The study formed part of a cross-sectional study into pain and osteoarthritis that was approved by the Nottingham Research Ethics Committee (Reference No: K13022014). Participants were recruited from a mixed population of local community-derived participants using a study advertisement. Implied consent to participate was obtained from all participants completing the electronic self-report instrument and written consent from those who attended for clinical assessment.
Development of the self-report instrument
A self-report instrument consisting of five line drawings (left and right elbow extension, knee extension, little finger extension, thumb extension, and trunk flexion) was created to depict the 9-point Beighton score criteria. Version one of the instrument depicted one degree of severity (positive test), and version two depicted two degrees of severity (positive and negative test) for each item in the Beighton score. The instrument including both the instructions and the line drawings underwent pilot testing and a review by a Patient Public Involvement (PPI) panel at Nottingham University Hospitals NHS Trust. The first 30 participants who agreed to take part were enrolled into the pilot study, and were invited to complete version one (n = 15) or version two (n = 15) of the instrument. The PPI panel consisted of six local residents who responded to a study advertisement. PPI members reviewed and completed the instrument and gave feedback during a focus group interview that was digitally recorded. Interviews were listened to and salient points were transcribed verbatim. Two sets of field notes were cross-referenced with the verbal recording and suitable recommendations were fed back into the redesign of the instrument. These amendments were verified by two research assistants following a review of their field notes, and by returning to the PPI members once more for verification. Four of the five line drawings (elbow extension, knee extension, little finger extension, and trunk flexion) were reconfigured to each include three intervals. The remaining item (thumb opposition) consists of two intervals.
The first line drawing was created to depict forward flexion of the trunk, with the knees straight, so that the palms of the hands rest flat on the floor. During early pilot testing, it became evident that the use of one and two gradations for trunk flexion yielded a number of false positive test results. PPI members explained that it was unclear if a positive test constituted placing their fingertips on the floor. The line drawing was reconfigured to include three intervals that illustrated: a) the inability to touch the floor (negative test result); b) being able to touch the floor with the fingertips (negative test result); and c) being able to place the palms of the hands flat on the floor (positive test result). A modified version of trunk flexion was also created to enable those who are unable to perform the manoeuver in standing, the opportunity to do so in sitting (see Appendix: Figure 1).
The second line drawing depicts extension of the tibiofemoral joint beyond −10 degrees. From a lateral view, a straight aligned knee was drawn with one interval (version one) and two interval changes (version two) of eleven degrees either side to illustrate knee-flexion and knee-extension. During early pilot work, it was noted that participants experienced difficulty in distinguishing between knee-extension range of movement (ROM) of 0–10 degrees (a negative test result), and knee-extension ROM beyond −10 degrees (a positive test result). Thus, three knee intervals were incorporated into the final instrument and two intervals were increased to twenty degrees to help distinguish between them more clearly. The line drawing was reconfigured to illustrate: a) knee-extension ROM of −20 degrees or greater (positive test result); b) knee-extension ROM of 0–19 degrees (negative test result); and c) knee-flexion ROM of 1 degree or greater (negative test result). Following PPI review, a red line was also drawn on the knee line drawing and the subsequent line drawings (thumb, elbow, and little finger) to illustrate the angle at the joints more clearly (see Appendix: Figure 2).
The third line drawing was created to illustrate the ability to passively extend the thumb and flex the wrist, so that the distal phalanx of the thumb can touch the distal radial side of the adjacent forearm. Following recommendations by PPI members, a line drawing consisting of two intervals depicting: a) the thumb touches the forearm (a positive test); and b) the thumb is unable to touch the forearm (a negative test) was incorporated into the final instrument (see Appendix: Figure 3).
The fourth line drawing illustrates the ability to extend the elbow joint beyond −10 degrees. A single line drawing was created with the elbow in a plane of 0 degrees of extension. Following pilot testing, two further line drawings were created with 11 degrees intervals in either direction. These two intervals were subsequently increased to 15 degrees to help distinguish between them more clearly. Thus, the line drawing consists of three intervals with varying degrees of elbow flexion-extension and is dichotomised into three outcomes: a) elbow flexion (a negative test), b) the elbow in a neutral plane (a negative test), and c) elbow extension (a positive test) (see Appendix: Figure 4).
The fifth line drawing was designed to replicate passive extension of the little finger beyond 90 degrees. Pilot testing revealed that line drawings incorporating one and two gradations yielded a high number of false negative test results. PPI members explained it was unclear if extending the little finger to 90 degrees also constituted a positive test. Thus, the final line drawing consists of three intervals to depict: a) the little finger extending beyond 90 degrees (positive test result), b) the little finger extending equal to 90 degrees (negative test result), and c) the little finger extending less than 90 degrees (negative test result) (see Appendix: Figure 5). Each item in the self-report instrument was accompanied by a set of instructions communicating to the participant how the line drawings should be used to determine GJH.
Validity and reliability of the self-report instrument
The final instrument was validated in 50 participants who twice completed the self-reported instrument online, a fortnight apart. Participants then subsequently attended for a first clinical assessment, and then again one week later for a second clinical assessment. The results of the previous assessment were not made available to the participants. The purpose of the clinical assessment was to determine the level of participant-observer agreement. One examiner completed the self-report instrument using a scoring card and goniometry measurements of the elbow, knee, and little finger were taken using standard practice guidelines [23]. Trunk flexion was assessed using a modified fingertip-to-floor distance (FFD) measurement. Goniometric measurements of the elbow and knee have been shown to have excellent reliability [24, 25]. The FFD has been shown to have excellent inter-test reliability of lumbar spine flexion [26], although it was adjusted to include the participants attempting to place their palms of the hands flat on the floor. So as to ensure that the observer was blinded to the results of the participants, the online self-report data was downloaded only after the clinical assessments had taken place. A second observer, who was blinded to the participants’ results, scored the self-report instrument using data from the observer’s scorecard. Data from each item in the self-report form was reduced by the second independent assessor to either a ‘positive’ or ‘negative’ result for joint hypermobility (JH) based on the Beighton criteria [6]. GJH was measured using a threshold cut-point of 4 of 9 to categorise participants as hypermobile, in line with previous studies on GJH. The results of the previous mechanical goniometry assessment were not made available to the assessor, or to the participants. There was no self-reported change in health status between the distributions of the two self-report forms or clinical assessments.
Statistical analysis
Validity of the self-report instrument was assessed by calculating the sensitivity, specificity, and the participant-observer agreement. Standard two-by-two tables were formulated to calculate sensitivity and specificity and their 95% confidence intervals (CIs). The Cohen’s unweighted kappa statistic (k) and its 95% CIs were used to calculate repeated measures between the participants and the observer – the reference standard. Reliability of the instrument was assessed by Cohen’s unweighted kappa statistics (k) (95% CI) for participant-repeatability (n = 50), and observer-repeatability (n = 50), and participant-observer agreement (n = 50) at two weekly intervals. Cohen’s kappa statistics were interpreted as follows: < 0 = poor, 0.01–0.20 = slight, 0.21–0.40 = fair, 0.41–0.60 = moderate, 0.61–0.80 = substantial, and 0.81–1 = almost perfect [27].
Validity and reliability were assessed for each item in the self-reported JH instrument separately, and for the sum of the total scores. To calculate these outcomes, the JH grade for each item of the instrument was dichotomised as positive (hypermobile) or negative (non-hypermobile) by classifying the most severe grade as present, specifically, category C for elbow extension, category A for extension of the little finger, knee and thumb, and category C and F for trunk flexion in standing or sitting, respectively (see Appendix: Figures 1-5). Analyses were performed using SPSS software version 22.0 and the 95% confidence intervals for Cohen’s kappa were calculated manually.