A validation study of the Cantonese Chinese version of short form McGill pain questionnaire 2 in Cantonese‐speaking patients with chronic pain in Hong Kong

The study tests the reliability and validity of the Cantonese Chinese version of Short Form McGill Pain Questionnaire 2 (SF‐MPQ‐2‐CC).


I N T RODUC T ION
The advances in modern pain research have led to the conceptualization of pain as a subjective, multidimensional experience, in which verbal report forms the basis for diagnosis and evaluation of pain. 1,2The McGill Pain Questionnaire and its short form versions (MPQ, SF-MPQ, and SF-MPQ-2) were among the commonest pain assessment tools used in clinical and research settings worldwide.There is strong evidence to support the clinical relevance of MPQ pain descriptors in a wide range of pain conditions. 3][6] SF-MPQ-2 categorizes pain descriptors into four subscales, namely, continuous pain (descriptors 1, 5, 6, 8, 9, and 10), intermittent pain (descriptors 2, 3, 4, 11, 16, and 18), predominantly neuropathic pain (descriptors 7, 17, 19, 20, 21, and 22), and affective (descriptors 12,  13, 14, and 15).The intensity scale for each item was expanded from 4 (0-3) to 11 (0-10) to increase its sensitivity to changes in response to treatment. 4,5Although it was not designed for any specific pain condition, it has been shown to reliably assess musculoskeletal pain conditions, [7][8][9] visceral pain, 10 cancer pain, 11 and neuropathic pain. 12F-MPQ-2 has been validated and translated into 33 languages in 54 regions, and a Standard Mandarin Chinese version (SF-MPQ-2-CN) had been validated for Mandarin-speaking Chinese patients in Hebei Province of China.10,13 However, according to the 2016 Hong Kong Population By-census data, 96.7% of the population uses Cantonese as their usual spoken language.14 This study is designed to develop a culturally appropriate Cantonese Chinese version of SF-MPQ-2 for Cantonese-speaking chronic pain patients in Hong Kong.

M ET HOD S A N D M AT ER I A L S Translation and procedures
After obtaining a formal translation agreement, the original English version of SF-MPQ-2 was translated into Cantonese Chinese according to the linguistic validation guidelines of the Mapi Research Trust. 15Forward translation was performed independently by two translators.Translator 1 was a consultant pain medicine specialist with anesthesia background who was native in Cantonese Chinese and bilingual in English.Translator 2 was a non-medical person with linguistic background, who was also native in Cantonese Chinese and bilingual in English.The translated version was written in Traditional Chinese, which is the commonest form of written Chinese in Hong Kong.The two translated versions were reconciled into one pooled version after a discussion between the lead author and two translators.Backward translation was performed by a third translator, a statistician at the Department of Anaesthesiology, University of Hong Kong who was bilingual in English and Cantonese Chinese.The translated version was then compared with the original English version, and all discrepancies were discussed and reconciled after discussion with the lead author and three translators to produce a revised Cantonese Chinese version.This revised version was then administered to five randomly selected Cantonese-speaking chronic pain patients recruited from the main study center (Queen Mary Hospital).All participants did not report any difficulty in interpreting the wordings, and this version is adopted as the finalized Cantonese Chinese version of SF-MPQ-2 (SF-MPQ-2-CC).

Comparison with other assessment tools
The SF-MPQ-2-CC was compared with validated Hong Kong Chinese versions of Identification Pain Questionnaire for Neuropathic Pain (ID Pain), Pain Catastrophizing Scale (PCS), and 36-item Short-Form Health Survey (SF-36) to assess its convergent or divergent validity.

Identification pain questionnaire (ID pain)
The Chinese version of Identification Pain Questionnaire for neuropathic pain (ID Pain) has been translated and validated in Hong Kong. 16ID Pain was developed as a screening tool for identifying the presence of neuropathic pain in respondents. 17It is a short questionnaire with six questions: (1) Did the pain feel like pins and needles (2) Did the pain feel hot/burning (3) Did the pain feel numb (4) Did the pain feel as though you had electric shocks (5) Is the pain made worse with the touch of clothing or bed sheets and (6) Is the pain limited to your joints?A score of three or more is suggestive of the presence of neuropathic pain.

Pain catastrophizing scale (PCS)
Pain catastrophizing scale is a self-reporting assessment tool designed to assess thoughts and feelings that individuals may experience when they are in pain. 18It contains 13 items of thoughts or feelings and respondents are asked to indicate the extent to which they experienced each of those in their past pain experience.The PCS conceptualizes pain catastrophizing as a multifaceted construct with three subscales: Rumination, Magnification, and Helplessness.The PCS has demonstrated good internal consistency (Cronbach's α = 0.87), test-retest reliability (r = 0.75), and construct validity. 18A Chinese version of PCS has been validated in Chinese patients in Hong Kong, and it is part of the standardized pain intake form used in public pain clinics in Hong Kong. 19ort form health survey (SF-36) The short form health survey was developed and validated as a self-reporting instrument for measuring health perception in general population. 20It consists of eight multi-item domains: Physical Functioning (PF), Role Physical (RP), Bodily Pain (BP), General Health (GH), Vitality (VT), Social Functioning (SF), Role Emotional (RE), and Mental Health (MH).A Chinese version of SF-36 had been validated in Hong Kong population and it is also part of the standardized pain intake form used in public pain clinics in Hong Kong. 18

Participants
This study was conducted in three major pain centers in Hong Kong: Queen Mary Hospital, United Christian Hospital, and Pamela Youde Nethersole Eastern Hospital.The study was registered at Clini calTr ials.gov (Identifier: NCT02513251).The study protocol was approved by Research Ethics Committee of all participating centers.Participants must fulfill the following inclusion criteria: presence of chronic pain for more than 3 months, aged 18 or above, native Cantonese Chinese speaker, ability to understand, read, and write Traditional Chinese, mentally capable to give consent, and ability to complete all study questionnaires.Those who are non-native Cantonese Chinese speaker, unable to read or write Traditional Chinese, aged below 18, mentally incapable to give consent or complete the questionnaires, their pain lasting for less than 3 months, or required active pain treatment within 2 weeks were excluded.After obtaining their written consent, all participants were interviewed by investigators for their demographic information, while their pain diagnoses and other relevant clinical data were retrieved from the electronic record system.They were then given two sets of identical questionnaires (SF-MPQ-2-CC, ID Pain, PCS, and SF-36), and were asked to complete one set on spot.The participants were then given an identical set of questionnaires to complete and return 2 weeks later via mail.

Statistical analysis
According to Kyriazos, a suggested ratio of 10 participants per item was used to calculate a sample size of 220 for this study. 21Statistical analysis was performed using IBM SPSS Statistics for Windows, Version 26.0 (IBM Corp., Armonk.NY, USA computer software) and R 4.0.2 with library -lavaan and library -VIM.Descriptive statistics were used to summarize demographic and clinical data, where continuous data are expressed as mean (SD) if data are normally distributed, while median (25th-75th percentile) will be used if data are not normally distributed.Categorical data were expressed as absolute proportions N (%).Internal consistency was expressed using Cronbach's alpha coefficient, with a value of 0.7 or above considered as having good internal consistency.Test-retest reliability was assessed by Wilcoxon signed-rank test and Spearman's rank correlation coefficient, as the data were not normally distributed.A p-value of Wilcoxon Signed-rank test over 0.05 and Spearman's rank correlation coefficient over 0.7 is considered satisfactory.Construct validity was assessed using confirmatory factor analysis (CFA).CFA was performed by the weighted least square mean and variance (WLSMV) estimation method based on polychoric correlation matrix.Missing values were imputed by K-nearest neighbor (KNN) method.Goodness of fit was evaluated by the following criteria: comparative fit index (CFI) > 0.9, Tucker-Lewis index (TLI) > 0.9, root mean square error of approximation (RMSEA) < 0.08, and standard root mean square residual (SRMR) < 0.08. 22,23onvergent and divergent validity were analyzed by Spearman's correlation coefficient, in which r > 0.5 and p < 0.05 could represent a satisfactory convergent validity. 24

R E SU LT S
A total of 339 participants were recruited from these three pain centers from July 2015 to May 2021, of which 333 of them completed the first set of questionnaires and 197 of them completed the second set of questionnaires (see Figure 1).

Characteristics of participants
Table 1 summarizes the demographic data of participants.The mean age of the participants was 52.44 years (SD 12.03).61.6% of the participants were female, 61.5% were married, and 88.1% had educational level of secondary school or above.Around half of them were employed (49.5%).21.9% of the participants reported their pain resulting from injury at work, of which 69.4% of them received compensation (15.2% of all participants).The pain diagnoses of participants were retrieved from an electronic medical record system and coded according to the International Classification of Diseases 11th Revision (ICD-11).The commonest pain diagnosis was chronic primary musculoskeletal pain (36.0%), followed by chronic neuropathic pain (24.6%) and chronic postsurgical or posttraumatic pain (13.8%).9.6% of the participants had more than one pain diagnosis.

Reliability
Table 2 summarizes the range of scores, Cronbach's alpha, Wilcoxon Signed-rank test, and Spearman's r coefficients of SF-MPQ-2-CC.The Cronbach's alpha coefficient of the total score of SF-MPQ-2-CC was 0.933, demonstrating excellent internal consistency.The internal consistencies of the subscales were good as well, with Cronbach's alpha coefficients ranging from 0.749 to 0.842 (Table 4).The Spearman's rank correlation coefficient was 0.875 (p < 0.05) for total score and 0.761-0.861(p < 0.05) for subscale scores together with p > 0.05 of Wilcoxon Signedrank test demonstrating good test-retest reliability.
Interclass correlation coefficient (ICC) was not used due to the skewed distribution of our data.

Construct validity
Table 3 summarizes the fit indices and Figure 2 shows the path diagram representing the four-factor model of CFA for the SF-MPQ-2-CC.Construct validity is the degree to which the scores of an instrument are consistent with the underlying hypothesis.A central concern of construct validity is to link the observed variables with the attributes of latent variables.The CFA of SF-MPQ-2-CC showed good fit with a second-order factor model (χ 2 = 826.51,p < 0.001, CFI = 0.92, TLI = 0.908, RMSEA = 0.097; SRMR = 0.063; error terms adjusted).second-order CFA was preferred as the result revealed a better fit than the first-order CFA.

Validity of reliability of SF-MPQ-2-CC
In contrast to other validation studies of translated versions of SF-MPQ-2 where participants with one pain condition were often used, this study included participants with a wide range of pain conditions, including both cancer and neuropathic pain, thus enhancing the applicability of our results to the chronic pain population in Hong Kong.Our analysis results show that SF-MPQ-2-CC has excellent internal consistency, with Cronbach's alpha coefficients of total and all subscale scores comparable to those in the original SF-MPQ-2 validation studies. 4,5est-retest reliability was demonstrated by Spearman's r correlation coefficients >0.7 for total and all subscales in 197 participants.Construct validity was established by a good fit to a four-factor CFA model.Intercorrelations within all four latent factors were observed, a phenomenon also observed in previous observations. 4,5,26Results from the second-order CFA suggested the presence of a "global" pain construct that jointly accounts for all four qualities (continuous, intermittent, neuropathic, and affective) of pain, which is similar to the finding in the study by Dworkin et al. 5 While this may suggest that all pains share a degree of common quality, it may also be worthwhile to tease out ambiguous descriptors to improve the discriminating power of SF-MPQ-2 in different types of pain.

Clinical utility of SF-MPQ-CC
The understanding and conceptualization of pain as a sensory and emotional experience.The current version of SF-MPQ-2 contains descriptors encompassing the sensory and affective aspects of pain experience with an improved scaling system, making it easier for clinical use.
Verbal descriptors of pain are important in providing semantic anchors to pain qualities, intensity, and unpleasantness. 2SF-MPQ-2-CC provides a standardized list of pain descriptors for the majority of Cantonese-speaking patients in Hong Kong.This list of culturally appropriate pain descriptors also enables patients to verbalize and communicate their pain experience.Although some descriptors may not be specific enough for diagnosing pain conditions (eg continuous and intermittent descriptors), the SF-MPQ-2 provided a broad coverage over sensory and affective aspects of pain experience, which may be useful in guiding treatment options and as an outcome measure for responsiveness to treatment. 3Recent studies suggested that there may be specific linguistic metaphorical markers for certain pain conditions such as endometriosis, complex regional pain syndrome, neuropathic pain, pain catastrophizing, and mood. 34The identification of these specific linguistic markers may help to refine pain descriptors in future revisions of the McGill Pain Questionnaires.It is also important to remember that pain is a multidimensional experience, and the qualitative descriptors as measured by SF-MPQ-2 should always be interpreted with a wider functional and contextual perspective. 3

SF-MPQ-2-CC as a screening tool for neuropathic pain
The neuropathic subscale of SF-MPQ-2-CC showed a positive correlation with ID pain, which is a commonly used screening tool for neuropathic pain.Similar findings have identified correlation between the neuropathic subscale with other neuropathic pain screening tools, such as Douleur Neuropathique 4 (DN4) and the diagnosis of neuropathic pain by pain physicians. 9,26A recent study by Packham et al suggested that the neuropathic subscale could be used as a "stand-alone" measure for neuropathic features of pain. 27However, SF-MPQ-2 does not generate a composite score of diagnostic value, hence it is not suitable for diagnostic use. 5,28Nevertheless, it may still serve as a flag signal to clinicians for the presence of neuropathic pain in respondents.

SF-MPQ-2-CC as a screening tool for negative pain cognitions in patients with pain
Our study has identified a positive correlation between the SF-MPQ-2-CC (total and affective subscale) and PCS (total and helplessness subscale).The three subscales of PCS (rumination, magnification, and helplessness) represent negative cognitions in patients, and it has been shown to predict pain intensity, disability, and psychological distress. 29,30Helplessness has been identified to predict greater pain severity, pain-related interference, and depressed mood. 31There is evidence that therapies that specifically target helplessness may be particularly useful in improving the outcomes of patients with pain conditions enrolled interdisciplinary pain rehabilitation program. 32SF-MPQ-2-CC and its affective subscale in particular may be able to detect the presence of sense of helplessness, which may help clinicians to identify patients suitable for appropriate treatment (eg interdisciplinary pain rehabilitation program).

Mandarin and Cantonese versions of SF-MPQ-2
SF-MPQ-2 was first translated into Mandarin Chinese (SF-MPQ-2-CN) in 2017. 10According to Ethnologue 2021 data, Chinese is the most spoken native language worldwide, with over 900 million users mostly residing in China. 25Chinese is a macrolanguage encompassing 13 languages, with Mandarin being the largest group, while Cantonese (or Yue Chinese) ranking third with 70 million users.Although Mandarin and Cantonese speakers can communicate with each other without much difficulty, there are considerable differences in vocabularies, pronunciations, and syntactic rules between Mandarin and Cantonese.There were examples of different translated versions of SF-MPQ-2 in other macrolanguages, such as English, Spanish, French, German, etc. 13 Furthermore, the Mandarin version of SF-MPQ-2 was written in Simplified Chinese, while SF-MPQ-2-CC uses Traditional Chinese characters (Hanzi), which is the major form of written Chinese in Hong Kong.SF-MPQ-2-CC is, therefore, culturally more appropriate for clinical use in Hong Kong.However, it is possible to reconcile the Mandarin and Cantonese versions of SF-MPQ-2, as seen in the example of Montreal Cognitive Assessment Chinese-Language Los Angeles version (MoCA-ChLA). 33

CONC LUSIONS
In conclusion, our study has shown that SF-MPQ-2-CC is valid and reliable to use in Cantonese-speaking patients in Hong Kong, based on results from a representative sample of participants encompassing a wide range of chronic pain conditions.It may be used as a clinical tool for initial assessment, treatment selection, and monitoring of responsiveness in Hong Kong pain clinics.

AC K NOW L E DGM E N T S
McGill Pain Questionnaires have evolved since its conception in the 1970s, and it has contributed to the T A B L E 2 Range of scores, internal consistency, and test-retest reliability of SF-MPQ-2-CC.SF-MPQ-Note: SF-MPQ-2-CC = Cantonese Chinese version of Revised Short Form McGill Pain Questionnaire. a Test scores expressed by Median (25th-75th percentile).b Wilcoxon Signed-rank test.c Spearman's r correlation coefficient.*p < 0.05.
This study was supported by Peter Hung Professorship granted to Professor Chi Wai Cheung, and by the Department of Anaesthesiology, the University of Hong Kong.The authors would like to thank Dr SL Tsui from the Department of Anaesthesiology, Queen Mary Hospital and Dr SW Choi from the Department of Anaesthesiology, University of Hong Kong for participating in the translation process of SF-MPQ-2, Mr Edward Choi for performing statistical analysis, and clinical staff from the Pain Clinic of Queen Mary Hospital (Ms Maggie Tsang and Ms YS Man), Kowloon East Pain Management Centre of United Christian Hospital (Dr HS Lim, Dr Febbie Chung, and Ms PC Chan), and Pain Clinic of Pamela Youde Nethersole Eastern Hospital (Dr Alfred Chan and T A B L E 4 Spearman's r correlation coefficient among SF-MPQ-2-CC, SF36, ID-Pain, PCS.

1
Flowchart of recruitment of participants.PYNEH, Pamela Youde Nethersole Eastern Hospital; QMH, Queen Mary Hospital; UCH, United Christian Hospital.339 parƟcipants were recruited in QMH, PYNEH, UCH during the period of 31 th July 2015 to 31 st May

Table 4
Incremental fit and absolute fit indices of SF-MPQ-2-CC.
T A B L E 3