The 27-item coping strategies Questionnaire – revised : confirmatory factor analysis , reliability and validity in italian-speaking subjects with chronic pain

1Physical Medicine and Rehabilitation Unit, Salvatore Maugeri Foundation IRCCS, Institute of Care and Research, Scientific Institute of Lissone; 2Neuroengineering and Medical Robotics Laboratory, Department of Electronics, Information, and Bioengineering, Politecnico di Milano, Milan; 3Operative Unit of Psychology, Salvatore Maugeri Foundation, Institute of Care and Research, Scientific Institute of Pavia, Pavia; 4Physical Medicine and Rehabilitation, Tor Vergata University of Rome, Rome, Italy Correspondence: Dr Marco Monticone, Via Monsignor Bernasconi 16, 20035 Lissone, Italy. Telephone 39-039-465-7277, fax 39-039-465-7279, e-mail marco.monticone@fsm.it Chronic pain is characterised by physical dysfunction, disability and mood alterations (1). However, not everyone experiencing chronic pain is physically disabled or invariably depressed; many continue their work and social activities, and rarely seek medical assistance or the help of a significant other (1). Therefore, it appears to be crucial to identify the cognitive factors that may promote adaptive functioning despite the presence of pain, such as coping strategies (ie, the use of cognitive and behavioural techniques to manage stressful events), because these can help to explain differences in adjustment among subjects experiencing chronic pain, determine cognitive strengths and weaknesses, identify treatment targets and predict outcomes (2). The Coping Strategies Questionnaire (CSQ) was developed in 1983 by Rosenstiel and Keefe (3) using a pool of items reflecting coping strategies frequently reported by patients and deemed to be important by researchers and clinicians involved in the management of pain. Although it had good reliability and validity, it showed an unstable structure (3-5) that has been hypothesized to be mainly due to difficulties in measuring differences in cognition between different clinical settings, disorders and pain problems (6). In 1996, an exploratory factor analysis of a large sample of subjects with chronic pain suggested a six-factor solution that was relatively supportive of the original scales, and showed satisfactory reliability and construct validity (7). A subsequent confirmatory factor analysis (CFA) established the good fit of this model (ie, the CSQ-Revised) (8). original arTicle

Translations of the CSQ have been validated and allow comparisons among different populations and countries (9)(10)(11).However, the CSQ-Revised version has yet to be adapted and psychometrically analyzed in Italian subjects, thus limiting the opportunities for researchers and clinicians to share the validated outcomes of chronic pain patients.
The aim of the present study was to describe the cultural adaptation of the CSQ-Revised and its validation in a large sample of subjects with chronic pain to enable its use in Italian-speaking subjects in Italy and abroad.It was hypothesized that internal consistency would be acceptable, that test-retest reliability would be good, and that CSQ-Revised maladaptive coping strategies would be statistically significantly and positively related to poor outcomes of pain intensity, disability and depression.It was also expected that the CSQ-Revised adaptive and maladaptive strategies would be statistically significantly and positively related to adaptive and maladaptive strategies taken from another coping questionnaire, respectively.

METHODS
The present cross-sectional study was approved by the Institutional Review Board of the authors' hospital (Salvatore Maugeri Foundation's Scientific Institute, Lissone, Italy), and the patients provided written informed consent to participate.

Subjects
Outpatients attending the Rehabilitation and the Psychology Units of Salvatore Maugeri Foundation's Scientific Institute were recruited between January 2011 and September 2012.The inclusion criteria were: chronic nonspecific pain (ie, lasting continuously for >3 months); age >18 years; and fluency in Italian.The exclusion criteria were: acute (ie, lasting <1 month) and subacute (ie, lasting one to three months) nonspecific pain; central or peripheral neurological signs; systemic illness (eg, tumours and rheumatological diseases); and cognitive impairment (ie, deficits in higher reasoning, forgetfulness, learning disabilities, concentration difficulties, decreased intelligence and other reductions in mental functions).Patients who had experienced recent cerebrovascular accidents, myocardial infarctions, or chronic lung or renal diseases were also excluded.
The subjects' demographic and clinical characteristics were recorded by a research assistant using a specific form.

CSQ-Revised
This 27-item questionnaire measures the use of strategies for coping with pain by assessing six domains: Distraction; Catastrophizing; Ignoring pain sensations; Distancing from pain; Coping selfstatements; and Praying.Patients rate the frequency of their use of the specific strategies using a seven-point Likert scale ranging from 0 "Never do that" to 6 "Always do that"; each domain is scored separately, with higher scores indicating greater use (7).The original form was tested in English-speaking subjects and showed acceptable internal consistency (Cronbach's alpha estimates ranging from 0.72 to 0.86) and satisfactory construct validity -ie, maladaptive strategies, such as Catastrophizing and Praying, were positively associated with negative effects and negatively associated with the level of activity, whereas adaptive strategies, such as Ignoring pain sensations and Coping self-statements, were positively associated with the level of activity (7).

Translation and cross-cultural adaptation
This procedure was performed in accordance with international guidelines (12).
Step 1.Translation into Italian: The items derived from the English CSQ-Revised were translated with the aim of retaining the concepts of the original while using culturally and clinically fitting expressions (7).Two translators, whose first language was Italian, each independently translated the English version into Italian, keeping the language colloquial and compatible with a reading age of 14 years.One of the translators was unfamiliar with the measure.The translators then reviewed the two Italian versions together, ensuring that items with poor wording were identified and improved by means of discussion.
Step 1 ended when a common adaptation was agreed on.None of the items were excluded.
Step 2. Back-translation into English: Two independent bilingual translators whose first language was English back-translated the initial translation; they did not have medical backgrounds and were unaware of the concepts being explored.The aim was to ensure the Italian version reflected the same item content as the original version and was conceptually equivalent.
Step 3. Expert committee: The translated versions were submitted to an expert committee of bilingual (Italian and English speaking) clinicians, methodologists, psychometricians and the translators.To identify any difficulties, inconsistencies or mistakes in translation, the committee explored the semantic, idiomatic and conceptual equivalence of the items and answers.Step 3 ended when the committee agreed on a penultimate version of the Italian CSQ-Revised.
Step 4. Testing the penultimate version: The CSQ-Revised was initially administered to 50 patients with chronic nonspecific pain to probe what was meant by each item and the chosen response.The expert committee re-evaluated all of the findings, although no further adjustment was required.

Analytical scale properties
All of the methodological criteria for investigating psychometric properties suggested by Terwee et al ( 13) were followed except for 'responsiveness' (because the present study was not longitudinal).Acceptability: The time needed to answer the questionnaire was recorded.The patients were asked about any problems they encountered and all of the data were checked for missing or multiple responses.Factor analysis: A CFA was performed and each of the items was specified to load on its respective subscale as originally hypothesized (7,8).Model fit was assessed using statistics, a comparative fit index, a normed-fit index, root mean square error of approximation (RMSEA) and the 90% CI of RMSEA (14).Ratios of <3 between χ 2 and df, comparative fit index and normed-fit index values ≥0.90, and RMSEA values ≤0.08 were considered to be indicative of a good fit (15).Floor/ceiling effects: Descriptive statistics were calculated to identify floor/ceiling effects, which were considered to be present when >15% of the subjects obtained the lowest or highest possible subscale scores (13).Reliability: This was tested by means of internal consistency, which reflects the inter-relatedness among the items (Cronbach's alpha, with a value >0.70 being considered acceptable), and test-retest stability, which measures reliability over time (intraclass correlation coefficient 2,1 -with good and excellent reliability being indicated by values of 0.70 to 0.85 and >0.85, respectively) (13).Test-retest reliability was investigated by administering the CSQ-Revised to the same subjects after seven days to avoid the natural fluctuations in symptoms associated with memory effects.Content validity: This assessment was based on the patients' yes/no answers to questions investigating the aim of the measurement ("Do you think pain coping strategies constitute the aim of this questionnaire?"), the target population ("Do you think the items described here may be related to your chronic pain?") and the concepts being measured, with special attention being devoted to relevance ("Do you think these items are relevant to evaluate your pain coping strategies?") and completeness ("Do you think that the items presented comprehensively reflect your pain coping strategies?").The hypotheses were considered acceptable if the rate of correct/affirmative answers was >90% (13).

Construct validity:
The extent to which an instrument's score relates to the score of the theoretical construct of another instrument as expected was investigated by means of hypothesis testing (13).It was hypothesised a priori that the CSQ-Revised maladaptive strategies (ie, Catastrophizing and Praying) would be statistically significantly and positively related to a measure of pain intensity (the 0 to 10 numerical rating scale) (16), a measure of disability (the Italian version of the Sickness Impact Profile -Roland scale) (17), and a measure of depression (the Italian version of the Center for Epidemiological Studies -Depression Scale) (18).It was also hypothesised that the CSQ-Revised maladaptive strategies would be statistically significantly and positively related to maladaptive strategies (Guarding, Resting, Asking for assistance) of another measure of coping (the Italian version of the Chronic Pain Coping Inventory [CPCI]) (19), and that the CSQ-Revised adaptive strategies would be statistically significantly and positively related to adaptive strategies of the CPCI (Relaxation, Task persistence, Exercise/Stretch, Seeking social support and Coping self-statements).Pearson's correlation coefficients were used.
The analyses were conducted using SPSS and AMOS (IBM Corporation, USA).

Other outcome measures Numerical Rating Scale:
This was an 11-point rating scale ranging from 0 (no pain at all) to 10 (the worst imaginable pain) (16).Sickness Impact Profile -Roland scale: This 23-item self-report scale, which was derived from the 136-item Sickness Impact Profile questionnaire, was developed to measure disability in patients with chronic pain.Each item is dichotomous (yes/no), and the sum of the 23 scores is expressed as a total score ranging from 0 (no disability) to 23 (maximum disability as perceived by the patient) (20).The Italian version was used, which has been shown to be reliable and valid (17).Center for Epidemiological Studies -Depression scale: This 20-item self-report questionnaire was developed to assess the presence and severity of depressive symptoms.Each item is answered using a fourpoint scale (0 = rarely; 3 = most of the time); thus, the total score ranges from 0 to 60 (higher scores indicate worse depressive symptoms) (21).The Italian version was used, which has been shown to be reliable and valid (18).CPCI: This 42-item self-report questionnaire asks patients to rate the frequency of their use of coping strategies during the previous week (22).There are eight subscales (Guarding, Resting, Asking for assistance, Relaxation, Task persistence, Exercise/Stretch, Seeking for social support, Coping self-statements) and each item is scored from 0 to 7, with higher scores representing greater use of coping strategies.For each subscale, the answers are summed and divided by the number of items for which a response was provided.Based on the findings of the original developers, Guarding, Resting and Asking for assistance were considered to be maladaptive strategies because they are more illness-focused, while the remaining five subscales were considered to be adaptive (23).The Italian version was introduced, which has been shown to be reliable and valid (19).

Subjects
A total of 380 patients were invited to participate, of whom 345 accepted, resulting in a response rate of 90.8%.There were 188 females (54.5%) and 157 males (45.5%) with a mean age of 58.4±15.8 years (range 21 to 88 years).The median duration of pain was 24 months (range three to 120 months).The sociodemographic characteristics of the participants are presented in Table 1.
The clinical and sociodemographic findings are largely consistent with those found by the original developers of the CSQ-Revised, being representative of subjects with chronic pain (6,7).

Translation and cross-cultural adaptation
The questionnaire was translated into Italian using a process of forward-backward translation involving four translators.It took two months to reach a culturally adapted version; all of the items were easily translated except two questions ("I try to feel distant from the pain almost as if the pain was in somebody else's body" and "I try not to think of it as my body, but rather as something separate from me"), but these difficulties were overcome by means of careful wording.A further review by experts and the testing of the penultimate version confirmed the correctness of the process of translation/back-translation and the content of the items.
The adapted questionnaire is reproduced in Appendix 1.

Analytical scale properties
Acceptability: All of the questions were well accepted.The CSQ-Revised was completed in 11.1±1.5 min, and there were no missing or multiple answers.There were no problems with regard to comprehension.Factor analysis: CFA met all of the fit criteria confirming the model on the present sample (Table 2) (7).The item-scale correlations were satisfactory (Distraction, 0.751 to 0.913; Catastrophizing, 0.849 to 0.891; Ignoring pain sensations, 0.821 to 0.957; Distancing from pain, 0.925 to 0.932; Coping self-statements, 0.829 to 0.926; and Praying, 0.781 to 0.982).
Floor/ceiling effects: No significant effects were found for any of the subscales (Table 3).Reliability: Cronbach's alpha was acceptable for all of the subscales (α=0.914 to 0.961).Test-retest reliability was measured in all of the subjects, and the domains showed good/excellent intraclass correlation coefficients (0.850 to 0.918) (Table 3).

Content validity:
The percentage of affirmative answers was >90% and, thus, the content of the items was considered to be adequate, appropriate for the target population, comprehensive and relevant for investigating coping strategies in this population.Construct validity: Most of the a priori hypotheses were confirmed.As expected, Catastrophizing (from r=0.27 to r=0.49) and Praying (from r=0.14 to r=0.23) were statistically significantly and positively related to other similar constructs.Catastrophizing was moderately correlated with Depression (r=0.49;P<0.01) (Table 4).When the coping questionnaires were compared, Catastrophizing and Praying were statistically significantly and positively related to CPCI maladaptive strategies (from r=0.11 to r=0.26, and from r=0.17 to r=0.29); weaker correlations were found in the case of adaptive strategies, except when CPCI Seeking social support was compared with Catastrophizing (r=0.14) and Praying (r=0.26).Likewise, Distraction, Distancing from pain and Coping self-statements were statistically significantly and positively related to CPCI adaptive strategies (from r=0.25 to r=0.50, from r=0.23 to r=0.43, and from r=0.13 to r=0.54) (Table 5).

DISCUSSION
The present study reports the adaptation of the CSQ-Revised and its validation in a sample of previously uninvestigated Italian patients with chronic pain.Analyzing the psychometric properties of an outcome measure is a continuous process that is strongly recommended to strengthen its properties and expand its applicability to specific populations and contexts (24).Our findings provide further evidence regarding the relationships between CSQ-Revised and CPCI, two widely used questionnaires that assess coping strategies individuals with chronic pain.
The meaning of the original items was adequately captured by the idiomatic translation of the CSQ-Revised.The difficulties encountered by the translators were overcome by means of careful wording.The questionnaire was acceptable and easily understood, and could be self-administered in approximately 10 min.It responded satisfactorily to the requirements of relevance and completeness, and appeared to be fully applicable to everyday clinical practice.No significant floor/ ceiling effects were found, which suggests the scale correctly assesses its construct.
The factorial structure of the CSQ-Revised was confirmed, and the satisfactory item-scale correlations enabled us to include all of the 27 items, as originally proposed (7).This model adequately fits the data obtained from our sample, which suggests that coping strategies can be thoroughly described as a process with six components.French researchers have also performed a CFA using the CSQ-Revised, and achieved satisfactory results consistent with our findings and those of Riley and Robinson (8,9).
Our internal consistency was satisfactory, thus confirming the extent to which the items assessed the same construct.Our estimates were higher than that of the developers of the CSQ-Revised (α=0.72 to 0.86) and the French adaptation (α=0.57to 0.83) (7,9).
The CSQ-Revised also showed satisfactory test-retest reliability in the investigated population and context; however, this psychometric property was not tested in the original and other adapted versions of the CSQ-Revised and, thus, no comparisons are possible.
Consistent with the English findings (8), our estimates of construct validity highlighted the adaptive (ie, Distraction, Ignoring pain sensations, Distancing from pain and Coping self-statements) and maladaptive properties (Catastrophizing and Praying) of most of the subscales (4,7,25).
The correlations between the CSQ-Revised and the CPCI contributed further evidence of the adaptive and maladaptive strategies investigated by both measures.Our findings also suggest that the

ICC Intraclass correlation coefficient
The 27-item CSQ-R CSQ-Revised and CPCI have different constructs, thus highlighting their distinctive contribution to multidisciplinary pain programs and confirming the intent of the original developers of the CPCI to create a questionnaire that investigated previously ignored coping strategies (22).As observed in previous studies (26), exceptions were the correlations between CPCI-Task persistence and CSQ-Ignoring pain sensations, CSQ-Distancing from pain and CSQ-Coping selfstatements; between CPCI-Relaxation and CSQ-Distraction and CSQ-Distancing from pain; and between CPCI-Coping self-statements and CSQ-Distraction and CSQ-Coping self-statements, which suggests that these constructs likely overlap in scale content and require further investigations.There were several limitations to the present study.First, its crosssectional design means that significant correlations should not be confused with causal effects.Second, the relationships between selfreported beliefs and objective measures of coping, such as behavioural observations or reports of cognitive coping during structured or standardized situations, were not considered because only self-administered measures were used.Third, additional studies of the properties of CSQ-Revised using modern test theory methods, such as Rasch measurement theory or item response theory, are recommended because only classical test theory psychometric properties were evaluated.

CONCLUSION
The Italian version of the CSQ-Revised confirmed the factor structure of the original English version and showed good psychometric properties.It can be recommended for use in chronic pain research and multidisciplinary pain assessments.

ACKNOWLEDGEMENTS:
The authors thank Kevin Smart for his help in preparing the English version of the manuscript.

DISCLOSURES:
The authors have no conflicts of interest to declare.

IRB APPROVAL:
The authors' Institutional Review Board approved the study, which was conducted in accordance with ethical and humane principles of research.