Translation, cross‐cultural adaptation and validation of the Pain Catastrophizing Scale (PCS) into Bengali in patients with chronic non‐malignant musculoskeletal pain

Abstract Aim To develop a culturally adapted and validated Bengali Pain Catastrophizing Scale (BePCS). Methods The English PCS was translated, adapted and back‐translated into and from Bengali, pre‐tested by 30 adult patients with chronic non‐malignant musculoskeletal pain. The BePCS was administered twice with 14 days interval to 90 patients. Convergent validity was measured by comparing the BePCS score with scores of the domains physical functioning and mental health of the Bengali Short Form 36, through Spearman's correlation coefficient. Test‐retest reliability was assessed by intraclass correlation coefficient (ICC) and Spearman's rank correlation coefficient and internal consistency by Cronbach's alpha. Content validity was assessed by index for content validity (ICV) and floor and ceiling effects. Results The BePCS was well accepted by the patients in the pre‐test. The content validity was excellent, both item ICV and scale ICV were 1. Construct validity: the convergent validity was −0.424 for physical functioning and −0.413 for mental health, indicating a moderate negative correlation. Total BePCS score showed excellent internal consistency with a mean Cronbach's α = 0.92. Internal consistency for subscales rumination, magnification and helplessness, were Cronbach's α 0.903, 0.72 and 0.872 respectively. The test‐retest reliability of total BePCS was 0.78 (P < .001) and for the subscales rumination 0.872 (P < .001), magnification 797 (P < .001) and helplessness 0.927 (P < .001), showing excellent test‐retest reliability. Conclusions The interviewer‐administered BePCS appears to be an acceptable, reliable and valid instrument for measuring health‐related quality of life in Bengali speaking patients with chronic non‐malignant musculoskeletal pain. Further evaluation in the general population and in different medical conditions should be done.


| INTRODUC TI ON
Pain as a symptom is now considered the 5th vital sign 1 ; it accounts for approximately 80% of physician visits and for an estimated US$ 100 billion annually regarding cost of healthcare and loss of productivity. 2 Chronic non-specific musculoskeletal pain is a burden for patients. It is associated with high socio-economic costs [3][4][5] and significantly affects the psychosocial status of affected people as well as their families and carers. 6 Chronic pain has complex underlying pathophysiology, and is determined by multiple psychological, social and biological factors.
One of these factors is pain catastrophizing, characterized by patients magnifying their feelings about painful situations and continually thinking about these situations. 7 Catastrophizing also involves feelings of helplessness and rumination about pain. Pain catastrophizing is related to multiple health outcomes like pain intensity, interference of pain with patients' lives, physical disability and mental well-being. 8 Pain catastrophizing causes a negative mental setting to bear actual or anticipated pain. 9 Pain feeling has been found to increase from 7% to 33% in pain ratings, depending on the extent of catastrophizing. 10 Catastrophizing plays an important role in pain chronicity and has a positive correlation with pain intensity and disability. 11 It not only causes an increased perception of pain and emotional stress, but also prolongs pain episodes and catastrophizing is a significant predictor of the severity of pain, and of the ways in which people cope with pain. 12,13 Catastrophizing thus influences various substantial pain-related outcomes including: greater pain intensity and chronicity, depression, anxiety, pain-related disability and analgesic use. 14 Pain catastrophizing has been associated with poor pain treatment response in patients with chronic pain. 9 Previous studies reveal that if pain catastrophizing diminishes, pain intensity, disability and chronic conditions would decrease. 15 It appeared possible to modify pain catastrophizing in patients undergoing surgery. 16 In psychological research it is postulated that pain catastrophizers may enact pain behaviors in order to receive support or empathy from their social environment. 17 It has been shown that higher levels of catastrophizing pain behavior were associated with a more intense inference of pain by the observers, which may lead to over-cautious treatment decisions by those who take care of these patients. 17,18 The Pain Catastrophizing Scale (PCS) was developed in 1995 by Sullivan et al to measure the individual degree of pain catastrophizing. The PCS is a multidimensional questionnaire, consisting of 3 subscales: helplessness, magnification, and rumination. The English version of the PCS has been investigated extensively, and its psychometric properties are good. 19,20 The psychometric properties of the questionnaire have been confirmed at least for 10 other languages, including German, Brazilian, Chinese, Portuguese, and Arabic. [21][22][23][24] There are more than 164 million people in Bangladesh 25 and about 265 million Bengali speaking people worldwide and it is the 7th language according to population. 26 A culturally adapted and validated Bengali version of the PCS for the people of Bangladesh is not yet available. The purpose of this study is the translation of the PCS into Bengali, cultural adaptation of the Bengali version and to test its validity and reliability in adult patients with chronic non-malignant musculoskeletal pain.

| Patients
The study has been conducted in the Department of Rheumatology The sample size of the study was 95 patients, as calculated by Study Size 3.0, a validated statistical software developed by Creostat HB35 HB in Sweden. 27 Our expected intraclass correlation coefficient (ICC) for the assessment of test-retest reliability was 0.9 and the minimal acceptable ICC was 0.7. So using a two-sided test with β = 0.2 (80% power) and α = 0.05, the sample size required was 22.873. Thus for the assessment of the test-retest reliability of the questionnaire and considering drop-out of some patients during retesting, a sample size of 32 was considered to be sufficient. These 32 patients were collected by simple random sampling from the 95 patients who were enrolled for the test.

| The PCS
The PCS was developed in 1995 at the University Center for

| Testing of prefinal version
The 2 prefinal Bengali versions of the PCS were tested in a sample of 30 adult patients with chronic non-malignant musculoskeletal pain.
Each subject completing the questionnaires was interviewed to find out what he or she thought was meant by each questionnaire item, and about the response they gave, and whether they had any further suggestions. If a participant was able to understand both of the translations of the same item, he or she was asked which translation (in the prefinal version -1/2) he or she would prefer. Based on the response of these participants, the adapted version was prepared.
The adapted version was administered twice with 14 days interval to 90 Bangladeshi patients who were suffering from chronic non-malignant musculoskeletal pain.
For measuring the physical functioning and mental health, these domains of the Bengali version of the Short Form-36 (SF-36) were applied. 30

| Questionnaire administration
The questionnaire was used as a self-administered one for literate participants and an interviewer-administered one in case of illiterate participants. The literate participants were allowed to read the questionnaire themselves and give the replies as per their own understanding. In case of illiterate participants, the interviewer read the questionnaire in a clearly audible voice, without giving explanation.
The responses were recorded by the interviewer.

| Statistical analysis
All data were assessed using SPSS 22.0 (SPSS Inc). All tests were 2-tailed and conducted at a 5% level of significance. There were no missing data for any items. Both the content validity and construct validity were assessed. Reliability was assessed through three ways: internal consistency, test-retest reliability and item to scale correlation. The internal consistency was measured using Cronbach's alpha. The internal consistency was considered acceptable when Cronbach's alpha was equal to or exceeded 0.70. 31 The item to scale correlation was assessed using Spearman's rank correlation (rho) between scale and their constituent items, taking a value of rho ≥ 0.40 as acceptable. 32 Test-retest reliability was assessed using ICC. An ICC between 0.60 and 0.74 was considered good, between 0.75 and 1.00 was excellent and considered acceptable for test-retest reliability. 33  The Mann-Whitney U test (also called the Mann-Whitney-Wilcoxon (MWW), Wilcoxon rank-sum test, or Wilcoxon-Mann-Whitney test) was used to compare between 2 groups with respect to a variable that does not follow a normal distribution.
The Kruskal-Wallis test (sometimes also called the "one-way analysis of variance on ranks") is a rank-based nonparametric test.
It was used to determine whether there are statistically significant differences between 2 or more groups of an independent variable on a continuous or ordinal dependent variable.
The ICC or the intraclass correlation, is a descriptive statistic. It was used when quantitative measurements are made on units that are organized into groups. It describes how strongly units in the same group resemble each other.

| Ethical clearance
The Institutional Review Board (IRB) of Bangabandhu Sheikh Mujib Medical University provided clearance to conduct the study (No. BSMMU IRB 11606). All the participants were informed in details about the nature of the study. Only the individuals willing to participate in the study were included. Informed written consent was taken from the participants. Every participant enjoyed his/her right to participate or refuse to participate and to withdraw participation at any time. The principal investigator maintained the confidentiality of the information obtained from the participants. Data were intended to be used solely for this study.

| Socio-demographic data
A total of 95 patients could be included in the study. Their mean age was 37 years (SD 13.01), 43 male (45.3%) and 52 female (54.7%). The rheumatological diagnoses are summarized in Table 1. There were 27 patients (24.8%) who were below the secondary education level and 68 patients (71.6%) were at secondary level and above (Table 1).
Thirty-eight patients (40%) came from a rural area and 57 (60%) patients were from the urban area. We found no significant difference between patients BePCS scores and their age (P = .971), gender (Table 2) or educational level (P = .145). Although BePCS scores were lower in people with higher educational levels, the differences were insignificant as per the Kruskal-Wallis test ( Table 3). The BePCS total and subscale scores were higher in females but this difference was not statistically significant ( Table 2).

| Content validity
The I-CVI and the S-CVI were the assessment tools of content validity. All items of the scale showed excellent content validity: both I-CVI and S-CVI were 1.

| Internal consistency and test-retest reliability
The total BePCS score showed an excellent total internal consistency with Cronbach's alpha of 0.92. Internal consistency for subscales rumination, magnification and helplessness, were Cronbach's α 0.903, 0.72 and 0.872, respectively ( Table 4). showing excellent test-retest reliability (Table 5).

| D ISCUSS I ON
The  Age is another factor evaluated in studies associated with PCS scores. 37,38,39 They did not find any correlation between age and the PCS score. In the present study also no significant correlation was found between age and the total PCS or PCS subscale scores. In our study, we looked for a possible relation between educational level and catastrophizing; the PCS scores of the lower literacy group were higher than those of the higher literacy group, but this was not sta- Pain catastrophizing has a social function and could affect family or significant others. It has been found in some previous studies that patients having higher PCS scores consumed higher amounts of analgesics and suffered from chronic and severe pain. 40,41 That is why the PCS has been developed into several other versions. 42

| Limitations
Our study showed some limitations. We could not study a possible correlation between various psychological scores, pain and disability (eg, Beck Depression Inventory, Pain Anxiety Symptom Scale-20 etc) as was done in some other studies (eg, Korean PCS) 13 as these scales have not yet been validated in Bengali. As our study was carried out in a tertiary level hospital, it may not be fully representative for the whole Bengali speaking population. Sensitivity to change could not be evaluated due to temporal constraint.
A strength of the study is that it is the first study in the Bengal language and it will create opportunities to study this important field of catastrophizing and chronic pain in 265 million Bengali speaking people. Our study showed acceptable validity and excellent internal consistency, construct and content validity and reliability of the Bengali version of the PCS.
In conclusion, the BePCS, being a valid and reliable tool, may be used to screen the probability of catastrophizing when suffering from chronic pain. The BePCS can be a valuable tool for patient education, treatment planning and to assess the need for psychological intervention.

ACK N OWLED G EM ENTS
We thank all the patients for their willingness to participate in the study, Mr Jashim Uddin, the Bengali teacher working in Dhaka