Joint hypermobility is defined as a condition in which one or more joints move beyond the normal physiological range of motion and may be associated with hereditary connective tissue disorders [1]. Asymptomatic hypermobility, also referred to as generalised joint hypermobility (GJH), is a common condition affecting between 2% and 57% of the population [2]. It is more frequent in women [3], children [4], and certain ethnic groups, such as Africans and Asians [5, 6]. Asymptomatic hypermobility can confer advantages in activities requiring high flexibility, including music performance, gymnastics, and ballet [7–9].

In contrast, a smaller subgroup presents with symptomatic hypermobility, formally classified as either hypermobility spectrum disorders or hypermobility Ehlers-Danlos syndrome. These individuals typically experience multisystemic clinical manifestations alongside joint hypermobility, such as joint instability [10], chronic pain [11], fatigue [12], autonomic dysfunction [13], gastrointestinal disorders [14], anxiety [15], abnormal skin elasticity and scarring [16], urinary complications [17], and recurrent falls and imbalance [18]. A recent study reported substantial healthcare costs in this population, with mean annual inpatient expenses of $47,900 per person and overall medical costs of $32,800 per person [19]. The evidence suggests that effective treatment for symptomatic hypermobility requires interdisciplinary integration [20–22]. Coordinated collaboration across rheumatology, gastroenterology, neurology, cardiology, pain medicine, rehabilitation, and mental health is essential to address the complex multisystemic nature of these conditions [22]. Such approaches may reduce fragmented care, prevent unnecessary hospitalisations, and improve long-term outcomes.

Currently, the Beighton score (BS), a nine-point scale, is one of the most widely used and validated tools for screening asymptomatic hypermobility, but it requires administration by trained clinicians or researchers [23, 24]. Five-Part Questionnaire (5PQ) is another frequently employed and self-reported assessment tool and has been validated in several populations [25–27]. To date, published translations are available in Brazilian-Portuguese and Swedish, but no validated Chinese version has been reported. Furthermore, although some recent studies have adopted a BS cut-off of ≥ 4 to define GJH in adults [28, 29], the 2017 international classification of the Ehlers-Danlos syndromes recommends a cut-off of ≥ 5 for adults [30]. Lower thresholds may potentially inflate false positive rates in young adults [31]. This inconsistency in diagnostic criteria raises concerns regarding the comparability and validity of prior findings.

To address this gap, the present study applied the internationally recommended cut-off (BS ≥ 5) as the reference standard to evaluate the validity and reliability of the Chinese version of the 5PQ (5PQ-CN) for assessing asymptomatic hypermobility in adults. This study aims to provide a convenient screening tool for the Chinese population and to support future clinical and epidemiological research.

Methods

Study design

This study was prospectively designed, meaning the plan for recruitment, testing, and analysis was finalized prior to data collection. The study procedure is illustrated in Fig. 1. A subgroup of participants was selected for test–retest reliability analysis. This reliability group completed the 5PQ-CN on two occasions: first at an initial visit and again seven days later during the main testing session alongside the rest of the participants. All other participants attended a single main testing session.

At the initial visit (reliability group only), participants completed the 5PQ-CN and a demographic questionnaire. At the main session seven days later, all participants completed the 5PQ-CN. Participants who were not in the reliability group also completed the demographic questionnaire at this time. Following the questionnaire completion, all participants underwent a physical assessment for GJH using the BS. The reliability group completed this physical assessment once, after their second 5PQ-CN administration.

(Fig. 1 HERE)

Title: Fig. 1. Testing procedure

Legend: 5PQ-CN: Chinese version of five-part questionnaire, BS: Beighton Score

Participants

Participants were recruited from two universities in Chengdu, China, in November 2021 using a convenience cluster sampling method. Researchers contacted lecturers in relevant disciplines (including rhythmic gymnastics, sports dance, basketball, football, tennis, sport science, and sport rehabilitation science) to recruit students from their classes. Inclusion criteria were: (1) in good general health; (2) provided consent form to participate; (3) no acute injuries or chronic conditions obstructing the BS test or 5PQ-CN; (4) aged 18 years or above. Exclusion criteria were: (1) declined to participate; (2) unable to undergo joint examinations; (3) physical disability affecting assessment; or (4) under 18 years of age.

A total of 615 participants were recruited for this cross-sectional study. Informed consent was obtained from all participants at the beginning of each testing session. The study was reviewed and approved by the Research Ethics Committee of Chengdu Sport University (Ethical Approval No. [2021] 46) and conducted in accordance with the Declaration of Helsinki. Sample size calculation was performed a priori using G*Power software (v3.1.9.7). Based on the prevalence, sensitivity, and specificity estimates reported by Glans et al. [26], with an alpha (α) of .05 and power (1 – β) of .9, the minimum required sample size was estimated to be 227. To account for potential participant drop-out, this number was doubled, resulting in a target sample size of 454.

Testing methods

The 5PQ-CN was the index test in this study. Originally developed in English by Hakim and Grahame [32], the 5PQ is a 5-item self-report tool asking participants whether they can perform five specific actions. A score of 1 is given for each "yes" response, and a cut-off point of ≥ 2 indicates GJH in adults. The 5PQ-CN was translated by a certified translator, back-translated by a second independent translator, and reviewed for accuracy. The original items and Chinese translations are presented in Table 1. Clarifications were made for Item 4 (explicitly specifying the patella) and Item 5 (adapted to “Do you consider your joints to be hypermobile [excessively flexible]?”).

(Table 1 HERE)

Title: Table 1. The five part questionnaire and its Chinese translation

Legend: 5PQ: five part questionnaire, 5PQ-CN: Chinese version of 5PQ

The BS was used as the reference standard for GJH [23, 31]. It consists of 9 assessments of upper limb, lower limb, and spinal flexibility (Fig. 2). Each item was judged positive (1 point) or negative (0), for a total of 0–9. A cut-off of ≥ 5 was used to define GJH, in line with the 2017 international classification [30]. Although the reliability of the BS has been questioned [23, 33–35], it remains the most widely adopted method. Assessments were conducted by trained researchers blinded to questionnaire results.

(Fig. 2 HERE)

Title: Fig. 2. Beighton Score assessment illustration

Legend: Positive criteria: 1) joint angle ≥ 90°; 2) thumb touches forearm; 3) elbow hyperextension ≥ 10°; 4) knee hyperextension ≥ 10°; 5) palms lay flat on the ground without bending knees. 1 to 4 assesses both sides of the limbs. Illustration adopted from Physiopedia. Hypermobility Syndrome [36].

Data were processed using Microsoft Excel (2019) and IBM SPSS Statistics (v25). Results are presented as mean ± standard deviation or as frequency and percentage with a 95% confidence interval (CI). Significance was set at p < .05. Diagnostic performance of the 5PQ-CN relative to the BS was evaluated by accuracy, sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and receiver operating characteristic (ROC) analysis, with the area under the curve (AUC) interpreted as: excellent (.9–1.0), considerable (.8–.89), fair (.7–.79), poor (.6–0.69), or fail (.5–.59) [37]. For item-specific analysis, chi-square (χ²) tests compared the frequency of positive responses between BS-defined groups, and odds ratios (OR) with 95% CI were calculated. Test–retest reliability was evaluated in the reliability subgroup using Cohen’s kappa (κ) for dichotomous items and the intraclass correlation coefficient (ICC, two-way mixed, absolute agreement) for total scores [38–39]. McNemar’s test was applied to identify response shifts.

The prevalence of GJH as identified by the BS and the 5PQ-CN is shown in Table 3. Based on a BS cutoff of ≥ 5, 308 participants (50.1%) were classified as GJH-positive. Using a 5PQ-CN cutoff of ≥ 2, 347 participants (56.4%) were classified as GJH-positive. A significant difference was observed in the GJH diagnosis between the two methods (χ² = 10.148, p = .001).

(Table 3 HERE)

Title: Table 3. GJH prevalence accessed by both tools

Legend: GJH: generalised joint hypermobility, BS: Beighton score, 5PQ-CN: Chinese version of five-part questionnaire

The diagnostic performance of the 5PQ-CN relative to the BS is summarized in Table 4. Among the 347 participants with a positive 5PQ-CN result, 242 were true positives and 105 were false positives, yielding a PPV of 69.74% (95% CI: 64.89–74.59). Among the 268 participants with a negative 5PQ-CN result, 202 were true negatives and 66 were false negatives, yielding a NPV of 75.37% (95% CI: 70.20–80.54). The overall accuracy of the 5PQ-CN was 72.19% (95% CI: 68.64–75.74), with a sensitivity of 78.57% (95% CI: 74.00–83.14) and a specificity of 65.80% (95% CI: 60.48–71.12). The AUC of ROC was 0.722, indicating moderate diagnostic accuracy for the 5PQ-CN.

(Table 4 HERE)

Title: Table 4. GJH Diagnosis by 5PQ-CN

Legend: GJH: generalised joint hypermobility

Table 5 presents the item-specific distribution of 5PQ-CN responses stratified by the BS diagnosis of GJH. χ² tests revealed a significant difference in the proportion of positive responses between GJH-positive and GJH-negative groups for items 1, 2, 3, and 5 (p < .001), but not for item 4 (p = .547). The highest OR was observed for item 2 (OR = 8.09, 95% CI: 4.86–13.45), followed by item 1 (OR = 5.24, 95% CI: 3.53–7.78) and item 3 (OR = 5.01, 95% CI: 3.55–7.05). Item 5 had the lowest OR (OR = 2.77, 95% CI: 1.74–4.41) among the items showing significant differences.

(Table 5 HERE)

Title: Table 5. 5PQ-CN item-specific distribution grouped by GJH status

Legend: *: p < .001 in χ2, GJH: generalised joint hypermobility, OR: odds ratio, CI: confidence interval

The consistency of the 5PQ-CN diagnoses between the two rounds of administration is shown in Table 6. In the first round, 179 participants (55.1%) were classified as GJH-positive and 146 (44.9%) as GJH-negative. In the second round, 171 (52.6%) were positive and 154 (47.4%) were negative. No significant difference was observed in the distribution of diagnoses or the mean cumulative scores between rounds (p ≥ .05). Cohen’s κ for the dichotomous diagnosis was .703, indicating substantial reliability. The ICC for the total score was 0.707 (95% CI: .649–.758; single-measurement, absolute agreement, two-way mixed-effects model), indicating moderate reliability.

(Table 6 HERE)

Title: Table 6. 5PQ-CN test-retest reliability

Item-specific reliability results are shown in Table 7. Item 3 had the highest proportion of response shifts, with 63 participants (19.4%) altering their response between rounds. This was followed by item 2 (48 participants, 14.8%) and item 1 (46 participants, 14.2%). Item 4 had the lowest number of response changes (18 participants, 5.5%). McNemar’s test indicated a significant difference in the response distribution for items 2 and 3 (p < .001). Cohen’s κ values for individual items ranged from .442 to .664, indicating moderate to substantial reliability.

(Table 7 HERE)

Title: Table 7. Item-specific distribution of 5PQ-CN in both round of test

Legend: *: p < .001 in McNemar’s test, ^: p < .001 for kappa

This study evaluated the validity and reliability of the Chinese version of the 5PQ-CN for screening GJH in young Chinese adults, using the BS with a cutoff of ≥ 5 as the reference standard. To our knowledge, this represents the first validation of the 5PQ in a Chinese population. The principal findings were: (1) a GJH prevalence of 49.9% using the BS and 56.4% using the 5PQ-CN, indicating a statistically significant difference between the two methods; (2) the 5PQ-CN demonstrated moderate validity, with an accuracy of 72.2%, sensitivity of 78.6%, specificity of 65.8%, and an AUC of .722; and (3) test–retest analysis revealed substantial reliability, with a Cohen’s κ of .703 and an ICC of .707.

The observed prevalence of GJH in this cohort is higher than the 10–30% typically reported in Western populations [23, 33], a variation that can be influenced by age, sex, and diagnostic criteria. Several factors may account for this elevated rate. Our cohort comprised young university students, a demographic known to exhibit greater joint laxity, and included a higher proportion of women, who are disproportionately affected by GJH. Ethnic differences may also be a contributor, as prior studies suggest a higher prevalence of GJH in Asian compared to Caucasian populations [2, 40]. Furthermore, a large portion of our sample had received systematic training in sports such as gymnastics and competitive dance, which require exceptional flexibility; this training background is a likely contributor to the high prevalence observed [7, 41].

The diagnostic performance of the 5PQ-CN was comparable to previously validated versions in Swedish and Brazilian Portuguese populations [26, 27]. The sensitivity in our study (78.6%) was higher than that of the Brazilian Portuguese (70.9%) and Swedish (72.3%) versions that used a BS cutoff of ≥ 4, though lower than the Swedish version (91%) that employed age-dependent BS cutoffs. Conversely, our specificity (65.8%) and AUC (.722) were lower than those reported in the other versions. These discrepancies may be attributed to differences in sample characteristics and, importantly, the different BS cutoff points used as the gold standard.

Item-specific analysis revealed a significant difference in response distribution between GJH-positive and GJH-negative participants for all items except item 4, for which few positive responses were recorded. This finding aligns with the prior validation [25], which attributed the low positive rate for item 4 to the non-clinical nature of their sample, where participants rarely present with musculoskeletal symptoms, unlike in the original 5PQ study [32]. In our study, a positive response to item 2 (OR = 8.09) was the strongest predictor of GJH, followed by item 1 (OR = 5.24), item 3 (OR = 5.01), and item 5 (OR = 2.77). Notably, the OR for item 5 in the Swedish version was substantially higher (23.82 and 9.76). Item 5 involves a subjective evaluation of overall joint hypermobility. We hypothesize that the high proportion of participants with a sports training background in our study may have normalized their own flexibility and that of their peers, leading them to underreport this item, whereas the more diverse Swedish sample may have provided greater contrast in subjective perception.

The BS and the 5PQ assess joint mobility with some overlap. While the BS includes 9 assessments (6 in the upper limbs, 2 in the lower limbs, and 1 for the trunk), the 5PQ's items 1 and 2 correspond to the trunk and thumb/wrist assessments of the BS, respectively. Item 4 of the 5PQ also addresses the shoulder joint, and item 3 assesses the hip, which is not evaluated by the BS. We observed contradictory responses between 5PQ item 1 and BS item 9 (spinal flexibility) in 18.5% of participants, and between 5PQ item 2 and BS items 3/4 (thumb/wrist) in 16.3%. For the trunk item, we suspect some participants reported a positive response if they could merely touch the ground with their fingertips, rather than placing their palms flat, as required by the BS. This discrepancy highlights a limitation of the self-reported nature of the 5PQ. For the thumb/wrist item, we noted that participants sometimes attempted the maneuver with the wrist extended in a more difficult position, and did not attempt the alternative, flexed-wrist position if they failed initially.

The test–retest results confirm the stability of the 5PQ-CN, with reliability indices (κ = .703; ICC = .707) consistent with earlier studies [25]. Item-level analysis further supported moderate to substantial agreement (κ: .442–.664), which is slightly lower than the agreement reported for the Brazilian Portuguese (.48–.7) or Swedish (.71–1.0) versions [26, 27].

Although the BS is a short and straightforward clinical tool, it requires trained personnel and is time-consuming to administer sequentially in large-scale epidemiological settings. In contrast, the 5PQ is self-administered and can be distributed in parallel, requiring no additional manpower. The high sensitivity of the 5PQ-CN suggests it is effective for initial screening; however, its moderate specificity indicates that some false positives should be anticipated. This is a relevant consideration for epidemiological studies, where prevalence estimates could be inflated. The AUC of .722 supports the tool's moderate diagnostic utility, indicating that the 5PQ-CN can usefully complement clinician-based assessments like the BS, particularly in large-scale or resource-limited settings. The large sample size and use of the internationally recommended BS cut-off of ≥ 5 strengthen the robustness of this study.

This study has several limitations. The sample consisted exclusively of young university students from a single region in south-western China, limiting the generalisability of the findings to older or more geographically and age-diverse populations. Furthermore, the use of a text-only version of the 5PQ-CN may have caused confusion, as observed with the thumb/wrist and spinal flexibility items.

Future research should explore the validity and reliability of the 5PQ-CN in more age-heterogeneous samples, using age-dependent BS cutoff points. The addition of illustrative diagrams to the questionnaire could enhance participant comprehension. Future studies should also extend validation to clinical cohorts, evaluate the tool's predictive validity for musculoskeletal outcomes, and explore cultural adaptations for broader Chinese populations.

The 5PQ-CN is a valid and reliable tool for screening GJH in young Chinese adults. Compared with the BS, the 5PQ-CN demonstrated moderate diagnostic accuracy, substantial test–retest reliability, and strong feasibility for large-scale application. Although not a replacement for clinician-administered assessments, it offers an efficient and accessible option for epidemiological studies and preliminary screening. Wider use of the 5PQ-CN may facilitate early recognition of GJH, improve understanding of its prevalence in China, and support future clinical and research applications.