Pain 127 (2007) 52–62

www.elsevier.com/locate/pain

Psychosocial factors associated with pain intensity,

pain-related interference, and psychological functioning in persons
with multiple sclerosis and pain
Travis L. Osborne *, Mark P. Jensen, Dawn M. Ehde, Marisol A. Hanley, George Kraft
Department of Rehabilitation Medicine, Box 356490, 1959 NE Pacific Street, University of Washington School of Medicine,
Seattle, WA 98104-2499, USA

Received 9 March 2006; received in revised form 14 July 2006; accepted 24 July 2006

Abstract

Biopsychosocial models of chronic pain that recognize psychological and environmental factors as important aspects of adjust-
ment to pain have been proposed for understanding chronic pain and related suffering in persons with multiple sclerosis (MS), but
such models have not been empirically tested. The objective of this study was to test such a model by evaluating the associations of
several psychosocial variables (i.e., pain-related catastrophizing, perceived social support, pain beliefs, and pain coping) with pain
intensity, pain interference with functioning, and psychological functioning in persons with chronic pain and MS, after controlling
for demographic and disease-related factors. Participants were 125 community-dwelling persons with MS and pain who completed a
mailed questionnaire that included measures of pain intensity and interference, psychological functioning, catastrophizing, social
support, and pain beliefs and coping. The psychosocial variables accounted for an additional 25% of the variance in average pain
intensity after controlling for demographic and disease-related variables (p < .001). These variables explained an additional 22% of
the variance in pain-related interference (p < .001) and 43% of the variance in psychological functioning (p < .001), after adjusting
for demographic and MS-related variables and average pain intensity. Catastrophizing was consistently and independently associ-
ated with all criterion measures, whereas social support, pain beliefs, and pain coping were associated with some criterion measures
but not others. The results provide empirical support for a biopsychosocial understanding of chronic pain in MS and suggest that
specific psychosocial factors (e.g., catastrophizing) may be important regarding adjustment to pain in persons with MS.
Ó 2006 International Association for the Study of Pain. Published by Elsevier B.V. All rights reserved.

Keywords: Multiple Sclerosis; Chronic pain; Biopsychosocial; Catastrophizing

1. Introduction MS, pain symptoms are chronic (Stenager et al., 1991;

Indaco et al., 1994; Ehde et al., 2003), and for a substan-
Pain is common in persons with multiple sclerosis tial minority, pain is severe (Ehde et al., in press). Pain
(MS), with the majority of studies indicating that 43– in MS can arise from multiple etiologies, including acute
80% of persons with MS report pain (e.g., Moulin, pain due to inflammation, neuropathic pain related to
1989; Svendsen et al., 2003; Beiske et al., 2004; Solaro central nervous system lesions, pain secondary to spas-
et al., 2004; Ehde et al., in press). For many persons with ticity, spasms, and muscle cramps, and musculoskeletal
pain (Kraft and Cui, 2005). Research has attempted to
*
Corresponding author. Tel.: +1 206 374 0109; fax: +1 206 374
identify factors associated with the prevalence and
0108. severity of pain in persons with MS. However, few of
E-mail address: [email protected] (T.L. Osborne). the demographic (i.e., age, gender) and disease-related

0304-3959/$32.00 Ó 2006 International Association for the Study of Pain. Published by Elsevier B.V. All rights reserved.
doi:10.1016/j.pain.2006.07.017
 T.L. Osborne et al. / Pain 127 (2007) 52–62 53

(i.e., disease subtype, severity and duration) variables the survey study in two primary ways. The majority of potential
that have been examined have shown consistent associ- participants were randomly selected from a group of persons
ations with MS-related pain (Ehde et al., 2005). More- who had previously been contacted through the Multiple Scle-
over, few studies have investigated whether various rosis Association (MSA) of King County, WA, as part of a sur-
vey that assessed multiple dimensions of MS (Chwastiak et al.,
psychosocial factors, shown to be associated with pain
2002; Ehde et al., 2003; Bombardier et al., 2004) and indicated
and functioning in other chronic pain populations, are
that they would be willing to be contacted about future research
related to pain in MS. opportunities. A small subset of potential participants who
Psychosocial factors that have demonstrated consis- heard about our research program by word of mouth also con-
tent associations with functioning and adjustment in tacted our office to inquire about research opportunities and
persons who present with primary chronic pain prob- were subsequently enrolled into the survey study. The current
lems include pain-related catastrophizing cognitions, survey was mailed to 300 individuals (287 surveys were sent
pain-related beliefs/attributions, pain coping strategies, to individuals randomly selected from the prior MSA sample
and social/environmental variables (i.e., social support, and an additional 13 surveys were sent to individuals who con-
spouse responses to pain behaviors) (e.g., Jensen et al., tacted our research office directly). A consent form and a cover
1991; Jensen et al., 1994a; Jensen et al., 1994b, Jensen letter inviting potential study participants to participate in the
study accompanied each survey. Respondents were paid $25
et al., 2001; Sullivan et al., 2001). Research also suggests
for completing and returning the consent forms and survey.
that several of these factors are associated with pain and
The University of Washington Human Subjects Review Com-
adjustment to pain in persons with pain secondary to a mittee approved all study procedures.
physical disability (e.g., spinal cord injury, amputation, Of the 300 surveys sent, 13 were returned due to incorrect
and cerebral palsy) (cf. Engel et al., 2000; Jensen et al., addresses, 4 individuals were deceased, and another 8 individ-
2002; Turner et al., 2002; Hanley et al., 2004). Although uals were not eligible to participate in the study because they
similar psychosocial factors may be associated with indicated that they did not have MS. Of the 275 remaining
pain-related outcomes in persons with MS, research is possible participants, 187 returned completed surveys, for a
needed to substantiate this hypothesis. Additionally, response rate of 68.0%. (Of these 187 participants, 180 individ-
the specific psychosocial variables associated with pain uals were from the prior MSA sample. Demographic data and
and functioning in other samples of persons with chron- information about the nature and scope of pain problems
reported by these individuals were recently reported; see Ehde
ic pain may, or may not, be associated with outcomes in
et al., in press.) Only participants who indicated a current or
persons with MS. Identifying the psychosocial factors
recent pain problem were included in the current study’s anal-
associated with chronic pain in persons with MS may yses (n = 125).
have important treatment implications; though there
have been calls for the use of adjunct psychosocial pain 2.2. Measures
treatments for persons with MS and pain (Kerns, 2000),
there is minimal empirical evidence to guide the develop- 2.2.1. Demographics
ment or implementation of such treatment approaches. Participants were asked to provide basic demographic
The current study’s primary objective was to conduct information, including information about their gender, age,
a preliminarily test of a biopsychosocial model of pain in race/ethnicity, marital status, and employment status.
MS. Although biopsychosocial models of pain in MS
2.2.2. MS-related factors
have been proposed (Kerns et al., 2002), we were unable
Several MS disease-related characteristics were assessed,
to identify any published empirical tests of such models.
including MS disease course, MS disease severity, and disease
We hypothesized that a set of psychosocial variables duration. MS disease course was assessed by asking partici-
(perceived social support, pain-related catastrophizing, pants to select a pictorial graph that most closely corresponded
pain beliefs, and pain coping strategies) would be signif- to their disease course over time (Lublin and Reingold, 1996).
icantly associated with several key pain-related out- These graphs were accompanied by written descriptions of the
comes (pain intensity, pain-related interference with four clinical courses of MS that were recognized at the time of
functioning, and psychological functioning), even after the study: relapsing/remitting, secondary progressive, primary
controlling for demographic and biological/disease-re- progressive, and progressive relapsing (Lublin and Reingold,
lated variables. 1996). The self-report version (Bowen et al., 2001) of the
Expanded Disability Status Scale (EDSS) (Kurtzke, 1983)
was used to assess disease severity/progression and is highly
2. Methods correlated (r = .89) with the physician-administered EDSS
(Bowen et al., 2001). Respondents’ EDSS scores were grouped
2.1. Participants and procedures into two categories based on whether they reported being inde-
pendent in ambulation (EDSS scores between 0 and 5.5) or
Participants were drawn from a sample of persons with MS reported requiring an assistive device (e.g., cane, walker,
who completed a survey on quality of life in persons with dis- wheelchair) for ambulation (EDSS scores between 6.0 and
abilities that included a number of pain-related assessment 9.5). Finally, participants were asked to provide the date of
measures (Ehde et al., in press). Individuals were recruited for their MS diagnosis in order to determine disease duration.
54 T.L. Osborne et al. / Pain 127 (2007) 52–62

2.2.3. Pain intensity as = .81–.95) and test–retest reliability (rs = .75–.80) (Ware
Survey respondents were asked to indicate the presence or et al., 2000). Scores on the Mental Health scale range from 0
absence of any recent pain problem(s) by answering the follow- to 100, with higher scores indicating better psychological func-
ing question: ‘‘Are you currently experiencing, or have you in tioning. Mental Health scale scores have shown a consistent
the past three months experienced, any pain (other than occa- pattern of association with various indices of mental health
sional headaches or menstrual cramps)?’’ Participants who (e.g., depressive symptoms, suicidal ideation, and inpatient
responded ‘‘yes’’ to this question were included in the current mental health treatment in the past 12 months) in the expected
study, and were then asked to rate their average pain in the directions (Ware et al., 2000). The scale demonstrated good
past week on a 0–10 Numeric Rating Scale (NRS), with internal consistency in the current sample (Cronbach’s
0 = ‘‘No pain’’ and 10 = ‘‘Pain as bad as could be.’’ Prior a = .87).
research supports the validity of such 0–10 scales; these scales
are strongly associated with other measures of pain intensity 2.2.6. Pain coping
and are also responsive to changes in pain in response to pain The Chronic Pain Coping Inventory (CPCI) (Jensen et al.,
treatments (Jensen and Karoly, 2001). 1995) and a measure of pacing as a coping strategy developed
to be administered with the CPCI items (Nielson et al., 2001)
2.2.4. Pain interference were used to assess the frequency with which respondents use
Pain-related interference with various domains of func- a variety of coping strategies in response to pain. The
tioning was assessed using a modified version of the Brief 70-items that make up the following nine coping domain
Pain Inventory (BPI) Interference scale (Daut et al., 1983; subscales were administered: Guarding, Resting, Asking for
Cleeland and Ryan, 1994). The original version of this scale Assistance, Relaxation, Task Persistence, Exercise/Strength,
asks respondents to rate the extent to which they perceive Seeking Social Support, Coping Self-Statements, and Pacing.
that pain interferes with seven activities of daily life, includ- The single CPCI item assessing medication use was not
ing general activity, mood, walking ability, normal work, administered. For each item, respondents indicate the number
relations with other people, sleep, and enjoyment of life. of days in the past week (0–7) that they used the coping strat-
We modified item 3 (‘‘Walking ability’’) to read ‘‘Mobility egy described at least once to cope with pain. Subscale scores
(ability to get around),’’ given that some participants in our are calculated by averaging the ratings for the items that
sample cannot walk. Consistent with our prior research with comprise each scale. Thus, the possible range of scores for
persons with disabilities and pain (Tyler et al., 2002; Hanley each subscale is 0–7, with higher scores indicating more fre-
et al., 2004; Jensen et al., 2005), we also added three items quent use of domain-specific coping strategies. The CPCI
(yielding a 10-item scale) to assess perceived pain interference subscales and the Pacing scale have been found to have
with additional areas of functioning: self-care, recreational adequate to excellent internal consistency (Cronbach’s as =
activities, and social activities. These items were added to .74–.91) and adequate test–retest reliability over a mean inter-
the scale in order to assess a wider range of functional val of 12.6 weeks (rs = .47–.78) (Jensen et al., 1995; Nielson
domains important to persons with disabilities that can be et al., 2001). The internal consistency of the CPCI subscales
affected by pain. The total score for the modified version of and Pacing scale in the current sample ranged from adequate
the BPI Interference scale is calculated by averaging the rat- to excellent (Cronbach’s as = .76–.91). Validity for the CPCI
ings for the 10 items. The possible range for the total score is subscales has been demonstrated by correlations between
0–10, with higher scores indicating a greater degree of pain- patient and significant other versions of the scale (rs =
related interference with functioning. The original 7-item .13–.65), as well as significant relationships between a number
BPI Interference scale has demonstrated excellent internal of the subscales and measures of disability and pain-related
consistency and validity in samples of persons with cancer distress (Jensen et al., 1995).
pain (Cleeland and Ryan, 1994), chronic non-malignant pain
(Tan et al., 2004), as well as arthritis and low back pain (Kel- 2.2.7. Pain cognitions and beliefs
ler et al., 2004). Additionally, the modified 10-item version of The Catastrophizing scale of the Coping Strategies Ques-
the scale has demonstrated excellent internal consistency tionnaire (CSQ; Rosenstiel and Keefe, 1983) was used to assess
(Cronbach’s as = .89–.95) and validity in samples of persons pain-related catastrophizing cognitions. The 6-item scale
with pain secondary to cerebral palsy (Tyler et al., 2002), assesses the frequency of catastrophizing cognitions in
limb loss (Hanley et al., 2004), and spinal cord injury (Jensen response to pain and items are rated on a 0 (‘‘never do that’’)
et al., 2005). The internal consistency of the modified 10-item to 6 (‘‘always do that’’) scale. The Catastrophizing scale score
version of the BPI Interference scale was excellent in the cur- is calculated by averaging the ratings across the 6 items. The
rent sample (Cronbach’s a = .95). The construct and concur- range of possible scores is 0–6, with higher scores indicating
rent validity of the scale in the current sample has previously more frequent pain-related catastrophizing. The CSQ Catas-
been reported (Osborne et al., in press). trophizing scale is frequently used in pain research and has
demonstrated adequate to excellent internal consistency across
2.2.5. Psychological functioning a number of samples (e.g., Rosenstiel and Keefe, 1983; Keefe
The 5-item Mental Health scale from the SF-36 (Ware et al., 1989). The internal consistency of the scale was good
et al., 2000) was used to assess global psychological function- in the current sample (Cronbach’s a = .89). The validity of
ing. The Mental Health scale is strongly correlated (r = .95) the Catastrophizing scale has been demonstrated by its associ-
with the longer 38-item Mental Health scale from which it ation with measures of depression and adjustment (Keefe and
was derived and is a commonly used measure of psychological Williams, 1990; Sullivan and D’Eon, 1990; Jensen et al., 1992;
functioning with established internal consistency (Cronbach’s Tan et al., 2001; Jensen et al., 2002; Turner et al., 2002) and its
 T.L. Osborne et al. / Pain 127 (2007) 52–62 55

sensitivity to change in response to pain treatments believed to were more frequent than high scores in this sample. This var-
alter catastrophizing (Jensen et al., 1994a). iable was transformed using a square root function to produce
The 57-item Survey of Pain Attitudes scale (SOPA) (Jensen a more normal distribution of scores. We then conducted the
et al., 1987; Jensen et al., 1994b) was used to assess pain-spe- primary regression analyses for the study (described below)
cific beliefs believed to be important in adjustment to chronic twice, once using the transformed variable and once using
pain. The SOPA is comprised of seven subscales that assess the non-transformed variable. As there were no substantive
the following domains of pain beliefs: Control (belief in one’s differences in the findings between the two sets of analyses,
control over pain), Disability (belief that one is disabled by we present the results for the analyses that included the non-
pain), Harm (belief that pain is a signal of bodily damage transformed CSQ Catastrophizing scores in order to facilitate
and that activity should be avoided), Emotion (belief that emo- interpretability of the findings.
tions influence pain), Medical Cure (belief in a medical cure for Descriptive statistics (e.g., means, standard deviations)
pain), Solicitude (belief that others should be solicitous when were calculated and reported for each of the primary variables
responding to pain behaviors), and Medication (belief in the of interest as, to our knowledge, prior studies on pain in MS
appropriateness of medications for the treatment of chronic have not utilized several of the measures that we included in
pain). Items are rated on a 0 (‘‘This is very untrue for me’’) our survey (e.g., CSQ Catastrophizing scale, SOPA, and
to 4 (‘‘This is very true for me’’) scale, reflecting the extent CPCI).
to which respondents agree with each of the pain beliefs To test the utility of a biopsychosocial model for under-
described. SOPA subscale scores are calculated by averaging standing pain in persons with MS, we conducted three hierar-
the ratings for the items that comprise each subscale. The chical multiple regression analyses, one for each of our
range of possible responses for each subscale is 0–4, with high- primary criterion variables: pain intensity, pain-related inter-
er scores indicating a greater degree of endorsement of the pain ference, and psychological functioning. For each analysis, we
beliefs. The SOPA subscales have demonstrated adequate to entered a set of demographic and MS-related (biological) vari-
good internal consistency (Cronbach’s as = .71–.81) and ade- ables as a block in the first step (gender, age, and disease sever-
quate test–retest reliability over an average of 6 weeks ity). In the analyses examining pain interference and
(rs = .63–.68) (Jensen et al., 1994b). Previous studies also pro- psychological functioning, we then entered average pain inten-
vide support for the subscales’ validity (Jensen et al., 1987; sity ratings in the second step. Finally, in all three analyses, the
Strong et al., 1992; Jensen et al., 1994b). Internal consistency psychosocial variables (measures of social support, pain-relat-
coefficients were adequate in the current sample for five of ed catastrophizing, pain beliefs, and pain coping strategies)
the subscales (Cronbach’s as = .72–.77), however, the Harm were entered as a block in the last step. We selected hierarchi-
and Medical Cure subscales demonstrated marginal internal cal regression as our multivariate analytic strategy given that
consistency (.59 and .60, respectively). our primary hypothesis was that the psychosocial variables
would account for a significant proportion of the variance in
2.2.8. Social support our criterion measures even after controlling for biological
Perceived social support was assessed using the Multidi- variables (demographic and disease-related variables).
mensional Scale of Perceived Social Support (MSPSS) (Zimet Due to the study’s sample size and the large number of sub-
et al., 1988). The 12-item scale assesses the extent to which scales that comprise the SOPA and CPCI, it was not possible
respondents perceive that various forms of support are avail- to enter each of these subscale scores into the block of the
able to them from members of their social support network. regression analyses containing the psychosocial variables and
Items are rated on a 1 (‘‘strongly disagree’’) to 7 (‘‘strongly still retain an adequate sample size to independent variable
agree’’) scale. The MSPSS yields a global perceived support ratio (Tabachnick and Fidell, 1996). Therefore, in order to
score, as well as subscale scores for support from family, reduce the number of variables representing these constructs,
friends, and a significant others. The total score was used in the subscale scores for these measures were subjected to two
the current study and is calculated by averaging the ratings principal components analyses (PCA), one for the SOPA and
for all 12 items. Thus, the range for the total MSPSS score is one of the CPCI. PCA was selected because this method pro-
1–7, with higher scores indicating greater perceived support. vides a unique solution that extracts maximum variance from a
The MSPSS total score has been shown to have excellent inter- data set, thereby providing a summary of the data (as opposed
nal consistency (Cronbach’s as = .88–.91), strong test–retest to Factor Analysis, which provides a theoretical solution; see
reliability over a 2–3 month period (r = .85), and established Tabachnick and Fidell, 1996). An orthogonal rotation (Vari-
validity as evidenced by a significant negative association with max) of the PCA solution was selected over other rotations
depressive symptoms (Zimet et al., 1988; Dahlem et al., 1991). because this approach maximizes the variance within compo-
Responses to the MSPSS have also been shown to be unrelated nents, across variables, thereby simplifying the components
to scores on a brief measure of social desirability (Dahlem and making them easier to interpret (Tabachnick and Fidell,
et al., 1991). The internal consistency of the MSPSS was excel- 1996). Component scores were computed and saved as vari-
lent in the current sample (Cronbach’s a = .93). ables (using the regression method for estimating factor scores
provided by SPSS 12.0) so that they could be included in the
2.3. Overview of data analysis primary regression analyses.
The PCA of the seven SOPA subscales resulted in a three-
We inspected the distributions of the primary variables of component solution. These three components accounted for
interest to ensure that they met the assumptions of the planned 66% of the variance in pain beliefs. The eigenvalues for the
statistical tests. The CSQ Catastrophizing scale was somewhat three components were 2.13, 1.43, and 1.09 and decreased to
negatively skewed, indicating that low scores on this measure 0.76 for the fourth component. Three subscales loaded on
56 T.L. Osborne et al. / Pain 127 (2007) 52–62

the first component (Control, .86; Disability, .65; Harm, .76), mean number of years since MS diagnosis in the sample
whereas two subscales loaded on the second (Emotion, .83; was 12.85 (SD = 10.04).
Solicitude, .76) and third (Medication, .83; Medical Cure, Participants reported having pain symptoms for an
.73) components. We labeled the components Pain as Illness average of 8.75 years (SD = 8.88). Regarding average
Beliefs, Affective and Social Beliefs, and Medically Focused
pain severity in the past week, 1.6% reported no pain
Beliefs, respectively. These three-component scores were
(0 on 0–10 rating scale), 46.4% reported mild pain
entered into the primary regression analyses.
Two components emerged in the PCA of the nine subscales (1–4), 27.2% reported moderate pain (5–6), and 24.8%
of the CPCI that accounted for 54% of the variance in pain reported severe pain (8–10).
coping. The eigenvalues for these components were 3.59 and
1.26 and decreased to 0.98 for the third component. Seven sub- 3.2. Descriptive statistics
scales loaded on the first component (Guarding, .81; Resting,
.77; Asking for Assistance, .63; Relaxation, .55; Coping Self- Means and standard deviations for the psychosocial
Statements, .60; Seeking Social Support, .66; Pacing, .63) variables and the primary criterion variables are present-
and two subscales loaded on the second component (Task Per- ed in Table 1.
sistence, .79; Exercise/Strength, .63). We labeled the first factor
General Pain Coping and the second factor Persistence/Exer-
3.3. Concurrent prediction of pain intensity, pain
cise Pain Coping. These two component scores were entered
interference, and psychological functioning
into the primary regression analyses.
To facilitate interpretation of the multivariate findings, we
also planned to conduct a series of univariate analyses to Results of the multiple regression analyses are pre-
examine the associations between our criterion variables (pain sented in Table 2. In the regression model predicting
intensity, pain-related interference, and psychological func- average pain intensity in the past week (assessed using
tioning) and the psychosocial variables, in the event that the a 0–10 NRS), the demographic and MS variables
regression analyses provided support for our primary hypoth- (gender, age, and disease severity, entered as a block)
esis. For these analyses, we computed partial correlation coef- only explained 6% of the variance (R2 = .06, F = 2.40,
ficients. We controlled for age, gender, and disease severity for ns). In contrast, the psychosocial variables (social sup-
all coefficients, and also controlled for average pain intensity port, pain-related catastrophizing, pain beliefs, and
for coefficients examining pain interference and psychological
pain coping, entered as a block) accounted for an
functioning.
additional and statistically significant 24% of the var-
iance in average pain intensity, even after controlling
3. Results for the demographic and MS-related variables
(R2D = .24, FD = 5.48, p < .001). Of the psychosocial
3.1. Sample characteristics variables, pain-related catastrophizing (b = 0.20,
p < .05), the Pain and Illness Beliefs (b = 0.32,
The sample (n = 125) was predominantly female p < .01) and Affective and Social Beliefs (b = 0.19,
(75.2%) with a mean age of 50.79 years (SD = 10.77 p < .05) components of the SOPA, and the Persis-
years). The majority of participants reported their eth- tence/Exercise Pain Coping component of the CPCI
nicity as Caucasian (96.8%), with the remaining partici- (b = 0.26, p < .01) were each independent concurrent
pants reporting their ethnicity as Asian-American predictors of pain intensity.
(1.6%), Native American (1.6%), African-American For the regression model predicting pain interference
(0.8%), Hispanic/Chicano (0.8%), and Pacific Islander scores (10-item modified BPI Interference scale), the
(0.8%). Most respondents were either married or living demographic and MS variables accounted for 10% of
with a significant other (68.0%). Only a quarter of the the variance (R2 = .10, F = 4.10, p < .01). However, of
sample was employed full or part-time (25.6%). Regard- these variables, only disease severity was a significant
ing the education level of the sample, the majority of independent predictor (b = 0.31, p < .001). After con-
participants were either college graduates (29.6%), had trolling for these variables, average pain intensity rat-
completed some college (28.8%), or attended graduate ings accounted for an additional and substantial 35%
school (24.0%). of the variance (R2D = .35, FD = 72.09, p < .001;
With regard to MS-related variables, participants b = 0.61, p < .001). Finally, when the psychosocial vari-
reported a range of MS disease courses: 53.3% relaps- ables were added to the model in the third step, they
ing/remitting, 23.8% secondary progressive, 13.1% pri- accounted for an additional 22% of the variance in pain
mary progressive, and 9.8% progressive relapsing. In interference scores (R2D = .22, FD = 10.37, p < .001).
terms of disease severity, approximately 46% of the sam- Among the psychosocial variables, pain-related catas-
ple reported being independent with ambulation (EDSS trophizing (b = 0.22, p < .01) and the General Pain Cop-
scores between 0 and 5.5), whereas the remaining 54% of ing component of the CPCI (b = 0.28, p < .001) were the
participants indicated the use of an assistive device for only significant independent concurrent predictors of
ambulating (EDSS scores between 6.0 and 9.5). The pain interference.
 T.L. Osborne et al. / Pain 127 (2007) 52–62 57

Table 1
Descriptive data for pain and psychosocial variables and partial correlations between psychosocial variables and pain intensity, pain interference, and
psychological functioning (n = 125)
Variable Mean (SD) Correlation with
Average pain intensity BPI Interference SF-36 Mental Health
Average pain intensity 4.80 (2.37) – – –
BPI Interference 3.44 (2.50) – – –
SF-36 Mental Health 66.21 (19.54) – – –
MSPSS 5.34 (1.25) .21b .30* .44*
CSQ Catastrophizing 1.14 (1.19) .31* .44* .53*
SOPA
Control 2.09 (0.73) .27a .02 .25a
Disability 2.08 (0.81) .27a .52* .36*
Harm 1.72 (0.66) .15 .10 .11
Emotion 1.83 (0.75) .16 .29* .48*
Medication 2.66 (0.82) .15 .10 .02
Solicitude 1.62 (0.86) .08 .29* .25a
Medical Cure 1.60 (0.64) .11 .06 .03
CPCI
Guarding 3.05 (1.66) .17 .45* .13
Resting 3.44 (2.02) .25a .35* .19b
Asking for Assistance 2.57 (2.28) .02 .08 .04
Relaxation 1.99 (1.55) .11 .16 .11
Task persistence 4.51 (1.71) .01 .24a .07
Exercise/Strength 2.73 (1.79) .13 .19b .28a
Coping self-statements 2.95 (1.96) .14 .27a .15
Seeking Social Support 2.07 (1.60) .07 .12 .03
Pacing 3.53 (1.97) .15 .18 .01
Note. BPI Interference, modified Brief Pain Inventory Interference scale; MSPSS, Multidimensional Scale of Perceived Social Support; CSQ, Coping
Strategies Questionnaire; CPCI, Chronic Coping Pain Inventory; SOPA, Survey of Pain Attitudes.
a
p < .01.
b
p < .05.
*
p < .0028 (Bonferroni correction).

In the final regression model predicting psychological each association was significantly different from zero.
functioning (SF-36 Mental Health scale), the demo- Using a Bonferroni correction to determine statistical
graphic and MS variables explained only 4% of the var- significance, in addition to computing partial correlation
iance (R2 = .04, F = 1.71, ns). After controlling for these coefficients, is a very conservative test of the associations
variables, average pain intensity ratings only accounted between the variables we examined. Therefore, we
for an additional 2% of the variance in psychological report on significant findings as well as non-significant
functioning scores (R2D = .02, FD = 2.57, ns). However, trends (p < .05) in order to avoid prematurely dismissing
when the psychosocial variables were added to the mod- variables that may be clinically meaningful and worthy
el in the final step, they accounted for an additional, and of additional research.
statistically significant, 43% of the variance in scores on Perceived social support was significantly associated
the SF-36 Mental Health scale (R2D = .43, FD = 13.26, with pain interference (r = .30) and psychological
p < .001). The psychosocial variables that were signifi- functioning (r = .44) at the univariate level. Greater
cant independent concurrent predictors of psychological levels of perceived social support were associated with
functioning included perceived social support (b = 0.26, decreased interference in functioning due to pain and
p < .001), pain-related catastrophizing (b = 0.38, better psychological functioning. A non-significant
p < .001), and the Affective and Social Beliefs compo- trend was also observed between greater perceived
nent of the SOPA (b = 0.27, p < .001). support and decreased pain intensity (r = .21,
p < .05).
3.4. Univariate analyses Pain-related catastrophizing was significantly related
to all three of our criterion measures. Catastrophizing
Partial correlations between the psychosocial vari- was moderately associated with average pain intensity
ables and our primary criterion variables are presented (r = .31) and strongly associated with both pain-related
in Table 1. Because of the large number of correlations interference with functioning (r = .44) and global psy-
performed (18 psychosocial predictor variables per crite- chological functioning (r = .53). In all three cases,
rion measure), we used a Bonferroni correction for each higher catastrophizing scores were related to poorer
criterion measure (.05/18 = .0028) to determine whether outcomes.
58 T.L. Osborne et al. / Pain 127 (2007) 52–62

Table 2
Concurrent prediction of average pain intensity, pain-related interference with functioning, and psychological functioning (n = 124)
R2 R2 change F change b
Average pain in the past week
1. Demographic/MS variables .06 .06 2.40
Gender .03
Age .02
Disease severity .24**
2. Psychosocial variables .30 .24 5.48***
Social support (MSPSS) .13
Catastrophizing (CSQ) .20*
Pain as Illness Beliefs (SOPA) .32**
Affective and Social Beliefs (SOPA) .19*
Medically Focused Beliefs (SOPA) .11
General Pain Coping (CPCI) .19
Persistence/Exercise Pain Coping (CPCI) .26**
Pain interference (BPI Interference scale)
1. Demographic/MS variables .10 .10 4.10**
Gender .09
Age .08
Disease severity .31***
2. Average pain intensity .44 .35 72.09*** .61***
3. Psychosocial variables .67 .22 10.37**
Social support (MSPSS) .12
Catastrophizing (CSQ) .22**
Pain as Illness Beliefs (SOPA) .03
Affective and Social Beliefs (SOPA) .11
Medically Focused Beliefs (SOPA) .04
General Pain Coping (CPCI) .28***
Persistence/Exercise Pain Coping (CPCI) .12
Psychological functioning (SF-36 Mental Health scale)
1. Demographic/MS variables .04 .04 1.71
Gender .03
Age .20*
Disease severity .10
2. Average pain intensity .06 .02 2.57 .15
3. Psychosocial variables .50 .43 13.26***
Social support (MSPSS) .26***
Catastrophizing (CSQ) .38***
Pain as Illness Beliefs (SOPA) .12
Affective and Social Beliefs (SOPA) .27***
Medically Focused Beliefs (SOPA) .10
General Pain Coping (CPCI) .00
Persistence/Exercise Pain Coping (CPCI) .00
Note. MSPSS, Multidimensional Scale of Perceived Social Support; CSQ, Coping Strategies Questionnaire; SOPA, Survey of Pain Attitudes; CPCI,
Chronic Pain Coping Inventory.
*
p < .05.
**
p < .01.
***
p < .001.

Several of the SOPA subscales demonstrated signifi- emotions impact the experience of pain) were also signif-
cant univariate relationships and trends in association icantly related to both greater pain interference (r = .29)
with the criterion variables, with the Disability subscale and poorer psychological functioning (r = .48). Fur-
demonstrating the most consistent pattern of associa- thermore, the Solicitude subscale was significantly relat-
tions. Greater endorsement of the belief that one is dis- ed to pain interference (r = .29), with more strongly held
abled by pain was significantly associated with increased beliefs that others should respond solicitously in
pain-related interference with functioning (r = .52) and response to expressions of pain being associated with
poorer psychological functioning (r = .36). There was greater pain interference. Higher scores on this subscale
also a trend toward significance for the association were also associated with poorer psychological function-
between stronger disability beliefs and increased pain ing, though this trend was not significant (r = .25,
intensity (r = .27, p < .01). Higher scores on the Emo- p < .01). Although the Control subscale was not signifi-
tion subscale (i.e., greater endorsement of the belief that cantly related to any of the three criterion variables,
 T.L. Osborne et al. / Pain 127 (2007) 52–62 59

non-significant trends were observed between beliefs for treating pain in persons with MS. Changes in catas-
that one has greater control over pain and decreased trophizing during multidisciplinary pain treatment are
pain intensity (r = .27, p < .01) and better global psy- associated with improved outcomes for persons with
chological functioning (r = .25, p < .01). chronic pain (Jensen et al., 1994a; Turner et al., 2000;
Few of the pain coping strategies measured by the Jensen et al., 2001) and a recent study suggests that cog-
CPCI demonstrated significant relationships with the nitive restructuring may reduce pain intensity in persons
dependent variables, although the Guarding (r = .45) with disabilities (Ehde and Jensen, 2004). Clinical trials
and Resting (r = .35) subscales demonstrated moderate are needed to determine whether these findings extend to
to strong relationships with pain-related interference persons with MS-related pain. Interestingly, less than
with functioning; increased use of each of these coping 10% of participants in the current study had high scores
strategies was associated with greater pain interference. on the CSQ Catastrophizing scale (scores greater than 3
Non-significant trends were also observed for associa- on a 0–6 scale), suggesting that frequent catastrophizing
tions between increased pain interference and decreased may be an important factor to assess and treat in per-
task persistence (r = .24, p < .01), decreased exercise sons with MS and pain, but may not be that common.
(r = .19, p < .05), and increased use of coping self- Research is needed to substantiate this hypothesis,
statements (r = .27, p < .01). Although none of the although the mean CSQ Catastrophizing score in our
CPCI subscales were significantly related to either aver- sample (M = 1.14, SD = 1.19) was comparable to those
age pain intensity or psychological functioning, several for samples of persons with cerebral palsy and pain
non-significant trends were evident. Greater frequency (M = 1.69, SD = 1.34; Engel et al., 2000) and limb loss
of resting in response to pain was associated with greater and pain (M = 1.12, SD = 1.54; Hill et al., 1995).
average pain intensity (r = .25, p < .01) and poorer psy- The role of the social environment in adjustment to
chological functioning (r = .19, p < .05). Also, pain in MS was less clear, as social support was related
increased exercise was associated with better psycholog- to only one of the criterion variables in the primary
ical functioning (r = .28, p < .01). analyses. Greater perceived social support was associat-
ed with better psychological functioning, even after con-
4. Discussion trolling for pain intensity, consistent with the previous
findings that increased social support is associated with
The major conclusion of this study is that psychoso- lower levels of depressed mood and anger in persons
cial factors contributed significantly to the concurrent with other pain problems (Feldman et al., 1999). Longi-
prediction of pain intensity, pain-related interference tudinal and prospective research is needed with persons
with functioning, and psychological functioning, provid- with MS and pain to determine whether perceived social
ing preliminary empirical evidence that a biopsycho- support impacts psychological functioning or whether
social model is viable for understanding pain and persons with poorer psychological functioning perceive
adjustment to pain in MS. Additionally, specific psycho- themselves as having less available social support. If
social variables may be more strongly related to pain, research supports the former assertion, this might sug-
functioning, and adjustment than others. These results gest a benefit to incorporating interventions designed
have important implications for understanding and to enhance social support into pain treatments for per-
treating pain in persons with MS. sons with MS.
Pain-related catastrophizing demonstrated the most Perceived social support was not associated with pain
robust pattern of associations with the criterion mea- intensity or pain interference in the multivariate analy-
sures. Increased catastrophizing was significantly associ- ses. This is inconsistent with studies indicating that low-
ated with greater pain intensity, increased pain er levels of social support are associated with increased
interference, and poorer psychological functioning. This pain in persons with cancer (Zaza and Baine, 2002)
is consistent with research indicating that catastrophiz- and regional pain syndrome (Feldman et al., 1999),
ing is associated with poorer outcomes in persons for and with both pain intensity and functional disability
whom chronic pain is the primary problem (Boothby in persons with rheumatoid arthritis (Evers et al.,
et al., 1999) or secondary to a disability (Hill et al., 2003). However, at least one study also found that per-
1995; Engel et al., 2000; Turner et al., 2002; Whyte ceived social support was unrelated to pain interference
and Carroll, 2004). Thus, catastrophizing may be related in persons with MS (Osborne et al., 2006), indicating
to pain and pain-associated outcomes in similar ways that more research is needed to determine whether the
for persons with MS and pain as it is for persons with findings regarding social support from the chronic pain
other chronic pain conditions. Although we cannot literature generalize to this population.
determine the direction of causality between catastro- We conducted principal components analyses of the
phizing and pain-related outcomes due to this study’s pain belief subscales of the SOPA and pain coping sub-
cross-sectional design, interventions that target catastro- scales of the CPCI to reduce the number of predictor
phizing (e.g., cognitive restructuring) may hold promise variables in our regression models. Consequently, the
60 T.L. Osborne et al. / Pain 127 (2007) 52–62

findings regarding the types of beliefs and coping that be made. Longitudinal and experimental research is
are related to our criterion measures are less clear, needed to better understand the nature of the relation-
because the predictor variables generated from these ships amongst the variables and to determine whether
analyses represent an amalgam of diverse pain beliefs changes in specific psychosocial variables are associated
and coping strategies. We therefore focus on general with changes in pain-related criterion measures. Second,
patterns in the univariate analyses to discuss specific our survey consisted only of self-report measures. Thus,
beliefs and coping styles that may be important when some of the associations observed between variables
conceptualizing and treating MS-related pain. However, may be due in part to shared method variance. Future
these findings are tentative due to the large number of research would benefit from the inclusion of data col-
analyses performed. lected from other sources (e.g., spouses/partners, health
Overall, pain beliefs were more strongly associated care providers), as well as observational measures of
with pain interference and psychological functioning functional ability. Third, we assessed MS disease sever-
than pain intensity. Three specific beliefs – that one is dis- ity using the self-report version of the EDSS. Though it
abled due to pain (Disability), that emotions influence may be preferable to assess disease severity via physician
pain (Emotion), and that others should be solicitous in examination, this might have precluded some people
response to expressions of pain (Solicitude) – were all sig- from participating; persons who live in remote areas
nificantly associated with greater levels of pain interfer- or who have medical issues that might make it difficult
ence. These beliefs were also associated with poorer for them to come to the medical center for such an eval-
psychological functioning, as was the belief that one can- uation could have been underrepresented. The self-re-
not control one’s pain (Control), though these latter port version of the EDSS also correlates highly with
associations were not statistically significant. Non-signif- the physician-administered version. Finally, the current
icant trends were also observed for the associations sample was drawn primarily from persons with MS
between beliefs about disability and control and greater who were registered with the MSA. Membership in the
pain intensity. These findings are consistent with prior MSA may not be reflective of the general MS popula-
research indicating that pain beliefs related to disability tion and this may have resulted in a biased sample.
and control are associated with poorer functioning and Additional research is needed to replicate the current
adjustment to chronic pain (e.g., Jensen et al., 1994a; findings and to determine whether they are generalizable
Jensen et al., 1994b, Turner et al., 2000) and suggest a to the general MS population.
need for further evaluation of the relationships between Despite these limitations, this study advances our
these beliefs and outcomes in persons with MS to deter- understanding of MS-related pain by providing preli-
mine whether pain treatments should target these beliefs. minary empirical support for the utility of a biopsycho-
A clear pattern emerged regarding the relationships social model for understanding pain in persons with MS
between pain coping and our pain-related criterion mea- and for specific psychosocial variables that may be relat-
sures: pain coping strategies were more strongly associat- ed to pain and adjustment in this population. Catastro-
ed with pain-related interference with functioning than phizing showed a consistent pattern of associations with
with pain intensity or psychological functioning. Specifi- the criterion measures, suggesting that this cognitive
cally, coping strategies involving guarding, resting, task attributional style may have particular importance in
persistence, exercise/strengthening, and the use of coping the experience of and adjustment to pain in MS (as with
self-statements were all associated with pain interference chronic pain more generally). The data also suggest that
(associations were non-significant for the latter three efforts to treat pain in MS from a purely biomedical per-
strategies). Consistent with the view of guarding and rest- spective may be insufficient, and that pain and suffering
ing as maladaptive pain coping strategies, greater use of may be better alleviated by adopting a biopsychosocial
these strategies was associated with greater pain interfer- approach that addresses underlying physiological causes
ence. Conversely, more frequent use of task persistence of pain, as well as relevant psychosocial factors. Incor-
and exercise was associated with decreased pain interfer- porating existing empirically supported psychosocial
ence. The relationship between greater use of coping self- interventions for pain (e.g., cognitive-behavioral thera-
statements and increased pain interference is unclear. py) that target specific factors associated with pain-relat-
This form of coping is generally perceived as adaptive; ed outcomes (e.g., catastrophizing) into treatment
however, people for whom pain interferes more signifi- protocols for MS-related pain may hold promise for
cantly with functioning may have a greater need to use enhancing pain management in this population.
coping statements. As these findings are tentative, more
research is needed to determine their generalizability.
Several methodological limitations of the current Acknowledgements
study warrant consideration. First, we utilized a cross-
sectional research design. Consequently, causal inferenc- This research was supported by a Grant ‘‘Manage-
es about the nature of the relationships observed cannot ment of Chronic Pain in Rehabilitation’’ PO1 HD/
 T.L. Osborne et al. / Pain 127 (2007) 52–62 61

NS33988, from the National Institutes of Health, Hanley MA, Jensen MP, Ehde DM, Hoffman AJ, Patterson DR,
National Institute of Child Health and Human Develop- Robinson LR. Psychosocial predictors of long-term adjustment to
lower-limb amputation and phantom limb pain. Disabil Rehabil
ment (National Center for Medical Rehabilitation Re- 2004;26:882–93.
search) and the National Institute of Neurological Hill A, Niven CA, Knussen C. The role of coping in adjustment to
Disorders and Stroke and Grant H13309800017 from phantom limb pain. Pain 1995;62:79–86.
the National Institute on Disability and Rehabilitation Indaco A, Iachetta C, Nappi C, Socci L, Carrieri PB. Chronic and
Research to the University of Washington MS Rehabil- acute pain syndromes in patients with multiple sclerosis. Acta
Neurol (Napoli) 1994;16:97–102.
itation Research and Training Center. The authors Jensen MP, Ehde DM, Hoffman AJ, Patterson DR, Czerniecki JM,
gratefully acknowledge the contributions of Amy Hoff- Robinson LR. Cognitions, coping and social environment predict
man, Lindsay Washington, Emily Phelps, Laura adjustment to phantom limb pain. Pain 2002;95:133–42.
Nishimura, Kristin McArthur, Kevin Gertz, Noel Pere- Jensen MP, Hoffman AJ, Cardenas D. Chronic pain in individuals
yra-Johnston, Silvia Amtmann, Joe Skala, and Kerry with spinal cord injury: a survey and longitudinal study. Spinal
Cord 2005;43:704–12.
Madrone, University of Washington School of Medicine Jensen MP, Karoly P. Self-report scales and procedures for assessing
Department of Rehabilitation Medicine, in data collec- pain in adults. In: Turk DC, Melzack R, editors. Handbook of pain
tion and database management. assessment. 2nd ed. New York: The Guilford Press; 2001. p.
15–34.
Jensen MP, Karoly P, Huger R. The development and preliminary
validation of an instrument to assess patients’ attitudes toward
References pain. J Psychosom Res 1987;31:393–400.
Jensen MP, Turner JA, Romano JM. Chronic pain coping measures:
Beiske AG, Pedersen ED, Czujko B, Myhr KM. Pain and sensory individual vs. composite scores. Pain 1992;51:273–80.
complaints in multiple sclerosis. Eur J Neurol 2004;11:479–82. Jensen MP, Turner JA, Romano JM. Correlates of improvement in
Bombardier CH, Blake KD, Ehde DM, Gibbons LE, Moore D, Kraft multidisciplinary treatment of chronic pain. J Consult Clin Psychol
GH. Alcohol and drug abuse among persons with multiple 1994a;62:172–9.
sclerosis. Mult Scler 2004;10:35–40. Jensen MP, Turner JA, Romano JM. Changes in beliefs, catastro-
Boothby JL, Thorn BE, Stroud MW, Jensen MP. Coping with pain. In: phizing, and coping are associated with improvement in multidis-
Gatchel RJ, Turk DC, editors. Psychosocial factors in pain: critical ciplinary pain treatment. J Consult Clin Psychol 2001;69:655–62.
perspectives. New York: The Guilford Press; 1999. p. 343–59. Jensen MP, Turner JA, Romano JM, Karoly P. Coping with chronic
Bowen J, Gibbons L, Gianas A, Kraft GH. Self-administered Expanded pain: a critical review of the literature. Pain 1991;47:249–83.
Disability Status Scale with functional system scores correlates well Jensen MP, Turner JA, Romano JM, Lawler BK. Relationship of
with a physician-administered test. Mult Scler 2001;7:201–6. pain-specific beliefs to chronic pain adjustment. Pain
Chwastiak L, Ehde DM, Gibbons LE, Sullivan M, Bowen JD, Kraft 1994b;57:301–9.
GH. Depressive symptoms and severity of illness in multiple Jensen MP, Turner JA, Romano JM, Strom SE. The Chronic Pain
sclerosis: epidemiologic study of a large community sample. Am J Coping Inventory: development and preliminary validation. Pain
Psychiatry 2002;159:1862–8. 1995;60:203–16.
Cleeland CS, Ryan KM. Pain assessment: global use of the Brief Pain Keefe FJ, Brown GK, Wallston KA, Caldwell DS. Coping with
Inventory. Ann Acad Med 1994;23:129–38. rheumatoid arthritis pain: catastrophizing as a maladaptive strat-
Dahlem NW, Zimet GD, Walker RR. The Multidimensional Scale of egy. Pain 1989;37:51–6.
Perceived Social Support: a confirmation study. J Clin Psychol Keefe FJ, Williams DA. A comparison of coping strategies in chronic
1991;47:756–61. pain patients in different age groups. J Gerontol 1990;45:161–5.
Daut RL, Cleeland CS, Flanery RC. Development of the Wisconsin Keller S, Bann CM, Dodd SL, Schein J, Mendoza TR, Cleeland CS.
Brief Pain Questionnaire to assess pain in cancer and other Validity of the brief pain inventory for use in documenting the
diseases. Pain 1983;17:197–210. outcomes of patients with noncancer pain. Clin J Pain
Ehde DM, Gibbons LE, Chwastiak L, Bombardier CH, Sullivan MD, 2004;20:309–18.
Kraft GH. Chronic pain in a large community sample of persons Kerns R. Psychosocial aspects of pain. Int J MS Care 2000;2:35–8.
with multiple sclerosis. Mult Scler 2003;9:605–11. Kerns RD, Kassirer M, Otis J. Pain in multiple sclerosis: a biopsy-
Ehde DM, Jensen MP. Feasibility of a cognitive-restructuring inter- chosocial perspective. J Rehabil Res Dev 2002;39:225–32.
vention for treatment of chronic pain in persons with disabilities. Kraft GH, Cui JY. Multiple sclerosis. In: DeLisa JA, editor. Physical
Rehabil Psychiatry 2004;49:254–8. medicine and rehabilitation: principles and practice. New
Ehde DM, Osborne TL, Hanley MA, Jensen MP, Kraft GH. The York: Lippincott, Williams, & Wilkins; 2005. p. 1753–69.
scope and nature of pain in persons with multiple sclerosis. Mult Kurtzke JF. Rating neurologic impairment in multiple sclerosis: an
Scler, in press. expanded disability status scale (EDSS). Neurology
Ehde DM, Osborne TL, Jensen MP. Chronic pain in persons with 1983;33:1444–52.
multiple sclerosis. Phys Med Rehabil Clin N Am 2005;16:503–12. Lublin FD, Reingold SC. Defining the clinical course of multiple
Engel JM, Schwartz L, Jensen MP, Johnson DR. Pain in cerebral palsy: sclerosis: results of an international survey. National Multiple
the relation of coping strategies to adjustment. Pain 2000;88:225–30. Sclerosis Society (USA) Advisory Committee on Clinical Trials of
Evers AW, Kraaimaat FW, Geenen R, Jacobs JW, Bijlsma JW. Pain New Agents in Multiple Sclerosis. Neurology 1996;46:907–11.
coping and social support as predictors of long-term functional Moulin DE. Pain in multiple sclerosis. Neurol Clin 1989;7:321–31.
disability and pain in early rheumatoid arthritis. Behav Res Ther Nielson WR, Jensen MP, Hill ML. An activity pacing subscale for the
2003;41:1295–310. chronic pain coping inventory: development in a sample of patients
Feldman SI, Downey G, Schaffer-Neitz R. Pain, negative mood, and with fibromyalgia syndrome. Pain 2001;89:111–5.
perceived support in chronic pain patients: a daily diary study of Osborne TL, Raichle KA, Jensen MP, Ehde DM, Kraft GH. The
people with reflex sympathetic dystrophy syndrome. J Consult Clin reliability and validity of pain interference measures in persons
Psychol 1999;67:776–85. with multiple sclerosis. J Pain Symptom Manage, in press.
62 T.L. Osborne et al. / Pain 127 (2007) 52–62

Osborne TL, Turner AP, Williams RM, Bowen JD, Hatzakis M, Tan G, Jensen MP, Thornby JI, Shanti BF. Validation of the Brief
Rodriquez A, et al. Correlates of pain interference in multiple Pain Inventory for chronic nonmalignant pain. J Pain
sclerosis. Rehabil Psychol 2006;51:166–74. 2004;5:133–7.
Rosenstiel AK, Keefe FJ. The use of coping strategies in chronic low Turner JA, Jensen MP, Romano JM. Do beliefs, coping, and
back pain patients: relationship to patient characteristics and catastrophizing independently predict functioning in patients with
current adjustment. Pain 1983;17:33–44. chronic pain? Pain 2000;85:115–25.
Solaro C, Brichetto G, Amato MP, Cocco E, Colombo B, D’Aleo G, Turner JA, Jensen MP, Warms CA, Cardenas DD. Catastrophizing is
et al. The prevalence of pain in multiple sclerosis: a multicenter associated with pain intensity, psychological distress, and pain-
cross-sectional study. Neurology 2004;63:919–21. related disability among individuals with chronic pain after spinal
Stenager E, Knudsen L, Jensen K. Acute and chronic pain syndromes cord injury. Pain 2002;98:127–34.
in multiple sclerosis. Acta Neurol Scand 1991;84:197–200. Tyler EJ, Jensen MP, Engel JM, Schwartz L. The reliability and
Strong J, Ashton R, Chant D. The measurement of attitudes towards validity of pain interference measures in persons with cerebral
and beliefs about pain. Pain 1992;48:227–36. palsy. Arch Phys Med Rehabil 2002;83:236–9.
Sullivan MJ, D’Eon JL. Relation between catastrophizing and depres- Ware J, Snow K, Kosinski M, Gandek B. SF-36 Health Survey:
sion in chronic pain patients. J Abnorm Psychol 1990;99:260–3. manual and interpretation guide. Lincoln (RI): QualityMetric
Sullivan MJ, Thorn B, Haythornthwaite JA, Keefe F, Martin M, Incorporated, 2000.
Bradley LA, et al. Theoretical perspectives on the relation between Whyte A, Carroll LJ. The relationship between catastrophizing and
catastrophizing and pain. Clin J Pain 2001;17:52–64. disability in amputees experiencing phantom pain. Disabil Rehabil
Svendsen KB, Jensen TS, Overvad K, Hansen HJ, Koch-Henriksen N, 2004;26:649–54.
Bach FW. Pain in patients with multiple sclerosis: a population- Zaza C, Baine N. Cancer pain and psychosocial factors: a critical
based study. Arch Neurol 2003;60:1089–94. review of the literature. J Pain Symptom Manage
Tabachnick BG, Fidell LS. Using multivariate statistics. Cam- 2002;24:526–42.
bridge: Harper & Row; 1996. Zimet GD, Dahlem NW, Zimet SG, Farley GK. The Multidimen-
Tan G, Jensen MP, Robinson-Whelen S, Thornby JI, Monga TN. Coping sional Scale of Perceived Social Support. J Pers Assess
with chronic pain: a comparison of two measures. Pain 2001;90:127–33. 1988;52:30–41.