|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
THE RELATIONSHIP OF FEAR-AVOIDANCE TO DISABILITY:
A QUANTITATIVE REVIEW
ALISA DIANE HASSINGER
A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE
UNIVERSITY OF FLORIDA
O 2007 Alisa Diane Hassinger
I thank my supervisory committee chair, Dr. Michael Robinson, for his invaluable
guidance and mentoring. I would also like to thank the members of the Center for Pain Research
for their help and support.
TABLE OF CONTENTS
ACKNOWLEDGMENTS .............. ...............3.....
LIST OF TABLES .........__.. ..... .__. ...............5....
AB S TRAC T ......_ ................. ............_........6
1 INTRODUCTION ................. ...............7.......... ......
Chronic Low Back Pain: A Statement of the Problem ................ ............... ......... ...7
The Fear-Avoidance Model of Exaggerated Pain Perception ................. ................ ...._.8
2 M ETHODS ................. ...............12.......... .....
3 RE SULT S ................. ...............18.......... .....
4 DISCUS SION AND CONCLUSIONS .............. ...............25....
LIST OF REFERENCES ................. ...............30................
BIOGRAPHICAL SKETCH .............. ...............35....
LIST OF TABLES
2-1 Excluded Studies............... ...............16
2-2 Included Studies............... ...............17
3-1 Baseline Correlations between Fear-Avoidance and Disability .............. ...................21
3-2 Significance and Effect Sizes of Studies with Continuous Variables............... ...............22
3-3 Significance and Effect Sizes of Studies with Dichotomous Variables ............................23
3-4 Significance and Effect Size of Studies Measuring Change ................. ......................24
Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science
THE RELATIONSHIPS OF FEAR-AVOIDANCE TO DISABILITY: A QUANTITATIVE
Alisa Diane Hassinger
Chair: Michael Robinson
Pain-related fear and its attendant avoidance behaviors have been identified as key
psychosocial variables in the development and maintenance of chronic low back pain. A
growing number of prospective studies have indicated that fear-avoidance models are useful in
predicting long-term disability. However, longitudinal data have yielded varying and sometimes
conflicting results. A systematic quantitative review of prospective studies was conducted to
examine the effect sizes of fear avoidance variables in the prediction of disability in patients with
low back pain. A literature search resulted in 12 studies which met initial inclusion criteria.
Analyses revealed that while fear avoidance variables typically share moderate zero order
correlations with disability measured at baseline, effect sizes decrease to small effects over time,
particularly when controlling for relevant variables such as initial pain intensity, or psychosocial
variables such as distress. This pattern of results indicates that fear avoidance variables may
share substantial statistical and conceptual overlap with other psychosocial variables and may be
best viewed as part of a larger psychosocial model, rather than as independent predictors of
Chronic Low Back Pain: A Statement of the Problem
Pain is one of the most common reasons people seek medical attention (Schappert, 1989).
When pain becomes chronic, there are serious negative consequences for physical,
psychological, and social functioning, with deleterious effects noted in both work and family life
Low Back Pain (LBP) is one of the most common forms of pain reported (Mantyselka et
al.., 2001; Sternbach, 1986), with 53% of adults in the general population experiencing some
disability from LBP in a six month period (Walker, Muller, & Grant, 2004). LBP is a significant
source of cost to both the individual and society, associated with increased healthcare
expenditures and lost productivity (Andersson, Ejlertsson, Leden, & Rosenberg, 1993; Stang,
Von Korff, & Galer, 1998; Stewart, Ricci, Chee, Morganstein, & Lipton, 2003; Tacci, Webster,
Hashemi, & Christiani, 1998; Walker et al.., 2004). It has long been asserted that only a small
percentage of patients who experience an acute episode will develop chronic LBP (Watson,
Main, Waddell, Gales, & Purcell-Jones, 1998). However, recent data suggests that acute LBP
evolves into a chronic or recurrent condition more frequently than previously suspected
(Hestback, Leboeuf-Yde, & Manniche, 2003).
Chronic LBP, with an estimated prevalence of 15 tol8% in the adult population, is of
special concern (Frank et al.., 1996). The cost of chronic LBP to society has been estimated at
$18 billion annually (Linton & van Tulder, 2001), a figure which distinguishes chronic LBP as
the most expensive benign medical condition in industrialized countries (Mayer, 1991). The
associated prevalence and costs highlight chronic LBP as an important public health concern,
and underscore the need for research focused on the identification and treatment of those at risk
for developing chronic LBP.
Because an underlying physical pathology responsible for the development and
maintenance of chronic LBP has not been discovered in the majority of patients who suffer with
this condition, the development of biomedically oriented prevention and treatment strategies has
not been successful (Hart, Deyo, & Cherkin, 1995; Linton & van Tulder, 2001; Loeser, 1994).
Randomized, clinical trials have repeatedly shown that commonly used strategies intended to
prevent the development of LBP, such as back schools, lumbar supports, and ergonomic
interventions, are ineffective (Linton & van Tulder, 2001; Scheer, Radack, & O'Brien, 1995).
However, studies have consistently identified psychosocial factors rather than physical
impairments as better predictors of which patients will develop chronic disability from an acute
episode of LBP (Burton, Tillotson, Main, & Hollis, 1995; Frank et al.., 1996; Gatchel, Polatin, &
Mayer, 1995; Pincus, Burton, Vogel, & Field, 2002)
The Fear-Avoidance Model of Exaggerated Pain Perception
The Fear-avoidance Model of Exaggerated Pain Perception (FAMEPP), first proposed by
Lethem and colleagues in 1983, has been posited as one explanation for why a subgroup of
patients develop chronic LBP and its attendant disability (Lethem, Slade, Troup, & Bentley,
1983; Slade, Troup, Lethem, & Bentley, 1983). This model proposes that fear-avoidance beliefs
and pain catastrophizing are the primary psychosocial factors responsible for the development of
chronic LBP symptoms. More specifically, a patient's response to LBP can fall along a
continuum of confrontation to avoidance. A confrontation response is viewed as adaptive as it
supports the patient's participation in daily social and vocational activities, encouraging a
recovery of function. An avoidance response to pain results from higher levels of pain related
fear and avoidance, and has negative psychological and physical consequences for the individual.
Psychologically, exaggerated pain perception results from elevated levels of fear (George,
Dannecker, & Robinson, 2006; George, Wittmer, Fillingim, & Robinson, 2007). Patients also
may develop depressive and anxious symptoms as a result of their perceived loss of function,
factors which are known to be associated with decreased pain tolerance (Vlaeyen & Linton,
2000). Physically, patients with LBP experience "disuse syndrome" (decreased spine range of
motion, loss of muscle force, and weight gain) as a result of their avoidant response to the fear of
pain (Bortz, 1984; Kottke, 1966; Lethem et al.., 1983; Slade et al.., 1983). A continuous
feedback loop is therefore created in which the patient's elevated fear results in avoidant
behavior, which consequently leads to increasing disability, emotional distress, and the
maintenance of the fear and avoidance.
Vlaeyen and colleagues have further elaborated on this model by proposing a Cognitive
Model of Fear of Movement/(Re)injury (Vlaeyen, Kole-Snijders, Boeren, & van Eek, 1995).
This model conceptualized the pain related fear as a specific fear of movement, particularly that
which the patient believes may result in further injury. This model also highlights the role of
pain catastrophizing, defined as the belief that the pain experience will lead to the worst possible
outcome (Sullivan et al.., 2001), as a psychosocial variable which increases the patient' s fear of
movement and subsequent avoidance behaviors (Vlaeyen & Linton, 2000).
Initial investigations have consistently demonstrated that elevated levels of pain related
fear and avoidance are associated with the development and maintenance of chronic low back
pain and its attendant disability. A study by Crombez and colleagues in 1998 categorized
subj ects as having either a confrontational or avoidance coping style, and found that those
classified as Avoiders had greater pain related fear and levels of disability, and more difficulty
with physical activities (Crombez, Vervaet, Lysens, Baeyens, & Eelen, 1998). Further evidence
of the relationship of pain related fear and functional limitations has been found in studies
investigating the performances of patients with CLBP during a task which measured walking
speed (Al-Obaidi, Al-Zoabi, Al-Shuwaie, Al-Zaabie, & Nelson, 2003) and a lifting task (Vlaeyen
et al.., 1995). A growing body of literature exists which has demonstrated a relationship
between higher levels of fear-avoidance and greater levels of disability (Asmundson, Norton, &
Allerdings, 1997; Crombez, Vlaeyen, Heuts, & Lysens, 1999; Severeijns, Vlaeyen, van den
Hout, & Weber, 2001; Waddell, Newton, Henderson, Somerville, & Main, 1993).
However, much of the research performed has been, until recently, cross sectional in
nature, and thus unable to establish a causal relationship between elevated pain related fear and
the development of chronic LBP disability. In 1995, Klenerman and colleagues published the
first prospective study examining the relationship between these variables, and found higher
levels of pain related fear to be a significant contributor to chronic LBP disability (Klenerman et
al.., 1995). Many of the prospective studies which followed also found this relationship
(Boersma & Linton, 2006; Fritz, George, & Delitto, 2001; George, Fritz, Bialosky, & Donald,
2003; Picavet, Vlaeyen, & Schouten, 2002; Woby, Watson, Roach, & Urmston, 2004).
However; the results of these studies were more mixed than those of the cross sectional
examinations. Several studies were unable to find significant prospective relationships between
pain related fear, and disability (Grotle, Vollestad, & Brox, 2006; Grotle, Vollestad, Veierod, &
Brox, 2004; Sieben et al.., 2005; Staerkle et al.., 2004).
Aims and Hypotheses
Although it appears that both cross sectional and prospective studies have indicated that
fear-avoidance beliefs may play a role in the development of chronic LBP disability, the
magnitude of these effects has not been systematically examined in the literature. With a
growing body of prospective literature, it is an opportune time to investigate the strength of the
relationship between fear-avoidance beliefs and disability in order to guide future research
endeavors. Our study proposes a quantitative review of the prospective literature to examine the
relationship between fear avoidance variables measured at baseline and a follow-up time point,
with the intention of assessing the magnitude of the relationship between pain related fear and
The first aim of our study was to examine the relationship of fear-avoidance variables to
initial disability ratings in patients with LBP. We hypothesize that fear-avoidance variables will
be highly correlated with measures of disability.
The second aim is too assess the effectiveness of the fear-avoidance model in predicting
later disability in patients with LBP. We hypothesize that fear-avoidance variables will
demonstrate large effect sizes in models predicting later LBP related disability.
A review of the literature was undertaken to identify prospective studies using Fear-
avoidance Beliefs as an independent variable predicting a measure of disability. A Medline
database search was conducted using the following search terms: Pain, Fear-avoidance; Pain,
Kinesiophobia; Pain, Fear of Movement/(Re)Injury. A publication cutoff date of September,
2006, was used; however, no backward cutoff date was specified. Additional articles were
identified using the bibliographies of the studies located through Medline database.
Twenty-Hyve studies were initially identified for inclusion in the review. Because this is a
small but increasing body of research available with a relatively new construct, liberal exclusion
criteria were developed. The first exclusion criterion precluded the inclusion of multiple studies
which used the same data set. When these studies were identified, only the first study published
using the data was included. This resulted in the exclusion of two studies. An exception to this
criterion was made for the two studies conducted by Grotle et al.., as the second study added an
additional population and used a different method of analysis.
The second exclusion criterion required that studies use standardized instruments to
measure Fear-avoidance Beliefs and Disability in order to be included in the review. Studies
with non-standardized instruments were permitted if they provided pilot data for the measure
they selected. Two studies were excluded for the use of non-standard measures. Because the
great maj ority of the literature has addressed only back pain, the use of a mixed pain population
was the third basis for exclusion, disqualifying three studies. The fourth criterion required that
studies use a sample size of 10 or more subj ects. This resulted in the exclusion of four studies
with sample sizes of four, six, six, and eight. After exclusion criteria were applied, 14 studies
were eligible for inclusion. A summary of the excluded studies can be found in Table 2-1.
Analyses were planned to examine the correlations between fear-avoidance variables at
baseline, and to calculate effect size estimates in order to examine the strength of the relationship
between the fear-avoidance variables and measures of disability after a period of time. Of the 14
eligible studies, two were deemed to contain sufficient data at publication to allow for the
strength of the relationship between fear-avoidance and disability to be examined at both
baseline and follow-up. Therefore, twelve e-mail requests were made to corresponding authors
to obtain the additional data required for the planned analyses. Of the twelve requests, one
author could not be reached, and two authors did not respond. However, one of these articles
contained enough information for one of the two planned analyses, and therefore was included.
In total, twelve articles were included in the final review, with ten studies examining fear-
avoidance variables as predictors of disability and two studies using an observed change in fear-
avoidance variables as predictors of change in disability.
The studies reviewed included both acute and chronic pain populations recruited from a
variety of settings, including primary care clinics, pain clinics, and general population surveys.
The number of subj ects included in the studies ranged from 54 to 1,571. Subj ects were followed
for time periods ranging from one month to two years. A summary of included studies can be
found in Table 2-2.
All studies measured the independent variable of pain related fear by the administration
of one of two standard self-report questionnaires: the Fear-avoidance Beliefs Questionnaire
(FABQ) or the Tampa Scale ofKinesiophobia (TSK). The Fear-avoidance Beliefs Questionnaire
(FABQ) is a measure described by Waddell and colleagues (Waddell et al.., 1993). It contains
16 items, each scored on a 6-point Likert scale with higher numbers indicating increased levels
of fear-avoidance beliefs. Two subscales have been identified within the FABQ: a seven-item
work subscale and a four-item physical activity subscale. The TSK (Miller RP, 1991) is a 17
item self report questionnaire with items scored on a 4-point Likert scale. Recent investigations
have indicated that the TSK can be divided into two subscale which describe the contracts of
somatic focus (five items) and activity avoidance (eight items)(Goubert et al.., 2004; Roelofs,
Goubert, Peters, Vlaeyen, & Crombez, 2004).
The dependent variable of disability was measured using a variety of methods, both
dichotomous and continuous. Dichotomous measures included work status, and self report
measures such as the TSK from which arbitrary categories were created. Continuous measures
included the Roland Morris Disability Questionnaire (RMDQ), Oswestry Disability Index (ODI),
or the Quebec Back Pain Disability Scale (QBPDS). The RMDQ (M. Roland, Morris R., 1983)
is a 24-item self-report questionnaire known to have excellent reliability and validity (M. Roland
& Fairbank, 2000; M. Roland, Morris R., 1983). Scores on the RMDQ range from 0 to 18, with
higher scores reflecting greater levels of disability. The ODI is a 10-item self-report scale
originally described by Fairbank and colleagues (Fairbank, 1980). Each item is scored on a
five-point Likert scale with higher numbers indicated greater disability. The QBPDS (Kopec et
al.., 1996) contains 20 items that describe different activities, such as getting dressed, climbing
the stairs, etc. The level of difficulty for each activity is expressed on a 6-point Likert scale, with
higher scores indicating greater difficulty.
In each study, the relationship of fear avoidance to disability was assessed at baseline by
examining the Pearson correlation coefficients and to obtain an estimate of the magnitude of the
relationship between the variables. Cohen's conventions were then used to designate the
relationship as small (. 1), medium (.3) or large (.5 or greater). (Cohen, 1988).
To examine the prospective relationship between fear-avoidance and disability, effect sizes
were calculated or estimated based on the type of analysis conducted in each study. All studies
using dichotomous outcome measures of disability conducted logistic regressions, which gave
rise to odds ratios as measures of effect size. Although no standard convention exists, it is
generally accepted that an odds ratio of 1.0 to 3.0 represents a weak relationship, whereas an
odds ratio over 3.0 for positive associations (or less than one-third for negative associations)
indicates a moderate to strong relationship (Haddock CK, 1998).
Studies in which continuous outcome measures of disability were assessed employed a
variety of regression analyses. When hierarchical linear regressions were used and fear
avoidance variables were entered as a separate block, Cohen's ? was calculated as a measure of
effect size. By convention, P effect sizes of 0.02, 0. 15, and 0.35 are considered small, medium,
and large, respectively (Cohen, 1988). One study made use of backwards stepwise regression
which produced a final model that did not include fear-avoidance variables, and a value of zero
assigned to the estimation of the effect size of fear-avoidance. In hierarchical linear regressions
where fear avoidance variables were not entered as a separate block but included with other
variables, or when simultaneous regression was used, an estimate of the unique proportion of the
variance explained by fear avoidance variables was obtained by squaring the standardized beta
weights and expressed in terms of a percentage of variance explained. Cohen's conventions of
.01 for a small effect, .06 for a medium effect, and .14 for a large effect were then used to assess
the magnitude of the relationship. (Cohen, 1988). This convention was also applied to a third
category of studies which used change in fear avoidance measures to predict change in disability
Table 2-1. Excluded Studies
Fritz & George
George, Fritz & McNeil
Klenerman et al.
Linton & Hallden
Letters et al.
Nederhand et al.
Turner et al.
N Reason for exclusion
78 2nd publication from data set
63 2nd publication from data set
300 Use of non-standard measure for
137 Mixed pain population
Mixed pain population
Head/neck pain population
Use of non-standard measure for
Fewer than ten participants
Fewer than ten participants
Fewer than ten participants
Fewer than ten participants
Vlaeyen et al.
Vlaeyen et al.
Jong et al.
Jong et al.
Table 2-2. Included Studies
Length of study
Boersma & Linton
Dionne, et al.
Fritz, George & Delito
George, et al.
Grotle, Vollestad & Brox
Grotle, et al.
Mannion, et al.
Picavet, Vlaeyen &
Sieben, et al.
Staerkle, et al.
Swinkels-Meewisse, et al.
Woby et al.
Table 3-1 summarizes the correlations between fear avoidance variables and disability at
baseline. Examination of the correlations of these studies indicated that, with the exception of
the fear-avoidance for work subscale of the FABQ in the acute population studied by Grotle et
al. (2005), fear avoidance variables were significantly correlated with disability. Applying
Cohen' s criteria, the subscale of FABQ-W in Grotle' s acute population, as well as both subscales
of the TSK examined in the acute population studied by Sieben et al. (2005), yielded small
correlations. Large correlations were noted by Staerkle et al. (2004) between the FABQ
sub scales and disability. Woby and colleagues (2004) also noted a large correlation between the
FABQ physical activity subscale (FA-p) and disability. The remaining studies all demonstrated
moderate correlations between the fear-avoidance variables and disability measures at baseline.
The studies were next examined to determine the unique effects of fear-avoidance in the
prediction of disability, after controlling for other relevant clinical variables. Common
covariates included variable such as initial disability rating, initial pain intensity, and
demographic variables such as age and gender. A review of the statistical significance of fear
avoidance variables in the various models predicting later disability yielded mixed results for
studies which used fear avoidance variables to predict disability. Two studies which used the
total FABQ or total TSK score rather than their subscales as the independent variable found the
fear avoidance variables to be significant predictors of disability (George et al.., 2003; Swinkels-
Meewisse et al.., 2006). Two studies which used the subscales of the FABQ or TSK were unable
to obtain significant results for any subscale (Grotle et al.., 2006; Sieben et al.., 2005). The
remaining studies, all of which used the FABQ and TSK subscales as independent predictors of
disability, found mixed significance. Of note is the study by Staerkle and colleagues (2004),
which examined the FABQ subscales as predictors of two dependent variables: perceived
disability and days of work lost. The FABQ-W was found to be a significant predictor of
reported days of work lost, although the FABQ-P was not. Neither subscale was significant in
predicting disability as measured by the RMDQ. Overall, three of the four analyses yielded non-
significant associations. Conversely, in the study by Dionne et al. (2006) which examined the
role of fear avoidance in predicting disability separately for each sex, the subscales of the FABQ
were independent predictors of disability, with the only exception being the FABQ-P in men.
The two studies which used changes in fear avoidance variables to predict changes in disability
both yielded significant results.
Examination of the effect sizes for the studies which used a continuous outcome measure
(including those analyzing change in disability resulting from change in fear avoidance) revealed
uniformly small effects of fear avoidance variables in the models. Cohen's ? effect sizes ranged
from .03 to .13, and are all considered small by Cohen's conventions. If only significant results
are considered, the effect sizes range from .03 to .10. In the five studies for which the proportion
of variance explained was used as an effect size estimate, results range from 0% (when the
variables were not included in the final model) to 6.6%. If only significant results are
considered, the estimates range from 3.7% to 6.6%.
Odds Ratios were used as measures of effect size in the studies which used dichotomous
outcome measures. Small effect sizes were noted in the study be Fritz et al. (Fritz et al.., 2001).
Picavet and colleagues divided the TSK into tertiles, and used the middle and highest tertile as
predictors of classification by the Quebec Back Pain Disability Questionnaire as disabled
(Picavet et al.., 2002). Although the odds ratios are considered small, the odds ratio of the TSK,
highest tertile [2.6(1.4-4)] approaches a moderate effect size. Dionne and colleagues examined
the subscales of the FABQ as predictors of a "return to work in good health," and examined the
results by gender (Dionne et al.., 2006). The subscale FA-P was not included in the final model
for men, and resulted in a small effect size estimate for women [1.98(1.01-3.89)]. The FA-W
subscale demonstrated moderate to large effects for both men [4.08(1.76-9.44)] and women
[3.01(1.14-7.9 1)]; however, examination of the confidence intervals indicate that the odds ratios
are unstable and must be interpreted with caution.
The data on statistical significance and effect size estimates for studies using continuous
variables, dichotomous variables, and studies measuring change are summarized in Tables 3-2,
3-3, and 3-4, respectively.
Table 3-1. Baseline Correlations between Fear-Avoidance and Disability
Disability: .56(FA-p)*, .57(FA-w)*
Work Loss: .42(FA-p)*, .47(FA-w)*
Boersma & Linton
Dionne, et al.
Fritz, George & Delito
41 TSK Subscales
48 FABQ Total
;71 TSK Total
George, et al.
Grotle, Vollestad &
Grotle, et al.
Mannion, et al.
Picavet, Vlaeyen &
Sieben, et al.
Staerkle, et al.
Woby et al.
158 TSK Subscales
431 TSK Total
* Indicates significant correlations
Table 3-2. Significance and Effect Sizes of Studies with Continuous Variables
Author N Significance estimate
Boersma & Linton 141 Not retained in model (TSK-sf) 0% (TSK-sf)
p <.01 (TSK-aa) 5.1% (TSK-aa)
Fritz, George & 69 p = .009 (FA-w) .10 (FA-w)
Delito p =.083 (FA-p) .04 (FA-p)
George, et al. 66 p = .049 (FABQ total at 1 month) .05 (FABQ)
p = .034 (FABQ total at 6 months) .08 (FABQ)
Grotle, Vollestad 123 Acute p = .23 (FA-w), p = .71 (FA-p) .07 (FABQ)
& Brox 47 Chronic p = .13 (FA-w), p = .07 (FA-p) .13 (FABQ)
Sieben, et al. 158 TSK not retained in model 0%
Staerkle, et al. 255 Disability: p > .05 for FABQ subscales .02(FABQ)
Work Loss: p < .05(FA-w), p > .05(FA-p) .03(FABQ)
Swinkels- 431 P < .01 (TSK total) 5.2%
Meewisse, et al.
Table 3-3. Significance and Effect Sizes of Studies with Dichotomous Variables
Effect size estimate
Author N Significance (odds ratio)
Dionne, et al. 269 female p <.05 for FABQ subscales FA-w 3.01(1.14 7.91)
FA-p 1.98(1.01 -3.89)
491 male p < .05 for FA-w, FA-p not retained FA-w 4.08(1.76 9.44)
Fritz, George & 78 p = .003 (FA-w) 1.17(1.04 1.31)
Delito p = .23 (FA-p) 1.14(.96 1.34)
Grotle, et al. 123 p > .05 (FA-w, high FA) 1.31(.44 3.85)
p > .05 (FA-p, high FA) 1.58(.57 4.4)
Picavet, Vlaeyen 1571 p < .05 (TSK, middle tertile) 1.3(.6 2.7)
& Schouten p < .05 (TSK, highest tertile) 2.6(1.4 4)
Table 3-4. Significance and Effect Size of Studies Measuring Change
Author N Significance Effect size estimate
(percent of variance)
Mannion, et al. 143 p <.05 3.7%
Woby, et al. 54 p <.05 (FA-w) 4.4%
p <.01 (FA-p) 6.6%
DISCUSSION AND CONCLUSIONS
The results of this review did not support the hypothesis that fear-avoidance variables
would be highly correlated with measures of disability at baseline. Of the 26 correlations
available for review, 20 fell within the moderately correlated range. Three high correlations and
three low correlations were noted.
The data provided by the studies also did not support the hypothesis that fear-avoidance
variables would demonstrate large effects in models predicting later disability. All eight Cohen's
f 2 effect sizes calculated were less than .13, and therefore are considered small by convention
(Cohen, 1988). The proportion of variance accounted for was used as an effect size estimate for
five studies. The estimates ranged from 0% variance to 6.6%, with an average of 3.6% of the
variance accounted for in the prediction of disability. If only significant results are considered,
the average increases to 5%. When Cohen's conventions are applied to these results, all studies
demonstrate small effects, with the exception of the moderate effect size of 6.6% reported by
Woby and colleagues (Woby et al.., 2004). The review of the dichotomous category of studies,
in which ORs were used as measures of effect size, continued to provide evidence that fear
avoidance variables demonstrate small effects in the estimation of disability. The ORs ranged
from 1.14 to 4.08, with three of the four studies reviewed demonstrating consistently small
effects (ORs less than three). The fourth study, by Dionne and colleagues (2006), contains
mixed results. The effect size of the FA-P subscale was not provided, as it was not significant.
The FA-P subscale for women demonstrated an OR of 1.98, which is generally considered small.
The ORs of FA-W for men and women, 4.08 and 3.01 respectively, could be considered
moderate to large. However, the confidence intervals for these two ORs are quite large, a factor
which indicates an unstable odds ratio, and may make these results difficult to replicate. Overall,
although this category of study demonstrated the most variation, the results of the studies using
dichotomous outcome measures are generally consistent with those of the studies using
continuous outcome measures in demonstrating small effect sizes for fear avoidance variables.
When the results are considered together, it appears that fear-avoidance initially correlates
moderately with disability, but when relevant variables (i.e. initial pain intensity or disability,
etc.) are controlled for, fear avoidance variables cannot account for significant variance in the
development and maintenance of disability from LBP. When considered within the context of
the Fear Avoidance Model of Exaggerated Pain Perception and the initial body of literature on
fear avoidance and LBP, these findings are unexpected. It is therefore important to consider
common factors which may have influenced the results of the studies.
There are some indications that measurement error may have been a significant factor
affecting the results in this review. Most of the studies used the subscales of the FABQ or TSK
as independent variables. However, three of the four studies which obtained consistently
significant results did so using the FABQ or TSK as a whole, rather than using the measure
subscales. This provides preliminary indication that the subscales of these measures may lack
the necessary stability to be reliably used in models predicting disability. Additionally, one third
of the studies used the TSK to assess the level of fear avoidance in their samples. Investigations
into the TSK have suggested several different factor structures, and indicated that the
measurement properties of the TSK may be problematic (Burwinkle, Robinson, & Turk, 2005;
French, France, Vigneau, French, & Evans, 2007; Roelofs et al.., 2004). However, this potential
problem in measurement is unlikely to provide a sufficient explanation for the small effect sizes
across studies, as eight of the twelve studies used the FABQ, but also demonstrated small effect
sizes for fear-avoidance.
A more compelling explanation of the Eindings can be found in examining the conceptual
and statistical overlap that may occur between the construct of fear-avoidance and other
psychosocial constructs, such as distress or depression. Systematic reviews of the literature
have consistently identified a variety of psychosocial variables as potentially important to the
transition from acute to chronic pain, to include stress, mood and depression, and cognitive
variables (i.e. coping style, self-efficacy) (Linton, 2000; Pincus et al.., 2002). Additionally,
Pincus and colleagues found moderate effect sizes demonstrated by depression and distress in
relation to a variety of outcomes, including disability (Pincus et al.., 2002). Four of the studies in
the present review examined the influence of other psychosocial beliefs on the outcome of
disability, either as individual predictors or in conjunction with fear-avoidance beliefs. Although
no baseline correlations between fear avoidance and distress were provided, Grotle and
colleagues, using a logistic regression analysis, found that, rather than fear-avoidance, distress
was the best psychosocial predictor of disability at three months, and therefore did not include
fear-avoidance in their Einal model (Grotle et al.., 2006). Staerkle et al. found that the FABQ
subscales shared moderate baseline correlations with the Zung Depression Scale, and that
depressive symptoms accounted for slightly more variance in the model predicting scores on the
RMDQ than did the FABQ subscales (R. sq. changes of .02 vs .01), although both variables
demonstrated small effects (Staerkle et al.., 2004). Mannion and colleagues conducted a
stepwise regression and found that depression and fear-avoidance beliefs accounted for 4. 1% and
3.7% of the variance, respectively (Mannion et al.., 2001). Finally, Boersma and Linton (2006)
found that, with the exception of the TSK-A subscale and pain expectancy, the TSK subscales
shared moderate correlations with pain and expectancy and negative affect, ranging from .384 -
.482. These variables were then entered together in a hierarchical regression predicting
disability, and resulted in moderate effect size (f 2 = .17). This led them to conclude, ... the
strong interrelationships between these variables should caution for treating them as separate
entities. ." (Boersma & Linton, 2006) This statement highlights the issue of the potential for
shared variance between fear-avoidance and other psychosocial variables.
Although this review followed standard recommendations for the interpretation of effect
sizes, there is a great deal of debate about the interpretation of effect sizes in the psychological
sciences. There are no universally accepted conventions by which these numbers can be
assigned a categorical ranking of small, medium, or large; rather, the judgment of effect size
must be considered within the context of the dependent variable. It can be argued that
questionnaire scores as a proxy for disability do not represent a clear and robust outcome
measure; therefore, greater proportions of the variances must be accounted for to judge the effect
size estimates as large. When considered within this context, it is reasonable to conclude that the
effect size estimates yielded by the proportion of the variance accounted for are small.
Examination of the aggregate results also supports the conclusion that the effect size for fear
avoidance variables in later disability is small. Although the reliability of the measures, the
heterogeneity of the populations studies, and the specific levels of the variable being examined
varied across studies, the effect sizes demonstrated remained consistently small.
Overall, the literature reviewed does not support the contention that fear- avoidance
variables play a central role in the development and maintenance of chronic LBP. However, it
would be premature to conclude that they are not related to LBP, and should not be a subj ect of
continued investigation. Relative to the population suffering with LBP, the sample sizes
examined in this review were small, as is the body of literature addressing this issue. Further
investigation examining different populations of sufferers may identify important interactions
between fear avoidance variables and population-specific factors. Additionally, this review
provides preliminary support to the idea that fear-avoidance variables may be better
conceptualized and studied as part of a multidimensional psychosocial model with interrelated
variables such as distress, pain catastrophizing, and negative expectancy. Because of variance
shared, it is likely that considering these variables together, as part of a larger "negative affect"
construct, will be better able to predict outcomes for patients with LBP. Finally, it is important
to continue to investigate fear-avoidance variables, as well as other psychosocial predictors,
because--due to the lack of biomedical explanations-psychosocial variables represent some of
the only intervention points currently available to patients with LBP. When compared to the
options of making no intervention or to expensive medical procedures, interventions based on
fear avoidance variables may prove to be cost effective and relatively useful, and should
therefore not be ignored as a possibility which warrants continued investigation.
LIST OF REFERENCES
Al-Obaidi, S. M., Al-Zoabi, B., Al-Shuwaie, N., Al-Zaabie, N., & Nelson, R. M. (2003). The
influence of pain and pain-related fear and disability beliefs on walking velocity in chronic
low back pain. hIt JRehabil Res, 26(2), 101-108.
Andersson, H. I., Ejlertsson, G., Leden, I., & Rosenberg, C. (1993). Chronic pain in a
geographically defined general population: studies of differences in age, gender, social
class, and pain localization. Clin JPain, 9(3), 174-182.
Asmundson, G. J., Norton, G. R., & Allerdings, M. D. (1997). Fear and avoidance in
dysfunctional chronic back pain patients. Pain, 69(3), 23 1-23 6.
Boersma, K., & Linton, S. J. (2006). Expectancy, fear and pain in the prediction of chronic pain
and disability: A prospective analysis. Eur JPain, 10(6), 551-557.
Bortz, W. M., 2nd. (1984). The disuse syndrome. West J2ed, 141(5), 691-694.
Burton, A. K., Tillotson, K. M., Main, C. J., & Hollis, S. (1995). Psychosocial predictors of
outcome in acute and subchronic low back trouble. Spine, 20(6), 722-728.
Burwinkle, T., Robinson, J. P., & Turk, D. C. (2005). Fear of movement: factor structure of the
tampa scale of kinesiophobia in patients with fibromyalgia syndrome. JPain, 6(6), 3 84-
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed. ed.). Hillsdale,
Crombez, G., Vervaet, L., Lysens, R., Baeyens, F., & Eelen, P. (1998). Avoidance and
confrontation of painful, back-straining movements in chronic back pain patients. Behav
M~odiJJ 22(1), 62-77.
Crombez, G., Vlaeyen, J. W., Heuts, P. H., & Lysens, R. (1999). Pain-related fear is more
disabling than pain itself: evidence on the role of pain-related fear in chronic back pain
disability. Pain, 80(1-2), 329-339.
Dionne, C. E., Bourbonnais, R., Fremont, P., Rossignol, M., Stock, S. R., Nouwen, A., et al..
(2006). Determinants of "return to work in good health" among workers with back pain
who consult in primary care settings: a 2-year prospective study. Eur Spine J.
Fairbank, J., Couper J, Davies JB, O'Brien JP. (1980). The Oswestry low back pain disability
questionnaire. Phys~ilotheiopy,, 66, 271-273.
Frank, J. W., Brooker, A. S., DeMaio, S. E., Kerr, M. S., Maetzel, A., Shannon, H. S., et al..
(1996). Disability resulting from occupational low back pain. Part II: What do we know
about secondary prevention? A review of the scientific evidence on prevention after
disability begins. Spine, 21(24), 2918-2929.
French, D. J., France, C. R., Vigneau, F., French, J. A., & Evans, R. T. (2007). Fear of
movement/(re)injury in chronic pain: a psychometric assessment of the original English
version of the Tampa scale for kinesiophobia (TSK). Pain, 12 7(1-2), 42-51.
Fritz, J. M., George, S. Z., & Delitto, A. (2001). The role of fear-avoidance beliefs in acute low
back pain: relationships with current and future disability and work status. Pain, 94(1), 7-
Gatchel, R. J., Polatin, P. B., & Mayer, T. G. (1995). The dominant role of psychosocial risk
factors in the development of chronic low back pain disability. Spine, 20(24), 2702-2709.
George, S. Z., Dannecker, E. A., & Robinson, M. E. (2006). Fear of pain, not pain
catastrophizing, predicts acute pain intensity, but neither factor predicts tolerance or blood
pressure reactivity: an experimental investigation in pain-free individuals. Eur JPain,
George, S. Z., Fritz, J. M., Bialosky, J. E., & Donald, D. A. (2003). The effect of a fear-
avoidance-based physical therapy intervention for patients with acute low back pain:
results of a randomized clinical trial. Spine, 28(23), 2551-2560.
George, S. Z., Wittmer, V. T., Fillingim, R. B., & Robinson, M. E. (2007). Sex and pain-related
psychological variables are associated with thermal pain sensitivity for patients with
chronic low back pain. JPain, 8(1), 2-10.
Goubert, L., Crombez, G., Van Damme, S., Vlaeyen, J. W., Bijttebier, P., & Roelofs, J. (2004).
Confirmatory factor analysis of the Tampa Scale for Kinesiophobia: invariant two-factor
model across low back pain patients and fibromyalgia patients. Clin JPain, 20(2), 103-
Grotle, M., Vollestad, N. K., & Brox, J. I. (2006). Clinical course and impact of fear-avoidance
beliefs in low back pain: prospective cohort study of acute and chronic low back pain: II.
Spine, 31(9), 1038-1046.
Grotle, M., Vollestad, N. K., Veierod, M. B., & Brox, J. I. (2004). Fear-avoidance beliefs and
distress in relation to disability in acute and chronic low back pain. Pain, 112(3), 343-352.
Haddock CK, R. D., Shadish WR. (1998). Using Odds Ratios as Effect Sizes for Meta-Analysis
of Dichotomous Data: A Primer on Methods and Issues. Psychological M~ethods, 3(3), 3 39-
Hart, L. G., Deyo, R. A., & Cherkin, D. C. (1995). Physician office visits for low back pain.
Frequency, clinical evaluation, and treatment patterns from a U.S. national survey. Spine,
Hestback, L., Leboeuf-Yde, C., & Manniche, C. (2003). Low back pain: what is the long-term
course? A review of studies of general patient populations. Eur Spine J, 12(2), 149-165.
Klenerman, L., Slade, P. D., Stanley, I. M., Pennie, B., Reilly, J. P., Atchison, L. E., et al..
(1995). The prediction of chronicity in patients with an acute attack of low back pain in a
general practice setting. Spine, 20(4), 478-484.
Kopec, J. A., Esdaile, J. M., Abrahamowicz, M., Abenhaim, L., Wood-Dauphinee, S., Lamping,
D. L., et al.. (1996). The Quebec Back Pain Disability Scale: conceptualization and
development. J Clin Epidemiol, 49(2), 151-161.
Kottke, F. J. (1966). The effects of limitation of activity upon the human body. JamaJJ~~~~~JJJJJ~~~~~ 196(10),
Lethem, J., Slade, P. D., Troup, J. D., & Bentley, G. (1983). Outline of a Fear-Avoidance Model
of exaggerated pain perception--I. Behav Res Ther, 21(4), 401-408.
Linton, S. J. (2000). A review of psychological risk factors in back and neck pain. Spine, 25(9),
Linton, S. J., & van Tulder, M. W. (2001). Preventive interventions for back and neck pain
problems: what is the evidence? Spine, 26(7), 778-787.
Loeser, J. D. (1994). The prevention of needless pain: research opportunities. Prev M~ed, 23(5),
Mannion, A. F., Junge, A., Taimela, S., Muntener, M., Lorenzo, K., & Dvorak, J. (2001). Active
therapy for chronic low back pain: part 3. Factors influencing self-rated disability and its
change following therapy. Spine, 26(8), 920-929.
Mantyselka, P., Kumpusalo, E., Ahonen, R., Kumpusalo, A., Kauhanen, J., Viinamaki, H., et al..
(2001). Pain as a reason to visit the doctor: a study in Finnish primary health care. Pain,
Mayer, T. G. (1991). Rational for Modern Spinal Care. In T. G. Mayer, Mooney V, & Gatchel
RJ (Ed.), Contemporary Conservative Care for Painful Spinal Disorders (pp. 3-9).
Philadelphia: Lea & Febiger.
Miller RP, K. S., Todd DD. (1991). Thze Tampa Scale for Kinesiophobia (Unpublished Report).
Picavet, H. S., Vlaeyen, J. W., & Schouten, J. S. (2002). Pain catastrophizing and kinesiophobia:
predictors of chronic low back pain. Am JEpidemiol, 156(11), 1028-1034.
Pincus, T., Burton, A. K., Vogel, S., & Field, A. P. (2002). A systematic review of psychological
factors as predictors of chronicity/disability in prospective cohorts of low back pain. Spine,
Roelofs, J., Goubert, L., Peters, M. L., Vlaeyen, J. W., & Crombez, G. (2004). The Tampa Scale
for Kinesiophobia: further examination of psychometric properties in patients with chronic
low back pain and fibromyalgia. Eur JPain, 8(5), 495-502.
Roland, M., & Fairbank, J. (2000). The Roland-Morris Disability Questionnaire and the
Oswestry Disability Questionnaire. Spine, 25(24), 3115-3124.
Roland, M., Morris R. (1983). 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, 141-144.
Schappert, S. (1989). National Ambulatory M~edical Care Survey (No. Vital Health Stat 13):
National Center for Health Statistics.
Scheer, S. J., Radack, K. L., & O'Brien, D. R., Jr. (1995). Randomized controlled trials in
industrial low back pain relating to return to work. Part 1. Acute interventions. Arch Phys
M~ed Rehabil, 76(10), 966-973.
Severeijns, R., Vlaeyen, J. W., van den Hout, M. A., & Weber, W. E. (2001). Pain
catastrophizing predicts pain intensity, disability, and psychological distress independent
of the level of physical impairment. Clin JPain, 1 7(2), 165-172.
Sieben, J. M., Vlaeyen, J. W., Portegijs, P. J., Verbunt, J. A., van Riet-Rutgers, S., Kester, A. D.,
et al.. (2005). A longitudinal study on the predictive validity of the fear-avoidance model
in low back pain. Pain, 117(1-2), 162-170.
Slade, P. D., Troup, J. D., Lethem, J., & Bentley, G. (1983). The Fear-Avoidance Model of
exaggerated pain perception--II. Behav Res Ther, 21(4), 409-416.
Staerkle, R., Mannion, A. F., Elfering, A., Junge, A., Semmer, N. K., Jacobshagen, N., et al..
(2004). Longitudinal validation of the fear-avoidance beliefs questionnaire (FABQ) in a
Swiss-German sample of low back pain patients. Eur Spine J, 13(4), 332-340.
Stang, P., Von Korff, M., & Galer, B. S. (1998). Reduced labor force participation among
primary care patients with headache. J Gen Intern M~ed, 13(5), 296-302.
Sternbach, R. A. (1986). Pain and 'hassles' in the United States: findings of the Nuprin pain
report. Pain, 27(1), 69-80.
Stewart, W. F., Ricci, J. A., Chee, E., Morganstein, D., & Lipton, R. (2003). Lost productive
time and cost due to common pain conditions in the US workforce. JamaJJ~~~~~JJJJJ~~~~~ 290(18), 2443-
Sullivan, M. J., Thorn, B., Haythornthwaite, J. A., Keefe, F., Martin, M., Bradley, L. A., et al..
(2001). Theoretical perspectives on the relation between catastrophizing and pain. Clin J
Pain, 1 7(1), 52-64.
Swinkels-Meewisse, I. E., Roelofs, J., Schouten, E. G., Verbeek, A. L., Oostendorp, R. A., &
Vlaeyen, J. W. (2006). Fear of movement/(re)injury predicting chronic disabling low back
pain: a prospective inception cohort study. Spine, 31(6), 658-664.
Tacci, J. A., Webster, B. S., Hashemi, L., & Christiani, D. C. (1998). Healthcare utilization and
referral patterns in the initial management of new-onset, uncomplicated, low back workers'
compensation disability claims. J Occup Environ M~ed, 40(1 1), 958-963.
Turk, D., Melzack R. (2001). The measurement of pain and the assessment of people
experiencing pain. In D. C. Turk, Melzack R. (Ed.), Handbook ofpain assessment (2nd ed.
ed.). New York: Guilford.
Vlaeyen, J. W., Kole-Snijders, A. M., Boeren, R. G., & van Eek, H. (1995). Fear of
movement/(re)injury in chronic low back pain and its relation to behavioral performance.
Pain, 62(3), 363-372.
Vlaeyen, J. W., & Linton, S. J. (2000). Fear-avoidance and its consequences in chronic
musculoskeletal pain: a state of the art. Pain, 85(3), 317-332.
Waddell, G., Newton, M., Henderson, I., Somerville, D., & Main, C. J. (1993). A Fear-
Avoidance Beliefs Questionnaire (FABQ) and the role of fear-avoidance beliefs in chronic
low back pain and disability. Pain, 52(2), 157-168.
Walker, B. F., Muller, R., & Grant, W. D. (2004). Low back pain in Australian adults:
prevalence and associated disability. J2anipulative Physiol Ther, 2 7(4), 23 8-244.
Watson, P. J., Main, C. J., Waddell, G., Gales, T. F., & Purcell-Jones, G. (1998). Medically
certified work loss, recurrence and costs of wage compensation for back pain: a follow-up
study of the working population of Jersey. Br JRheumatol, 3 7(1), 82-86.
Woby, S. R., Watson, P. J., Roach, N. K., & Urmston, M. (2004). Are changes in fear-avoidance
beliefs, catastrophizing, and appraisals of control, predictive of changes in chronic low
back pain and disability? Eur JPain, 8(3), 201-210.
Alisa Diane Hassinger was born on May 21, 1973 in Harrisburg, Pennsylvania. The
youngest of two, she grew up in the greater Harrisburg area, graduating from Millersburg Area
High School in 1991. She earned her B.A. in psychology from Indiana University of
Pennsylvania. Upon graduating with honors in 1995, she began graduate studies in social work
at Arizona State University. Alisa graduated with Master of Social Work in 1997, and
subsequently accepted a commission to serve in the United States Army as a social work officer.
Upon her honorable discharge from the military in 2003, Alisa continued to work for the
Department of Defense at Fort Benning, Georgia as a case manager and therapist for soldiers
with deployment related mental health concerns. Alisa relocated to Gainesville, Florida in
August of 2005 to begin her graduate career in clinical psychology. Upon completion of the
M. S., Alisa will continue on in her program and apply for doctoral candidacy in the Department
of Clinical and Health Psychology at the University of Florida.