Original article 125
Measuring avoidance of pain: validation of the Acceptance and Action Questionnaire II-pain version Michiel F. Renemana, Marco Kleenb, Hester R. Trompetterc, Henrica R. Schiphorst Preupera, Albe`re Ko¨kee, Bianca van Baalenf and Karlein M.G. Schreursc,d Psychometric research on widely used questionnaires aimed at measuring experiential avoidance of chronic pain has led to inconclusive results. To test the structural validity, internal consistency, and construct validity of a recently developed short questionnaire: the Acceptance and Action Questionnaire II-pain version (AAQ-II-P). Cross-sectional validation study among 388 adult patients with chronic nonspecific musculoskeletal pain admitted for multidisciplinary pain rehabilitation in four tertiary rehabilitation centers in the Netherlands. Cronbach’s a was calculated to analyze internal consistency. Principal component analysis was performed to analyze factor structure. Construct validity was analyzed by examining the association between acceptance of pain and measures of psychological flexibility (two scales and sum), pain catastrophizing (three scales and sum), and mental and physical functioning. Interpretation was based on a-priori defined hypotheses. The compound of the seven items of the AAQ-II-P shows a Cronbach’s a of 0.87. The single component explained 56.2% of the total variance. Correlations ranged from r = – 0.21 to 0.73. Two of the
Introduction A growing body of evidence shows that experiential avoidance of pain plays an important role in the decreased functional status of people with chronic pain. Experiential avoidance of pain is defined as any overt or covert behavior aimed at diminishing or controlling pain in all of its facets, including physical experiences of pain and pain-related thoughts, emotions or cognitions. Pain can be avoided experientially by carrying out certain activities less often or carrying out other activities more often. Short-term consequences of experiential avoidance may be immediate relief and positive attention from the social environment. These consequences serve as reinforcers for future experiential avoidance. However, experiential avoidance may also lead to restriction of normal behavioral repertoire and to rigid maladaptive behaviors such as physical inactivity, social isolation, work absence, or dependence on pain medication. In the long run, the choice for experiential avoidant strategies to deal with negative experiences blocks the way to a vital and healthy life. In the majority of studies investigating the effects of experiential avoidance in chronic pain the Chronic Pain c 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins 0342-5282
predefined hypotheses were rejected and seven were not rejected. The AAQ-II-P measures a single component and has good internal consistency, and construct validity is not rejected. Thus, the construct validity of the AAQ-II-P sum scores as indicator of experiential avoidance of pain was supported. International Journal of Rehabilitation c 2014 Wolters Kluwer Health | Research 37:125–129 Lippincott Williams & Wilkins. International Journal of Rehabilitation Research 2014, 37:125–129 Keywords: acceptance, Acceptance and Action Questionnaire, acceptance and commitment therapy, experiential avoidance, pain a Department of Rehabilitation Medicine, Center for Rehabilitation, University Medical Center Groningen, bOCRN Child and Youth Psychiatry, Groningen, c Faculty of Behavioural Sciences, University of Twente, dRoessingh Research and Development, Enschede, eAdelante Rehabilitation Center, Hoensbroek and f Rijndam Rehabilitation Center, Rotterdam, The Netherlands
Correspondence to Michiel F. Reneman, PhD, Department of Rehabilitation Medicine, Center for Rehabilitation, University Medical Center Groningen, PO Box 30.002, 9750 RA Haren, Groningen, The Netherlands Tel: + 31 50 5338550; fax: + 31 50 5338570; e-mail: [email protected] Received 5 July 2013 Accepted 12 November 2013
Acceptance Questionnaire (CPAQ) is used (Reneman et al., 2010). Research on the psychometric properties of the CPAQ draws inconsistent conclusions. Although earlier studies report adequate reliability and validity (McCracken, 1999, 2004; Vowles et al., 2008), recent studies report psychometric inconsistencies. In studies on the factorial validity no evidence for sufficient goodness-of-fit for the two-factor solution of the CPAQ was observed (Nicholas and Asghari, 2006; Bernini et al., 2010; Fish et al., 2010; Trompetter et al., 2011). In addition, weak predictive validity for physical disability as well as severity of depression has been reported (Nicholas and Asghari, 2006). These inconsistent findings, and concerns about its length, motivated researchers to develop a pain version of the most widely used experiential avoidance questionnaire, the Acceptance and Action Questionnaire II (AAQ-II; Bond et al., 2011). The aim of the current study was to analyze the internal consistency and construct validity of the AAQ-II-P in a clinical context. We hypothesized that: (1) the items of the AAQ-II-P represent a single internally consistent component; DOI: 10.1097/MRR.0000000000000044
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
126 International Journal of Rehabilitation Research 2014, Vol 37 No 2
(2) the AAQ-II-P sum scores correlate weakly or moderately positively with questionnaires measuring other positively measured concepts related to experiential avoidance: cognitive fusion, avoidance of activity, and pain catastrophizing; (3) the AAQ-II-P sum scores correlate moderately or strongly negatively with positively measured outcome measures: functional mental or physical status and perceived mental health. A moderate-to-strong relation was hypothesized, because acceptance is not completely dependent from physical and mental functioning, yet also not independent, as it may be challenging to accept pain while experiencing high amounts of disfunction. Moreover, they all represent measures at the level of ‘functioning’.
Methods Patients
The sample consisted of 388 adults who were referred to one of four Dutch rehabilitation centers specializing in treatments for chronic pain.
‘Pain gets in the way of my success’. Results indicate that the AAQ-II-P sum scores of respondents recruited from the general, mainly pain-free, population are internally consistent and valid. The total score of the AAQ-II-P can range from 0 to 42, with higher scores representing higher self-reported experiential avoidance of pain. The internal consistency and validity of the AAQ-II-P were established in a community sample. Psychological inflexibility of pain
The Dutch version of the Psychological Inflexibility in Pain Scale (PIPS; Wicksell et al., 2008) was used to measure psychological inflexibility of pain. The PIPS is a 12-item instrument reflecting two aspects of psychological inflexibility of pain: avoidance (eight items) and cognitive fusion (four items). Higher scores indicate more psychological inflexibility. Participants rate the items on a seven-point Likert-type scale ranging from ‘never true’ (1) to ‘always true’ (7). In a recent study its psychometric qualities were confirmed (Wicksell et al., 2010). Pain catastrophizing
Procedure Design
A cross-sectional validation study with data from care as usual during the admission phase of multidisciplinary pain rehabilitation. In an admission procedure the patient signed an informed consent form in which the goals and procedures regarding confidentiality of the personal information were provided. Patients filled out the questionnaires either at the rehabilitation center or at home; in the latter case a prepaid return envelope was included with the questionnaire. One designated person at each rehabilitation center organized the logistics surrounding the questionnaires. Finally, anonymous datasets of the rehabilitation centers were merged in the final database by the research group. Because data were derived from care as usual and patients agreed to their data being used for research, formal ethical approval was not needed. Due to practical reasons not all measures were available for all four rehabilitation centers. The four participating rehabilitation centers are located throughout the Netherlands. Questionnaires Acceptance of pain
The AAQ-II-P (Appendix) is a seven-item questionnaire, aimed at measuring experiential avoidance of pain. Respondents are asked to rate their answers on a sixpoint Likert scale ranging from 0 (never true) to 6 (always true). A lower score on the AAQ-II-P represents higher acceptance of pain and a higher score represents higher self-reported experiential avoidance of pain. The AAQ-II-P was developed by modifying the content of items from the general AAQ-II, for instance by changing the wording ‘Worries get in the way of my success’ to
The Pain Catastrophizing Scale (PCS; Sullivan et al., 1995) is a 13-item questionnaire that can be scored on a five-point Likert scale, ranging from 0 (never) to 4 (always). The PCS measures the extent to which people overreact to experiencing pain in a sum score and three subscales: rumination, magnifying pain, and feelings of helplessness resulting from pain. In the current study, the validated Dutch version of the PCS was used (Van Damme et al., 2002). Physical functioning and mental health
The RAND-36 Health Survey is a generic measure of health status (Ware and Sherbourne, 1992). The survey consists of 36 questions organized in eight scales. In the current study only the subscales ‘physical functioning’ (PF) and ‘mental health’ (MH) were used. The scale PF considers limitations in performing daily activities, such as climbing stairs, carrying groceries, bathing, or dressing. The scale MH considers general feelings of MH over the past 4 weeks. Both raw scale scores are linearly converted to a 0–100 scale, with higher scores indicating higher levels of physical functioning and mental well-being. The subscales of the Dutch version of the RAND-36 have been shown to have good psychometric qualities (Aaronson et al., 1998). Statistical analysis
To test the hypothesized single factor underlying the items of the AAQ-II-P, a principal component analysis was performed. Internal consistency was analyzed by calculating Cronbach’s a, using a value of more than 0.70 to be considered acceptable. Construct validity was analyzed by examining the association between acceptance of pain and the other variables, as indicated by Pearson product
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Measuring avoidance of pain Reneman et al.
moment correlation coefficients. Correlations over 0.75 were considered strong, between 0.50 and 0.75 moderate, between 0.25 and 0.50 weak, and below 0.25 slight or nonexistant. Significance of differences of AAQ-II-P scores between demographic groups (sex, age, work status, education) was tested using independent groups t-tests and one-way analyses of variance. Cases with missing values were deleted from the analyses pairwise.
Results Patients
The patients’ mean age was 43.6 years (SD 12.5; range 17–75) and 73% were female. The majority of patients reported pain in the neck and chronic low back pain, and other less frequently reported complaints were whiplash-associated disorder and chronic fatigue. Mean pain intensity was 64 (SD = 19, VAS 0–100, higher scores indicate more pain). Of the patients, 52% were working either full time or part time, 52% were in an intimate relationship, and 18% had higher education. Clinical characteristics as measured with questionnaires are presented in Table 1.
127
Construct validity
The frequency distribution of AAQ-II-P sum scores follows a normal curve (Fig. 2). Floor or ceiling effects were not apparent. Males scored significantly higher than females [two tailed, t = 2.1; d.f. = 328; P < 0.05; males mean 29.6 (SD 8.6), females mean 27.3 (SD 9.0)]. However, because this difference was less than half a SD of the total AAQ-II-P sum scores, further separate analyses for men and women were not performed. Relations between AAQ-II-P sum scores and age, education, and work status were nonsignificant. Correlations between AAQ-II-P sum scores and theoretically related variables are shown in Table 3. Two of the predefined hypotheses were rejected and seven were not rejected, indicating adequate support for the construct validity of the AAQ-II-P sum scores as indicator of experiential avoidance of pain.
Discussion The results of this study indicate that the Dutchlanguage version of the AAQ-II-P measures a single component and has good internal consistency, and
Item analysis
There were no AAQ-II-P items with an item-total correlation below 0.3. Visual inspection of the frequency distributions of the seven items showed no extreme skewed responses or low variability.
Fig. 1
4
Component structure and internal consistency
Table 1
3 Eigen value
After justification by Kaiser–Meyer–Olkin measure of sampling adequacy (KMO = 0.86) and Bartlett’s test of sphericity (P < 0.001) a principal component analysis was carried out. Both the scree plot and Kaiser’s eigenvalue of more than 1 criterion indicated a one-component solution (Fig. 1). The single component explained 56.2% of the total variance. All component loadings of the items of the AAQ-II-P were above the minimum of 0.4 (Table 2). The compound of the seven items of the AAQ-II-P show a Cronbach’s a of 0.87. Frequency distribution of the total scores does not indicate floor or ceiling effects (Fig. 2). No items were deleted from the analysis.
2
1
0 1
2
3 4 5 Component number
6
7
Screeplot with Eigenvalues for each component.
Clinical characteristics of the patient sample (n = 388)
AAQ-II-P sum PIPS sum PIPS fusion PIPS avoidance PCS sum PCS ruminating PCS magnification PCS helplessness RAND-36 MH RAND-36 PF
n
Score range
Mean
SD
388 388 388 388 159 90 90 90 279 282
0–42 0–84 0–28 0–56 0–52
28 4.4 5.4 3.9 27.6 8.3 4.0 11.1 62.2 46.1
8.9 1.0 1.1 1.2 12.5 4.2 3.7 6.2 16.7 21.6
0–100 0–100
AAQ-II-P, Acceptance and Action Questionnaire II-pain version; MH, mental health; PCS, Pain Catastrophizing Scale; PF, physical functioning; PIPS, Psychological Inflexibility in Pain Scale.
Table 2 Items of the Acceptance and Action Questionnaire II-pain version and their item-factor loadings
Item 1. My pain makes it difficult for me to live a life that I would value 2. I’m afraid of my pain 3. I worry about not being able to control my pain 4. My pain prevents me from having a fulfilling life 5. Pain causes problems in my life 6. It seems like most people are handling their lives better than I am 7. Pain gets in the way of my success
Item-factor loadings 0.79 0.65 0.79 0.79 0.70 0.74 0.79
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
128 International Journal of Rehabilitation Research 2014, Vol 37 No 2
Fig. 2
40
Frequency
30
20
10
0 0.00
10.00
20.00
30.00
40.00
50.00
Frequency distribution of Acceptance and Action Questionnaire II-pain version sum scores.
Table 3 Correlations between Acceptance and Action Questionnaire II-pain version sum scores and other (subscale) scores
PIPS sum PIPS fusion PIPS avoidance PCS sum PCS ruminating PCS magnification PCS helplessness RAND-36 MH RAND-36 PF
n
r
Hypothesis
Interpretation
388 388 388 159 90 90 90 275 272
0.73 0.47 0.71 0.59 0.49 0.63 0.65 – 0.45 – 0.21
Weak–moderate Weak–moderate Weak–moderate Weak–moderate Weak–moderate Weak–moderate Weak–moderate Moderate–strong Moderate–strong
Not rejected Not rejected Not rejected Not rejected Not rejected Not rejected Not rejected Rejected Rejected
MH, mental health; n, number of participants in calculation; PCS, Pain Catastrophizing Scale; PF, physical functioning; PIPS, Psychological Inflexibility in Pain Scale; r, Pearson correlation coefficient. All correlations: P < 0.01.
construct validity is not rejected, because all nine correlations with other constructs appear in the expected direction and seven of nine in the expected magnitude. A closer look at the correlations between AAQ-II-P and the PIPS shows that the correlations were very high. Apparently, there is substantial shared variance between experiential pain avoidance (AAQ-II-P) and avoidance (PIPS) of pain. The correlation of AAQ-II-P with RAND36 PF was slight or nonexistent, which was lower than hypothesized. There may be several explanations for this. This may be a correct finding, indicating that this relation is indeed slight or nonexistent. If this observation is replicated in a different setting, this may indicate that the hypothesis is incorrect. Alternatively, the RAND-36 PF did not measure functional activities that were avoided by the participants of this study.
The outcomes of this study confirm that the AAQ-II-P is a useful alternative to the CPAQ (McCracken et al., 2004) in measuring components of experiential avoidance of chronic pain. As described previously, several statistical and theoretical problems have been reported with the CPAQ, most importantly with regard to the fit of a twofactor structure and the specific theoretical content of the pain willingness subscale (Nicholas and Asghari, 2006; Bernini et al., 2010; Fish et al., 2010; Trompetter et al., 2011). The AAQ-II-P circumvents these issues as experiential avoidance is measured as a generic construct with one scale, without splitting this construct into several subconstructs. Additionally, the straightforward but successful modifications applied in this study to adjust the original and widely used AAQ-II (Bond et al., 2011; Fledderus et al., 2012) for pain populations offer opportunities for modifying the AAQ-II toward other clinical populations. Applying modified versions of the AAQ-II in different populations will stimulate more straightforward comparisons of research findings. Furthermore, this will ensure a homogeneous operationalization of experiential avoidance in different research and clinical settings. This study may help the scientific and clinical fields of chronic pain to measure a relatively new construct. Now that these first steps have been taken, the AAQ-II-P should be studied further to analyze its reliability and predictive validity, and to develop clinically meaningful reference values. One should note that the AAQ-II-P does not explicitly measure experiential avoidance of chronic pain, as none of the items explicitly focus on pain duration. For this reason the name AAQ-II-P was preferred over CPAQ-II. Future research could clarify the relation between pain duration and experiential avoidance of pain. The psychometric results of the current study are similar to those observed in a recent study on the psychometric properties of the AAQ-II-P in a community sample. In both samples one single component underlies the scores on seven internally consistent items, which justifies comparison of AAQ-II-P sum scores between both samples. Although in the community sample pain did occur, compared with the community sample, the AAQII-P sum scores of the clinical sample were notably higher. Perhaps not surprisingly, experiential avoidance in people suffering from chronic pain seems to be much more common than experiential avoidance of pain in people in the general community. Further research could focus on the differences between people suffering from chronic pain who do or do not maintain their daily functioning and its relationship with experiential avoidance of pain. In addition, the difference between the scores of the clinical and nonclinical sample may be indicative for the discriminative power of the AAQ-II-P as clinical instrument. An interesting subject of investiga-
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Measuring avoidance of pain Reneman et al. 129
tion may be the significant difference in AAQ-II-P sum scores between men and women in the current clinical sample. This indicates that while being in a position to seek medical help for chronic pain, women report being better able to accept their pain than men. Future longitudinal studies in which acceptance of pain is investigated as a process variable may shed light on the predictive value of sex differences in amount of selfreported experiential avoidance of pain. The current study was performed within a fairly large sample of people referred to pain rehabilitation programs for chronic pain, and this may be interpreted as an advantage for the generalizability of its results to other tertiary care clinical populations. A limitation of the study is that it is only performed within a Dutch sample. Although not all measures were available for all four rehabilitation centers, the sample is still sufficiently powered for its purpose. Further research may be needed to investigate the psychometric properties of the AAQ-II-P in non-Dutch samples, which may also be used to replicate the results of this study and to further investigate and develop the psychometric properties of the AAQ-II-P. Based on the results of the current study we conclude that the Dutchlanguage version of the AAQ-II-P can be used as an internally consistent and valid instrument to measure experiential avoidance of pain.
Acknowledgements All authors discussed the results and commented on the manuscript; contributed substantially to conception and design, or acquisition of data, or analysis and interpretation of data; assisted in drafting the article or revising it critically for important intellectual content; and gave final approval of the version to be published.
Conflicts of interest
There are no conflicts of interest.
References Aaronson NK, Muller M, Cohen PD, Essink-Bot M, Fekkes M, Sanderman R, et al. (1998). Translation, validation, and norming of the Dutch language version of the SF-36 health survey in community and chronic disease populations. J Clin Epidemiol 51:1055–1068. Bernini O, Pennato T, Cosci F, Berrocal C (2010). The psychometric properties of the chronic pain acceptance questionnaire in Italian patients with chronic pain. J Health Psychol 15:1236–1245. Bond FW, Hayes SC, Baer RA, Carpenter KM, Guenole N, Orcutt HK, et al. (2011). Preliminary psychometric properties of the Acceptance and Action Questionnaire-II: a revised measure of psychological inflexibility and experiential avoidance. Behav Ther 42:676–688. Fish RA, McGuire B, Hogan M, Morrison TG, Stewart I (2010). Validation of the chronic pain acceptance questionnaire (CPAQ) in an internet sample and development and preliminary validation of the CPAQ-8. Pain 149: 435–443. Fledderus M, Oude Voshaar MA, Ten Klooster PM, Bohlmeijer ET (2012). Further evaluation of the psychometric properties of the Acceptance and Action Questionnaire-II. Psychol Assess 24:925–936. McCracken LN (1999). Behavioral constituents of chronic pain acceptance: results from factor analysis of the Chronic Pain Acceptance Questionnaire. J Back Musculoskeletal Rehabil 13:93–100. McCracken LM, Vowles KE, Eccleston C (2004). Acceptance of chronic pain: component analysis and a revised assessment method. Pain 107: 159–166. Nicholas MK, Asghari A (2006). Investigating acceptance in adjustment to chronic pain: is acceptance broader than we thought? Pain 124:269–279. Reneman MF, Dijkstra A, Geertzen JH, Dijkstra PU (2010). Psychometric properties of chronic pain acceptance questionnaires: a systematic review. Eur J Pain 14:457–465. Sullivan MJL, Bishop SR, Pivik J (1995). The pain catastrophizing scale: development and validation. Psychol Assess 7:524–532. Trompetter H, Klooster PM ten, Ko¨ke A, Schreurs KMG (2011). Acceptance of pain; problems with the factorial validity of the Dutch translation of the chronic pain acceptance questionnaire (CPAQ) [in Dutch]. Psychologie Gezondheid 39:292–299. Van Damme S, Crombez G, Bijttebier P, Goubert L, Van Houdenhove B (2002). A confirmatory factor analysis of the pain catastrophizing scale: invariant factor structure across clinical and non-clinical populations. Pain 96:319–324. Vowles KE, McCracken LM, McLeod C, Eccleston C (2008). The chronic pain acceptance questionnaire: confirmatory factor analysis and identification of patient subgroups. Pain 140:284–291. Ware JEJ, Sherbourne CD (1992). The MOS 36-item short-form health survey (SF-36). I. conceptual framework and item selection. Med Care 30:473–483. Wicksell RK, Renofalt J, Olsson GL, Bond FW, Melin L (2008). Avoidance and cognitive fusion – central components in pain related disability? Development and preliminary validation of the psychological inflexibility in pain scale (PIPS). Eur J Pain 12:491–500. Wicksell RK, Lekander M, Sorjonen K, Olsson GL (2010). The psychological inflexibility in pain scale (PIPS). Statistical properties and model fit of an instrument to assess change processes in pain related disability. Eur J Pain 14:771.e1–771.e14.
Appendix Below you will find a list of statements. Please rate how true each statement is for you by circling a number next to it. Use the scale below to make your choice. Table A1 1 Never true
Acceptance and Action Questionnaire II-pain version 2
3
4
5
Very seldom true
Seldom true
Sometimes true
Frequently true
1. My pain makes it difficult for me to live a life that I would value 2. I’m afraid of my pain 3. I worry about not being able to control my pain 4. My pain prevents me from having a fulfilling life 5. Pain causes problems in my life 6. It seems like most people are handling their lives better than I am 7. Pain gets in the way of my success
6
7
Almost always true 1 1 1 1 1 1 1
2 2 2 2 2 2 2
Always true 3 3 3 3 3 3 3
4 4 4 4 4 4 4
5 5 5 5 5 5 5
6 6 6 6 6 6 6
7 7 7 7 7 7 7
This is a one-factor measure of experiential avoidance of pain. Score the scale by summing the seven items. Higher scores equal greater levels of experiential avoidance of pain.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Measuring avoidance of pain: validation of the Acceptance and Action Questionnaire II-pain version Michiel F. Renemana, Marco Kleenb, Hester R. Trompetterc, Henrica R. Schiphorst Preupera, Albe`re Ko¨kee, Bianca van Baalenf and Karlein M.G. Schreursc,d Psychometric research on widely used questionnaires aimed at measuring experiential avoidance of chronic pain has led to inconclusive results. To test the structural validity, internal consistency, and construct validity of a recently developed short questionnaire: the Acceptance and Action Questionnaire II-pain version (AAQ-II-P). Cross-sectional validation study among 388 adult patients with chronic nonspecific musculoskeletal pain admitted for multidisciplinary pain rehabilitation in four tertiary rehabilitation centers in the Netherlands. Cronbach’s a was calculated to analyze internal consistency. Principal component analysis was performed to analyze factor structure. Construct validity was analyzed by examining the association between acceptance of pain and measures of psychological flexibility (two scales and sum), pain catastrophizing (three scales and sum), and mental and physical functioning. Interpretation was based on a-priori defined hypotheses. The compound of the seven items of the AAQ-II-P shows a Cronbach’s a of 0.87. The single component explained 56.2% of the total variance. Correlations ranged from r = – 0.21 to 0.73. Two of the
Introduction A growing body of evidence shows that experiential avoidance of pain plays an important role in the decreased functional status of people with chronic pain. Experiential avoidance of pain is defined as any overt or covert behavior aimed at diminishing or controlling pain in all of its facets, including physical experiences of pain and pain-related thoughts, emotions or cognitions. Pain can be avoided experientially by carrying out certain activities less often or carrying out other activities more often. Short-term consequences of experiential avoidance may be immediate relief and positive attention from the social environment. These consequences serve as reinforcers for future experiential avoidance. However, experiential avoidance may also lead to restriction of normal behavioral repertoire and to rigid maladaptive behaviors such as physical inactivity, social isolation, work absence, or dependence on pain medication. In the long run, the choice for experiential avoidant strategies to deal with negative experiences blocks the way to a vital and healthy life. In the majority of studies investigating the effects of experiential avoidance in chronic pain the Chronic Pain c 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins 0342-5282
predefined hypotheses were rejected and seven were not rejected. The AAQ-II-P measures a single component and has good internal consistency, and construct validity is not rejected. Thus, the construct validity of the AAQ-II-P sum scores as indicator of experiential avoidance of pain was supported. International Journal of Rehabilitation c 2014 Wolters Kluwer Health | Research 37:125–129 Lippincott Williams & Wilkins. International Journal of Rehabilitation Research 2014, 37:125–129 Keywords: acceptance, Acceptance and Action Questionnaire, acceptance and commitment therapy, experiential avoidance, pain a Department of Rehabilitation Medicine, Center for Rehabilitation, University Medical Center Groningen, bOCRN Child and Youth Psychiatry, Groningen, c Faculty of Behavioural Sciences, University of Twente, dRoessingh Research and Development, Enschede, eAdelante Rehabilitation Center, Hoensbroek and f Rijndam Rehabilitation Center, Rotterdam, The Netherlands
Correspondence to Michiel F. Reneman, PhD, Department of Rehabilitation Medicine, Center for Rehabilitation, University Medical Center Groningen, PO Box 30.002, 9750 RA Haren, Groningen, The Netherlands Tel: + 31 50 5338550; fax: + 31 50 5338570; e-mail: [email protected] Received 5 July 2013 Accepted 12 November 2013
Acceptance Questionnaire (CPAQ) is used (Reneman et al., 2010). Research on the psychometric properties of the CPAQ draws inconsistent conclusions. Although earlier studies report adequate reliability and validity (McCracken, 1999, 2004; Vowles et al., 2008), recent studies report psychometric inconsistencies. In studies on the factorial validity no evidence for sufficient goodness-of-fit for the two-factor solution of the CPAQ was observed (Nicholas and Asghari, 2006; Bernini et al., 2010; Fish et al., 2010; Trompetter et al., 2011). In addition, weak predictive validity for physical disability as well as severity of depression has been reported (Nicholas and Asghari, 2006). These inconsistent findings, and concerns about its length, motivated researchers to develop a pain version of the most widely used experiential avoidance questionnaire, the Acceptance and Action Questionnaire II (AAQ-II; Bond et al., 2011). The aim of the current study was to analyze the internal consistency and construct validity of the AAQ-II-P in a clinical context. We hypothesized that: (1) the items of the AAQ-II-P represent a single internally consistent component; DOI: 10.1097/MRR.0000000000000044
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
126 International Journal of Rehabilitation Research 2014, Vol 37 No 2
(2) the AAQ-II-P sum scores correlate weakly or moderately positively with questionnaires measuring other positively measured concepts related to experiential avoidance: cognitive fusion, avoidance of activity, and pain catastrophizing; (3) the AAQ-II-P sum scores correlate moderately or strongly negatively with positively measured outcome measures: functional mental or physical status and perceived mental health. A moderate-to-strong relation was hypothesized, because acceptance is not completely dependent from physical and mental functioning, yet also not independent, as it may be challenging to accept pain while experiencing high amounts of disfunction. Moreover, they all represent measures at the level of ‘functioning’.
Methods Patients
The sample consisted of 388 adults who were referred to one of four Dutch rehabilitation centers specializing in treatments for chronic pain.
‘Pain gets in the way of my success’. Results indicate that the AAQ-II-P sum scores of respondents recruited from the general, mainly pain-free, population are internally consistent and valid. The total score of the AAQ-II-P can range from 0 to 42, with higher scores representing higher self-reported experiential avoidance of pain. The internal consistency and validity of the AAQ-II-P were established in a community sample. Psychological inflexibility of pain
The Dutch version of the Psychological Inflexibility in Pain Scale (PIPS; Wicksell et al., 2008) was used to measure psychological inflexibility of pain. The PIPS is a 12-item instrument reflecting two aspects of psychological inflexibility of pain: avoidance (eight items) and cognitive fusion (four items). Higher scores indicate more psychological inflexibility. Participants rate the items on a seven-point Likert-type scale ranging from ‘never true’ (1) to ‘always true’ (7). In a recent study its psychometric qualities were confirmed (Wicksell et al., 2010). Pain catastrophizing
Procedure Design
A cross-sectional validation study with data from care as usual during the admission phase of multidisciplinary pain rehabilitation. In an admission procedure the patient signed an informed consent form in which the goals and procedures regarding confidentiality of the personal information were provided. Patients filled out the questionnaires either at the rehabilitation center or at home; in the latter case a prepaid return envelope was included with the questionnaire. One designated person at each rehabilitation center organized the logistics surrounding the questionnaires. Finally, anonymous datasets of the rehabilitation centers were merged in the final database by the research group. Because data were derived from care as usual and patients agreed to their data being used for research, formal ethical approval was not needed. Due to practical reasons not all measures were available for all four rehabilitation centers. The four participating rehabilitation centers are located throughout the Netherlands. Questionnaires Acceptance of pain
The AAQ-II-P (Appendix) is a seven-item questionnaire, aimed at measuring experiential avoidance of pain. Respondents are asked to rate their answers on a sixpoint Likert scale ranging from 0 (never true) to 6 (always true). A lower score on the AAQ-II-P represents higher acceptance of pain and a higher score represents higher self-reported experiential avoidance of pain. The AAQ-II-P was developed by modifying the content of items from the general AAQ-II, for instance by changing the wording ‘Worries get in the way of my success’ to
The Pain Catastrophizing Scale (PCS; Sullivan et al., 1995) is a 13-item questionnaire that can be scored on a five-point Likert scale, ranging from 0 (never) to 4 (always). The PCS measures the extent to which people overreact to experiencing pain in a sum score and three subscales: rumination, magnifying pain, and feelings of helplessness resulting from pain. In the current study, the validated Dutch version of the PCS was used (Van Damme et al., 2002). Physical functioning and mental health
The RAND-36 Health Survey is a generic measure of health status (Ware and Sherbourne, 1992). The survey consists of 36 questions organized in eight scales. In the current study only the subscales ‘physical functioning’ (PF) and ‘mental health’ (MH) were used. The scale PF considers limitations in performing daily activities, such as climbing stairs, carrying groceries, bathing, or dressing. The scale MH considers general feelings of MH over the past 4 weeks. Both raw scale scores are linearly converted to a 0–100 scale, with higher scores indicating higher levels of physical functioning and mental well-being. The subscales of the Dutch version of the RAND-36 have been shown to have good psychometric qualities (Aaronson et al., 1998). Statistical analysis
To test the hypothesized single factor underlying the items of the AAQ-II-P, a principal component analysis was performed. Internal consistency was analyzed by calculating Cronbach’s a, using a value of more than 0.70 to be considered acceptable. Construct validity was analyzed by examining the association between acceptance of pain and the other variables, as indicated by Pearson product
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Measuring avoidance of pain Reneman et al.
moment correlation coefficients. Correlations over 0.75 were considered strong, between 0.50 and 0.75 moderate, between 0.25 and 0.50 weak, and below 0.25 slight or nonexistant. Significance of differences of AAQ-II-P scores between demographic groups (sex, age, work status, education) was tested using independent groups t-tests and one-way analyses of variance. Cases with missing values were deleted from the analyses pairwise.
Results Patients
The patients’ mean age was 43.6 years (SD 12.5; range 17–75) and 73% were female. The majority of patients reported pain in the neck and chronic low back pain, and other less frequently reported complaints were whiplash-associated disorder and chronic fatigue. Mean pain intensity was 64 (SD = 19, VAS 0–100, higher scores indicate more pain). Of the patients, 52% were working either full time or part time, 52% were in an intimate relationship, and 18% had higher education. Clinical characteristics as measured with questionnaires are presented in Table 1.
127
Construct validity
The frequency distribution of AAQ-II-P sum scores follows a normal curve (Fig. 2). Floor or ceiling effects were not apparent. Males scored significantly higher than females [two tailed, t = 2.1; d.f. = 328; P < 0.05; males mean 29.6 (SD 8.6), females mean 27.3 (SD 9.0)]. However, because this difference was less than half a SD of the total AAQ-II-P sum scores, further separate analyses for men and women were not performed. Relations between AAQ-II-P sum scores and age, education, and work status were nonsignificant. Correlations between AAQ-II-P sum scores and theoretically related variables are shown in Table 3. Two of the predefined hypotheses were rejected and seven were not rejected, indicating adequate support for the construct validity of the AAQ-II-P sum scores as indicator of experiential avoidance of pain.
Discussion The results of this study indicate that the Dutchlanguage version of the AAQ-II-P measures a single component and has good internal consistency, and
Item analysis
There were no AAQ-II-P items with an item-total correlation below 0.3. Visual inspection of the frequency distributions of the seven items showed no extreme skewed responses or low variability.
Fig. 1
4
Component structure and internal consistency
Table 1
3 Eigen value
After justification by Kaiser–Meyer–Olkin measure of sampling adequacy (KMO = 0.86) and Bartlett’s test of sphericity (P < 0.001) a principal component analysis was carried out. Both the scree plot and Kaiser’s eigenvalue of more than 1 criterion indicated a one-component solution (Fig. 1). The single component explained 56.2% of the total variance. All component loadings of the items of the AAQ-II-P were above the minimum of 0.4 (Table 2). The compound of the seven items of the AAQ-II-P show a Cronbach’s a of 0.87. Frequency distribution of the total scores does not indicate floor or ceiling effects (Fig. 2). No items were deleted from the analysis.
2
1
0 1
2
3 4 5 Component number
6
7
Screeplot with Eigenvalues for each component.
Clinical characteristics of the patient sample (n = 388)
AAQ-II-P sum PIPS sum PIPS fusion PIPS avoidance PCS sum PCS ruminating PCS magnification PCS helplessness RAND-36 MH RAND-36 PF
n
Score range
Mean
SD
388 388 388 388 159 90 90 90 279 282
0–42 0–84 0–28 0–56 0–52
28 4.4 5.4 3.9 27.6 8.3 4.0 11.1 62.2 46.1
8.9 1.0 1.1 1.2 12.5 4.2 3.7 6.2 16.7 21.6
0–100 0–100
AAQ-II-P, Acceptance and Action Questionnaire II-pain version; MH, mental health; PCS, Pain Catastrophizing Scale; PF, physical functioning; PIPS, Psychological Inflexibility in Pain Scale.
Table 2 Items of the Acceptance and Action Questionnaire II-pain version and their item-factor loadings
Item 1. My pain makes it difficult for me to live a life that I would value 2. I’m afraid of my pain 3. I worry about not being able to control my pain 4. My pain prevents me from having a fulfilling life 5. Pain causes problems in my life 6. It seems like most people are handling their lives better than I am 7. Pain gets in the way of my success
Item-factor loadings 0.79 0.65 0.79 0.79 0.70 0.74 0.79
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
128 International Journal of Rehabilitation Research 2014, Vol 37 No 2
Fig. 2
40
Frequency
30
20
10
0 0.00
10.00
20.00
30.00
40.00
50.00
Frequency distribution of Acceptance and Action Questionnaire II-pain version sum scores.
Table 3 Correlations between Acceptance and Action Questionnaire II-pain version sum scores and other (subscale) scores
PIPS sum PIPS fusion PIPS avoidance PCS sum PCS ruminating PCS magnification PCS helplessness RAND-36 MH RAND-36 PF
n
r
Hypothesis
Interpretation
388 388 388 159 90 90 90 275 272
0.73 0.47 0.71 0.59 0.49 0.63 0.65 – 0.45 – 0.21
Weak–moderate Weak–moderate Weak–moderate Weak–moderate Weak–moderate Weak–moderate Weak–moderate Moderate–strong Moderate–strong
Not rejected Not rejected Not rejected Not rejected Not rejected Not rejected Not rejected Rejected Rejected
MH, mental health; n, number of participants in calculation; PCS, Pain Catastrophizing Scale; PF, physical functioning; PIPS, Psychological Inflexibility in Pain Scale; r, Pearson correlation coefficient. All correlations: P < 0.01.
construct validity is not rejected, because all nine correlations with other constructs appear in the expected direction and seven of nine in the expected magnitude. A closer look at the correlations between AAQ-II-P and the PIPS shows that the correlations were very high. Apparently, there is substantial shared variance between experiential pain avoidance (AAQ-II-P) and avoidance (PIPS) of pain. The correlation of AAQ-II-P with RAND36 PF was slight or nonexistent, which was lower than hypothesized. There may be several explanations for this. This may be a correct finding, indicating that this relation is indeed slight or nonexistent. If this observation is replicated in a different setting, this may indicate that the hypothesis is incorrect. Alternatively, the RAND-36 PF did not measure functional activities that were avoided by the participants of this study.
The outcomes of this study confirm that the AAQ-II-P is a useful alternative to the CPAQ (McCracken et al., 2004) in measuring components of experiential avoidance of chronic pain. As described previously, several statistical and theoretical problems have been reported with the CPAQ, most importantly with regard to the fit of a twofactor structure and the specific theoretical content of the pain willingness subscale (Nicholas and Asghari, 2006; Bernini et al., 2010; Fish et al., 2010; Trompetter et al., 2011). The AAQ-II-P circumvents these issues as experiential avoidance is measured as a generic construct with one scale, without splitting this construct into several subconstructs. Additionally, the straightforward but successful modifications applied in this study to adjust the original and widely used AAQ-II (Bond et al., 2011; Fledderus et al., 2012) for pain populations offer opportunities for modifying the AAQ-II toward other clinical populations. Applying modified versions of the AAQ-II in different populations will stimulate more straightforward comparisons of research findings. Furthermore, this will ensure a homogeneous operationalization of experiential avoidance in different research and clinical settings. This study may help the scientific and clinical fields of chronic pain to measure a relatively new construct. Now that these first steps have been taken, the AAQ-II-P should be studied further to analyze its reliability and predictive validity, and to develop clinically meaningful reference values. One should note that the AAQ-II-P does not explicitly measure experiential avoidance of chronic pain, as none of the items explicitly focus on pain duration. For this reason the name AAQ-II-P was preferred over CPAQ-II. Future research could clarify the relation between pain duration and experiential avoidance of pain. The psychometric results of the current study are similar to those observed in a recent study on the psychometric properties of the AAQ-II-P in a community sample. In both samples one single component underlies the scores on seven internally consistent items, which justifies comparison of AAQ-II-P sum scores between both samples. Although in the community sample pain did occur, compared with the community sample, the AAQII-P sum scores of the clinical sample were notably higher. Perhaps not surprisingly, experiential avoidance in people suffering from chronic pain seems to be much more common than experiential avoidance of pain in people in the general community. Further research could focus on the differences between people suffering from chronic pain who do or do not maintain their daily functioning and its relationship with experiential avoidance of pain. In addition, the difference between the scores of the clinical and nonclinical sample may be indicative for the discriminative power of the AAQ-II-P as clinical instrument. An interesting subject of investiga-
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Measuring avoidance of pain Reneman et al. 129
tion may be the significant difference in AAQ-II-P sum scores between men and women in the current clinical sample. This indicates that while being in a position to seek medical help for chronic pain, women report being better able to accept their pain than men. Future longitudinal studies in which acceptance of pain is investigated as a process variable may shed light on the predictive value of sex differences in amount of selfreported experiential avoidance of pain. The current study was performed within a fairly large sample of people referred to pain rehabilitation programs for chronic pain, and this may be interpreted as an advantage for the generalizability of its results to other tertiary care clinical populations. A limitation of the study is that it is only performed within a Dutch sample. Although not all measures were available for all four rehabilitation centers, the sample is still sufficiently powered for its purpose. Further research may be needed to investigate the psychometric properties of the AAQ-II-P in non-Dutch samples, which may also be used to replicate the results of this study and to further investigate and develop the psychometric properties of the AAQ-II-P. Based on the results of the current study we conclude that the Dutchlanguage version of the AAQ-II-P can be used as an internally consistent and valid instrument to measure experiential avoidance of pain.
Acknowledgements All authors discussed the results and commented on the manuscript; contributed substantially to conception and design, or acquisition of data, or analysis and interpretation of data; assisted in drafting the article or revising it critically for important intellectual content; and gave final approval of the version to be published.
Conflicts of interest
There are no conflicts of interest.
References Aaronson NK, Muller M, Cohen PD, Essink-Bot M, Fekkes M, Sanderman R, et al. (1998). Translation, validation, and norming of the Dutch language version of the SF-36 health survey in community and chronic disease populations. J Clin Epidemiol 51:1055–1068. Bernini O, Pennato T, Cosci F, Berrocal C (2010). The psychometric properties of the chronic pain acceptance questionnaire in Italian patients with chronic pain. J Health Psychol 15:1236–1245. Bond FW, Hayes SC, Baer RA, Carpenter KM, Guenole N, Orcutt HK, et al. (2011). Preliminary psychometric properties of the Acceptance and Action Questionnaire-II: a revised measure of psychological inflexibility and experiential avoidance. Behav Ther 42:676–688. Fish RA, McGuire B, Hogan M, Morrison TG, Stewart I (2010). Validation of the chronic pain acceptance questionnaire (CPAQ) in an internet sample and development and preliminary validation of the CPAQ-8. Pain 149: 435–443. Fledderus M, Oude Voshaar MA, Ten Klooster PM, Bohlmeijer ET (2012). Further evaluation of the psychometric properties of the Acceptance and Action Questionnaire-II. Psychol Assess 24:925–936. McCracken LN (1999). Behavioral constituents of chronic pain acceptance: results from factor analysis of the Chronic Pain Acceptance Questionnaire. J Back Musculoskeletal Rehabil 13:93–100. McCracken LM, Vowles KE, Eccleston C (2004). Acceptance of chronic pain: component analysis and a revised assessment method. Pain 107: 159–166. Nicholas MK, Asghari A (2006). Investigating acceptance in adjustment to chronic pain: is acceptance broader than we thought? Pain 124:269–279. Reneman MF, Dijkstra A, Geertzen JH, Dijkstra PU (2010). Psychometric properties of chronic pain acceptance questionnaires: a systematic review. Eur J Pain 14:457–465. Sullivan MJL, Bishop SR, Pivik J (1995). The pain catastrophizing scale: development and validation. Psychol Assess 7:524–532. Trompetter H, Klooster PM ten, Ko¨ke A, Schreurs KMG (2011). Acceptance of pain; problems with the factorial validity of the Dutch translation of the chronic pain acceptance questionnaire (CPAQ) [in Dutch]. Psychologie Gezondheid 39:292–299. Van Damme S, Crombez G, Bijttebier P, Goubert L, Van Houdenhove B (2002). A confirmatory factor analysis of the pain catastrophizing scale: invariant factor structure across clinical and non-clinical populations. Pain 96:319–324. Vowles KE, McCracken LM, McLeod C, Eccleston C (2008). The chronic pain acceptance questionnaire: confirmatory factor analysis and identification of patient subgroups. Pain 140:284–291. Ware JEJ, Sherbourne CD (1992). The MOS 36-item short-form health survey (SF-36). I. conceptual framework and item selection. Med Care 30:473–483. Wicksell RK, Renofalt J, Olsson GL, Bond FW, Melin L (2008). Avoidance and cognitive fusion – central components in pain related disability? Development and preliminary validation of the psychological inflexibility in pain scale (PIPS). Eur J Pain 12:491–500. Wicksell RK, Lekander M, Sorjonen K, Olsson GL (2010). The psychological inflexibility in pain scale (PIPS). Statistical properties and model fit of an instrument to assess change processes in pain related disability. Eur J Pain 14:771.e1–771.e14.
Appendix Below you will find a list of statements. Please rate how true each statement is for you by circling a number next to it. Use the scale below to make your choice. Table A1 1 Never true
Acceptance and Action Questionnaire II-pain version 2
3
4
5
Very seldom true
Seldom true
Sometimes true
Frequently true
1. My pain makes it difficult for me to live a life that I would value 2. I’m afraid of my pain 3. I worry about not being able to control my pain 4. My pain prevents me from having a fulfilling life 5. Pain causes problems in my life 6. It seems like most people are handling their lives better than I am 7. Pain gets in the way of my success
6
7
Almost always true 1 1 1 1 1 1 1
2 2 2 2 2 2 2
Always true 3 3 3 3 3 3 3
4 4 4 4 4 4 4
5 5 5 5 5 5 5
6 6 6 6 6 6 6
7 7 7 7 7 7 7
This is a one-factor measure of experiential avoidance of pain. Score the scale by summing the seven items. Higher scores equal greater levels of experiential avoidance of pain.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
