sacks, D.S., Crepeau, L.J. and Abbott, B.B., The
psycho- and neuro-biology of fear systems in the brain. In: M.R.
Denny (Ed.), Fear, Avoidance and Phobias: a Fundamental Analysis,
Lawrence Erlbaum, Hillsdale, NJ, 1991, 7-59.
Penfield, W.. Foreword. In: J. White and W.H. Sweet, (Eds.), Pain
and the Neurosurgeon, Thomas, Springfield, IL, 1969, vii-viii.
Philips, H.C., Avoidance learning and its role in sustaining chronic
pain, Behav. Res. Ther., 25 (1987) 273-279.
Pilowsky, I. and Spence, N.D., Patterns of illness behaviour in
patients with intractable pain, J. Psychosom. Res., 19 (1975)
Pilowsky, I. and Spence, N.D., Manual for the Illness Behaviour
Questionnaire (IBQ), 2nd edn., University of Adelaide, 1983.
Pope. M.H., Rosen, J.C., Wilder, D.G. and Frymoyer, J.W., The
relation between biomechanical and psychological factors in patients
with low-back pain, Spine, 5 (1980) 173-178.
Rachman, S. and Lopatka, C., Accurate and inaccurate predictions
of pain, Behav. Res. Ther., 26 (1988) 291-296.
Reesor, K.A., Medically Incongruent Back Pain Presentation: an
Indication of Physical Restriction, Suffering and Ineffective Coping
with Pain. PhD Thesis, University of British Columbia, Vancouver,
Reesor, K.A. and Craig, K.D., Medically incongruent chronic back
pain: physical limitations, suffering and ineffective coping, Pain,
32 (1988) 35-45.
Riley, J.F., Ahern, D.K. and Follick, M.J., Chronic pain and functional
impairment: assessing beliefs about their relationship, Arch.
Phys. Med. Rehab., 69 (1988) 579-582.
Roland, M. and Morris, R., A study of the natural history of back
pain. Part I. Development of a reliable and sensitive measure of
disability in low back pain, Spine, 8 (1983) 141-144.
Rosensteil, A.K. and Keefe, F.J., The use of coping strategies in
chronic low back pain: relationships to patient characteristics and
current adjustment, Pain, 17 (1983) 33-44.
Rudy, T.E., Kerns, R.D. and Turk, D.C., Chronic pain and depression:
toward a cognitive-behavioural mediation model, Pain, 35
Sandstrom, J. and Esbjornsson, E., Return to work after rehabilitation:
the significance of the patient’s own prediction, Stand. J.
Rehab. Med., 18 (1986) 29-33.
Schmidt, A.J., Cognitive factors in the performance level of chronic
low back pain patients, J. Psychosom. Res., 29 (1985) 183-189.
Slade, P.D., Troup, J.D.G., Lethem, J. and Bentley, G., The fear avoidance model of exaggerated pain perception. 11. Preliminary
studies of coping strategies for pain, Behav. Res. Ther.. 21 (1983)
Slater, M.A., Hall. H.F.. Atkinson. J.H. and Garfin, S.R.. Pain and
impairment beliefs in chronic low back pain: validation of the
Pain and Impairment Relationship Scale (PAIRS), Pain, 44 (1991)
Smith. T.W., Follick. M.J., Ahern, D.K. and Adams, A., Cognitive
distortion and disability in chronic low back pain. Cogn. Ther.
Res., 10 (1986) 201-210.
Strong, J., Ashton, R. and Chant. D., The measurement of attitudes
towards and beliefs about pain. Pain, 48 (1992) 227-236.
Troup, J.D.G., The perception of musculoskeletal pain and incapacity
for work: prevention and early treatment, Physiotherapy, 74
Troup. J.D.G.. Foreman. T.K.. Baxter. C.E. and Brown. D., The
perception of back pain and the role of psychophysical tests of
lifting capacity, Spine, I? (1987) 645-hS7.
Turk, D.C. and Flor, H., Etiological theories and treatments for
chronic back pain. II. Psychological models and interventions.
Pain, 19 (1984) 209-233.
Turk, D.C., Meichenbaum, D. and Genest, M.. Pain and behavioural
medicine, Guildford Press. New York, 1983.
Turk, D.C., Wack, J.T. and Kerns, R.D.. An empirical examination
of the pain-behaviour construct, J. Behav. Med., 8 (19%) 119-130.
Turner, J.A. and Romano, J.M.. Cognitive-behavioral therapy. In:
J.J. Bonica (Ed.), The Management of Pain, Vol. 2, 2nd edn., Lea
and Feabiger, Philadelphia. PA. 1990, 171 l- 1721.
Volinn. E., Van Koevering, D. and Loeser, J.D.. Back sprain in
industry: the role of socioeconomic factors in chronicity. Spine,
I6 (1991) 542-548.
Waddell, G.. An approach to backache, Br. J. Hosp. Med., 28 (1987)
Waddell, G., Clinical assessment of lumbar impairment, Clin. Orthop.,
221 (1987a) 110-120.
A New Clinical Model For The Treatment Of Low-back Pain
Spine (Phila Pa 1976) 1987 (Sep); 12 (7): 632-644
Waddell, G. and Main, C.J., Assessment of severity in low hack
disorders, Spine, 9 (1984) 204-208.
Waddell. G. and Richardson, J., Clinical assessment of overt pain
behaviour by physicians during routine clinical examination, J.
Psychosom. Res., 3h (1991) 77-X7.
Waddell. G.. Main, C.J.. Morris. E.W.. Di Paola, M.P. and Gray,
I.C.M.. Chronic low hack pain, psychologic distress. and illness
hehaviour. Spine, Y (1984) 209-213.
Waddell, G.. Morris, E.W.. Di Pa&. M.P., Bircher, M. and Finlayson.
D.. A concept of illness tested as an improved basis for
surgical decisions in low hack disorders. Spine. II (19X6) 712-719.
Waddell, G., Pilowsky. I. and Bond. M.. Clinical assessment and
interpretation of abnormal illness behaviour in low hack pain.
Pain. 39 (1989) 41-53.
Waddell, G.. Sommerville. D.. Henderson. I. and Newton. M.. Ohjective
clinical evaluation of physical impairment in chronic low
hack pain. Spine. I7 (1902) 617~628.
Wallston, K.. Wallston, B. and DeVellis. R.. Development of the
multidimensional Health Locus of Control (MHLC) Scales, Hlth
Educ. Monogr., 6 (lY78) IhI--170.
World Health Organisation. International classification of impairments.
di?ahilities and handicaps. WHO. Geneva, 1980. 47.
Zung, W.W.K.. A self-rated depression scale, Arch. Gen. Psychiat..
32 (196.5) 63-70.
Return to OUTCOME ASSESSMENT