Cross-cultural adaptation and measurement properties of the Malay Shoulder Pain and Disability Index

Objective The purpose of this study is to cross-culturally adapt the Shoulder Pain and Disability Index from English to Malay, and to evaluate the measurement properties of the Malay version among Malay speakers with shoulder pain. Methods Cross-cultural adaptation of the Malay version of Shoulder Pain and Disability Index (M-SPADI) was conducted according to international guidelines. 260 participants (Shoulder pain = 130, No shoulder pain = 130) completed the M-SPADI, the Numerical Rating Scale (NRS), and measurement of shoulder active range of motion (AROM). 54 participants repeated M-SPADI within a mean of 9.2 days. Results Cross-cultural adaptation of M-SPADI had no major issues. The M-SPADI had good face validity; item and scale content validity indexes (I-CVI, S-CVI) were >0.79 except for Disability Item 3 (I-CVI = 0.75), and exploratory factor analysis showed that M-SPADI had a bidimensional structure. There was a strong positive correlation between M-SPADI and NRS (rPain = 0.845, rDisability = 0.722, rTotal = 0.795, p <0.001) and a negative correlation between M-SPADI and shoulder AROM with the following correlation ranges (rPain = -0.316 to -0.637, rDisability = -0.419 to -0.708, rTotal = -0404 to -0.697, p<0.001). M-SPADI’s total score was higher in participants with shoulder pain (Mdn: 33.8, IQR = 37.3) compared to no shoulder pain (Mdn:0, IQR = 0.8) and the difference was statistically significant (U = 238.5, z = -13.89, p<0.001). M-SPADI had no floor or ceiling effects (floor/ceiling <15%), high internal consistency (Cronbach’s αPain = 0.914, Cronbach’s αDisability = 0.945) and good to excellent test-retest reliability (ICCPain = 0.922, ICCDisability = 0.859, ICCTotal = 0.895). Conclusion M-SPADI has a bi-dimensional structure with no floor or ceiling effects, established face, content and construct validity, internal consistency, and test-retest reliability. M-SPADI is a reliable and valid tool for assessing Malay-speaking individuals with shoulder pain in clinical and research settings.

Objective: 23 The purpose of this study is to cross-culturally adapt the Malay version of the Shoulder Pain 24 and Disability Index (M-SPADI) and evaluate its measurement properties among Malay 25 speakers with shoulder pain.

26
Methods: 27 Cross-cultural adaptation of M-SPADI was conducted according to international guidelines.    M-SPADI's total score was higher in participants with shoulder pain (  Shoulder pain is a common musculoskeletal disorder with a lifetime prevalence of 7 to 67%

62
(1, 2). Associated symptoms include restricted shoulder motion, disturbed sleep, and impaired 63 activities of daily living (3)(4)(5). Globally, it causes work absence, disability, and increased 64 healthcare costs (2,6). In Malaysia, shoulder injury is ranked third in musculoskeletal disorders 65 causing disability and fourth in the total cost of workers' compensation claims per body part 66 (7).

67
Health-related patient-reported outcome measures (PROM) are essential to patient-centred care 68 and research (8). decline to give consent, and psychiatric illness. 106 We calculated a sample size of 50 participants for the pilot study and 60 participants for test-107 retest (27,30). Based on a participant to item ratio of 10: 1, the validation study sample size 108 was 130 participants (28,30). To assess known group validity, another 130 participants with 109 no shoulder pain were recruited (28,30). Coltman et al., the author considered SPADI a reflective construct (31,32).

118
Each subscale score is calculated using the formula: 9).

120
The maximal possible score excludes any unmarked item but requires at least 3/5 pain items 121 and 6/8 disability items answered for SPADI to be scored (10). The total SPADI score is the 122 unweighted mean of pain and disability domain scores (9). The scores range from 0= the best 123 to 100= the worst with no cut-off point to indicate severity as it was designed to measure current 124 status and change over time (1,9). The Numerical Rating Scale (NRS) 127 The NRS is an instrument for pain intensity assessment where individuals are asked to select a 128 number from 0 to 10 that best describes their pain intensity (33). The anchors are zero for no 129 pain and ten for the worst pain ever possible (33).  160 We submitted all documentation and a report of the adaptation process to the original author.  and factor loading ≥0.5 was considered significant(39).

196
As there is no gold standard shoulder-specific questionnaire in the Malay language, criterion 197 validity could not be tested (28). Hypotheses testing for construct validity was assessed using 198 convergent validity and known-group validity (28,29). Spearman's correlation was used to  Test-retest reliability was assessed using the intraclass correlation coefficient (ICC) with a 95% 213 confidence interval based on average measurement, absolute agreement, 2-way mixed-effects 214 model (29,41,42). The selected interval between repeated measures was seven days to prevent 215 recall but ensure no clinical change had occurred (29).    Table   264 3.

313
Overall Mann-Whitney U test revealed that M-SPADI pain scores, disability score, and total 314 score were significantly higher in the shoulder pain group compared to the no shoulder pain 315 group and the difference was statistically significant with a large effect size (Table 5). Less than 15% of participants with shoulder pain achieved the lowest or highest possible scores 324 in the pain subscale (floor=0%, ceiling=0%), disability subscale (floor=1.5%, ceiling=0.8%), 325 or total score (floor=0%, ceiling=0%) ( Table 6).  (Table 6). The cross-cultural adaptation of M-SPADI adhered strictly to recommended guidelines (28, 342 29). During the expert committee review, there were two main issues. The first was Disability 343 Item 3: "putting on an undershirt or jumper," which scored I-CVI=0.75. This item had issues 344 with the word undershirt and jumper, thus requiring multiple amendments and a trial in the 345 pilot study before being accepted by the expert committee. This was a similar issue faced by 346 the Brazilian-Portuguese study, which substituted the word "jumper" for "T-shirt" followed by MacDermid et al. and may suggest that respondents cannot distinguish between pain and 365 disability in some functional items as these two factors are closely related (44,45).

366
Other studies which had similar EFA results of two factors with some items cross-loading were  MeanTamil:52.60) (16,17). Mean scores of AROM were also much higher in the M-SPADI 396 study than the Telugu SPADI and Tamil SPADI study (Example: AbductionMalay=134.61°, 397 AbductionTelugu=83.00°, AbductionTamil =103°) (16,17).  Known-group validity testing demonstrated that the M-SPADI subscale and total scores were 408 higher in the group with shoulder pain compared to the group without shoulder pain, and the 409 difference was statistically significant. This confirmed that M-SPADI could discriminate 410 between different groups, in this case, participants with and without shoulder pain. Other 411 studies with reported known-group validity for SPADI include Dutch SPADI, Slovene SPADI, 412 and Danish SPADI. They compared high and low initial pain scores (23), work absence versus 413 no work absence (23), different severities of self-reported perceived disability (22), and 414 working versus non-working participants (20). 415 Cronbach α for total score was not calculated as our EFA concluded that M-SPADI is 416 bidimensional (26,28). Instead, we reported Cronbach α for each unidimensional domain: the 417 pain subscale (Cronbach's α=0.914) and the disability subscale (Cronbach's α=0.945). These 418 results are comparable to those reported in the original SPADI and other SPADI translations 419 (1, 4,13,14,16,18). These results show that the M-SPADI has a high internal consistency and 420 does not have item redundancy as it did not cross the Cronbach's α>0.95 threshold.

421
The test-retest reliability for M-SPADI was rated good to excellent (ICCPain=0.922,

424
Original SPADI findings could be due to its small sample size of 37 males. A recent systematic 425 review reports test-retest reliability of SPADI ranges from ICC=0.850-0.922, reflecting the 426 findings of this study (9).

427
The strengths of this study were its large sample size fulfilling the COSMIN guidelines 428 requirements (26,28). Moreover, this study adhered to the standard methods and guidelines for 429 all procedures. In addition, this study includes known-group validity comparing participants 430 with and without shoulder pain which has not been performed before.

431
The limitation of this study was that it was conducted in a single urban centre, with 75% of 432 participants having tertiary education, which may cause bias in the results. A multicentre study 433 conducted in rural and urban settings with participants of different educational backgrounds 434 could yield different results. We also noted that the mean time interval for the test-retest was 435 9.2 ± 3.8 days, which was longer than the recommended seven days (9). Like the Danish SPADI 436 study, the M-SPADI results did not seem to be affected by this prolonged interval, and the 437 results were in keeping with other SPADI translation studies (13,14,(16)(17)(18)(20)(21)(22). 438 We recommend future prospective studies for the M-SPADI to examine measurement