Cross-cultural adaptation, reliability, and validity of a Chinese version of the pelvic girdle questionnaire

The PGQ is a simple instrument consisting of 25 items, including 20 items that assess activity limitations (activity subscale) and 5 items that evaluate symptoms (symptom subscale) [14]. Each item has 4 points and is scored using a Likert-type scale ranging from 0 to 3. The maximum possible score for the PGQ is 75, including 60 for the activity subscale and 15 for the symptom subscale. For an item without an answer or with a response of “not applicable”, three points are subtracted from the total possible score. The final data are presented as a percentage ranging from 0 (without disability) to 100 (severe disability). A higher score indicates a worse outcome.

The ODI (2.0 version) is one of the most commonly used tools in clinical practice for assessment of low back pain [11, 25]. Thus, we applied this instrument in our study to serve as a possible control in analyzing the relevance of the PGQ and ODI. This scale has 10 items, including pain intensity, personal care, lifting, walking, sitting, standing, sleeping, sexual life, social life and traveling. Each item has six response categories, ranging from 0 to 5 points [25]. The final data are divided into five categories (without disability, moderate disability, severe disability, crippled, and bedbound or exaggerating their symptoms) presented as a percentage score ranging from 0–100% with a 20% interval.

The 11-point NRS (0–10) is a tool commonly used to measure pain intensity [19]. It is labeled from 0 to 10, with 0 representing no pain and 10 reflecting the worst pain possible.

Approval was obtained from one of the authors of the original questionnaire before the translation process was started. The translation of the English version of the PGQ into Chinese was completed according to a 5-stage approach based on the guidelines for translation and cultural adaptation of questionnaires [26]. The detailed steps are as follows. The English version of the PGQ was independently translated into Chinese by two native Chinese speakers with a good command of English. Neither of the translators were previously familiar with the PGQ or PGP. One of the two translators was an orthopedic surgeon, and the other was a physiatrist. After translation, the two translated PGQs (Chinese versions) were compared with each other and synthesized, and then compared with the English version. Meanwhile, the synthesized Chinese version was conversely translated into English by two native English speakers independently. Both English translators, a professional translator and a physiotherapist, were not familiar with the original instrument before translation. After that, the two English translators were then introduced to the PGQ. A back-translation version (English) was obtained after an expert committee discussed the translations and resolved discrepancies. Experts proficient in psychometry, linguistics and PGP were included in this expert committee. For the purpose of capturing the accurate meaning of each item, the committee could refer back to the original Chinese version when difficulties were encountered in translations. A total of 18 women with PGP (11 pregnant women and 7 postpartum women) were enrolled to evaluate the prefinal version of the Chinese version of the PGQ. All the participants were required to explain the meaning of each point in the questionnaire and to describe any difficulty they experienced in understanding the question or statement. In addition, the participants were asked to reflect on the relevance of each item in the PGQ to their PGP situation. The women were asked to re-state the item if they encountered difficulties. Finally, all inconsistencies and difficulties were discussed until a consensus was reached by the expert committee.

A total of 151 women suffering from PGP and attending our hospital were recruited for the second part of the study. All of the participants were clinically examined and confirmed to have PGP based on the recommendations from the European guidelines [24]. All the participants were enrolled from May 2019 to January 2020. Twenty-one participants were removed from the study because they did not complete the PGQ (missing items were found after the PGQ was collected). Finally, 130 participants were included in this study, and 67 of the included participants completed the PGQ again within 7 days. These data were used to evaluate the test–retest reliability. None of the patients enrolled in this study received any treatment for PGP.

The inclusion and exclusion criteria for the study were: PGP presenting in pregnancy or within 3 weeks after delivery; pain between the posterior iliac crest and the gluteal fold, particularly in the vicinity of the sacroiliac joint; pain radiating in the posterior thigh and also occurring in conjunction with/or separately from the symphysis; positive result on posterior pelvic pain provocation test and active straight leg raise test; positive result on the long dorsal sacroiliac ligament (LDL) test or symphysis pain palpation test, or positive for pain provocation of the symphysis by Modified Trendelenburg’s test; absence of hip disease, spondylolisthesis, lumbar disease or any autoimmune disease; no prior surgery related to the spine, pelvis or lower limbs; absence of radicular pain below the knee; absence of urinary disease; and absence of neurosensory disorders. Patients who met all the above criteria were diagnosed with PGP.

The construct validity was calculated to test the extent to which scores on the Chinese version of the PGQ related to the theoretical concept of disability in women with PGP. The method reported by Annelie et al. was used to assess the divergent and convergent evidence of the construct validity [19], including five hypotheses: i) high correlation between the scores on the ODI and the activity subscale of the Chinese version of the PGQ; ii) moderate correlation between the scores on the ODI and the symptom subscale of the Chinese version of the PGQ; iii) high correlation between the scores for pain intensity and the total score on the Chinese version of the PGQ; iv) high correlation between the scores for pain intensity and the symptom subscale of the Chinese version of the PGQ; and v) moderate correlation between the scores on the NRS and the total score on the Chinese version of the PGQ.

The area under the receiver operating characteristic (ROC) curve (AUC) was computed to evaluate the discriminative validity of the instrument for the following two aspects. First, the ROC curve was used to discriminate women with treatment or treatment needs reported in the instrument from women without treatment needs. Second, the ROC curve was applied to distinguish women during pregnancy from women in the postpartum period.

To further confirm the reliability of the questionnaire, 67 patients were asked to complete the questionnaire twice within 7 days. The intraclass correlation coefficient (ICC) was calculated to assess the sensitivity and reproducibility of the test. An ICC value greater than 0.80 indicated satisfactory reproducibility [27, 28]. The standard error of measurement (SEM) was also calculated based on the ICC value and its standard deviation using a previously reported formula [29, 30].

Spearman’s correlation coefficients were computed to assess the association between the PGQ and ODI or NRS. Values less than 0.3 indicate low correlation; values from 0.3 to 0.6 indicate moderate correlation; and values greater than 0.6 indicate high correlation [19]. The internal consistency was computed using Cronbach’s alphas, and values ranging from 0.7 to 0.95 implied a good internal consistency. ROC curve analysis was performed to evaluate the discriminative ability of the Chinese PGQ. All statistical analyses were conducted using SPSS software (version 22.0, SPSS, Inc., Chicago, IL, USA). All results are presented as mean values with standard deviations.

All procedures conducted in this study involving human participants complied with the ethical standards of the Handbook for Good Clinical Research Practice of the World Health Organization and the Declaration of Helsinki principles (https://​www.​wma.​net/​policies-post/​wma-declaration-of-helsinki-ethical-principles-for-medical-research-involving-human-subjects/​).

The final analysis included 80 pregnant women (61.5%) and 50 postpartum women (38.5%). The mean age of these women was 32.2 years (SD 3.7 years), and the average gestational age in weeks was 29.7 (SD 8.1). One hundred twenty-five (96.2%) patients had a bachelor’s degree or higher. The detailed characteristics of the participants enrolled in the second part of this study are presented in Table 1. In general, most participants considered the items of the Chinese version of the PGQ to be concise, easy to understand, and strongly relevant to the symptoms of PGP. Two participants considered it hard to correctly understand the introductory text, the 16th item “Carry out sporting activities”, and the 19th item “Have a normal sex life”. Four participants had difficulties in correctly interpreting the meaning of the 23th item “Has your leg/have your legs given way”.

None of the tools included in this study showed ceiling or floor effects for either the total score or subscales scores of the PGQ. Eighteen women chose not applicable for the 16th item “Carry out sporting activities”, and 18 women chose not applicable for the 19th item “Have a normal sex life”. Thirty-four women did not provide an answer for the item about sex in the ODI.

The correlations among the Chinese version of the PGQ, ODI, and NRS for pain intensity are presented in detail in Table 2. The correlation coefficients for the relationships between the PGQ total score and ODI and NRS scores were 0.84 and 0.73, respectively. In addition, high correlations were found between the PGQ activity and symptom scores and the ODI score (0.82 and 0.71, respectively) and NRS score (0.80 and 0.72, respectively; Table 2). For the construct validity of the translated PGQ, five formulated hypotheses were evaluated, and four of them were confirmed: PGQ total to ODI, PGQ activity subscale to ODI, PGQ total to NRS, and PGQ symptom subscale to NRS (Table 3). For the discriminative validity of the translated PGQ, the results from ROC curve analysis showed that the Chinese version of the PGQ could significantly distinguish pregnant/postpartum women with PGP who received or requested treatment from those who do not need treatment (Fig. 2, Table 4). The AUC values for the PGQ total score, activity subscale score and symptom subscale score ranged from 0.772–0.918 (p < 0.001). The sensitivity and specificity of the translated PGQ for correctly predicting treatment and non-treatment were 0.57–0.70 and 0.82–0.97, respectively. However, the ROC curve analysis showed no significant discriminative ability of the PDQ for distinguishing pregnant women and postpartum women with PGP (Supplementary Table 1).

For the factor analysis, the Kaiser–Meyer–Olkin value was 0.824 and the p value of the Bartlett sphericity test was less than 0.001. A forced solution with three components was rotated using the Varimax method (Supplementary Table 2, Supplementary Figure 1).

Test–retest reliability was calculated with a random sub-sample of 67 patients who repeated the Chinese version of the PGQ after 1 week. The result of the internal consistency test showed satisfactory reproducibility (ICC > 0.77) for the total PGQ (ICC = 0.93), the PGQ activity subscale (ICC = 0.92), and the PGQ symptom subscale (ICC = 0.77) (Table 5). The SEM was 3.94%, indicating a low discrepancy between the results of a particular evaluation and the average of all the results hypothetically possible for a patient.