Design and setting
This study used a cross-sectional study design with convenience samples and conducted from September 2017 to February 2019. A two-phase process was used in this study to evaluate psychometric properties. Phase 1 was a linguistic validation. Phase 2 was a psychometric validation. Written permission was granted to the translation by one of the original authors of the WORC.
Procedure for evaluating psychometric properties
Linguistic validation
The T-WORC was translated according to the original author’s recommendations [23]. Forward translation from the English version of the WORC to the T-WORC was done by two independent individuals. One translator was a doctoral student of nursing and the other translator was a physician of orthopaedics, and then two forward translators and researchers had a discussion, by exchanging and discussing, a group consensus was reached and "forward translation" was established. In order to assess conceptual equivalence with the original version. The T-WORC was translated back into English by an English teacher who is bilingual, and who never saw the original version of the questionnaire. The semantic equivalence of the original version of WORC and translated T-WORC was compared by a native English-speaking professional of nursing. Then "backward translation" was completed. Finally, an expert committee consisted of five professionals (a doctoral nursing student, a nursing teacher, a nurse practitioner, a nursing supervisor, and an attending physician) reviewed each item of T-WORC, and rated on a 4-point Likert scale of each item for meaning clarity, understanding, and wording. If any item was rated as 2 or below it was retranslated, if it was rated as three, it would be revised according to the experts' suggestions. Among the 21 items, item 9 was rated by experts as 3 in ‘‘understanding’’ ‘‘how much difficulty do you have with someone or something coming in contact with your affected shoulder?’’, the answer ranged from ‘‘no fear’’ to ‘‘extremely fear’’. According to the expert committee's suggestion and the original author's consent, we replaced ‘‘fear’’ with ‘‘difficult’’ to answer the question of item 9 of the T-WORC. Then the final version of T-WORC was completed and adopted for the study.
Participants
All patients with the disease of rotator cuff tears were recruited from a teaching hospital in northern Taiwan. The inclusion criteria were: (a) aged 20 years or older; (b) the diagnosis was a rotator cuff tear and waiting for surgery; (c) able to speak Mandarin or Taiwanese language and have the reading ability at the elementary school level. The exclusion criteria were (a) the patient who had clavicle, scapula, and humeral fractures; (b) cognitive impairment; (c) and rheumatic disease.
The data were collected from inpatient and outpatient departments. In the inpatient department, while the patient admitted to the ward, the first author who then contacted them directly and explained the purpose of this research. If patients met the screening criteria and agreed to participate, they were requested to complete a written consent form and asked to fill in the questionnaires. Thirty patients were measured at two different time points for testing the test-retest reliability. Time 1 was in the outpatient department and time 2 was 2 weeks later. Patients diagnosed with RCT were referred by the physician to a well-trained research assistant and informed about the study. T2 (2-week later), the patients needed to complete the T-WORC questionnaire at home, and returned the questionnaire to the researcher by mail, or brought it to the researcher when they were hospitalized. The 2-week interval for assessing test-retest reliability was chosen because the patients unlikely to change much during the 2 weeks while they were waiting for surgery, and it was long enough for patients to forget the previous experience of answering.
Sample size
Based on the Polit & Beck’s (2012) instruction calculated sample size, 10 participants per each item, it totally need a sample size of at least 210 [24]. A total of 218 patients met the sampling criteria, eight of whom refused to participate because of loss of interest. Finally, a total of 210 patients were recruited. The refusal rate was 3.7%. The sample power was .89 with sample size of 210, α was set at .05 and effect size was set at .2.
Measurements
Six instruments were used in this study including demographic and disease characteristics form, the T-WORC, the Quick Disabilities of the Arm, Shoulder, and Hand (QuickDASH), the Short Form-12 (SF-12), the State Anger Scale (SAS), and the Numerical Rating Scales (NRS) of pain.
Participants demographic and disease characteristics
The RCT patient’s demographic and disease details included age, gender, marital status, education, employment, affected side, dominant side, and pain duration.
Western Ontario Rotator Cuff index (WORC)
The WORC was developed to measure disease-related QoL of patients who suffer from RCTs disorders [15]. It has 21 items with five domains. Each item in WORC has a possible score from 0-100-mm visual analogue scale (VAS). The maximum score of WORC is 2100, and the minimum score is 0, the higher the score, the worse the QoL. The total score can also be converted into a percentage score, range from 0% to 100%. The total score is subtracted from 2100 and divided by 2100 x 100 for the % [15]. The WORC demonstrated good internal consistency reliability (Cronbach’s α = .98) [21], and test-retest reliability with Intra-class correlation coefficient (ICC) of .96 [15]. The Chinese version’s Cronbach’s α was .95 and the ICC was .96 [25].
Quick Disabilities of the Arm, Shoulder, and Hand (QuickDASH)
The 11-item QuickDASH is a shortened version of the 30-item Disabilities of the Arm, Shoulder and Hand [26]. It measures symptoms and physical function in patients with disorders of the upper extremity including total shoulder arthroplasty, rotator cuff tendinopathy, carpal tunnel syndrome, wrist/hand tendinitis. Each item scored 0-5, transforming score from 0 (no disability) to 100 (most severe disability). A higher score indicates greater disability. The QuickDASH has good internal consistency reliability (α = .92) [26]. In this study, Cronbach’s α was .896.
Short Form-12 (SF-12)
The SF-12 is an instrument for measuring the QoL. There are 12 items, which is a subset of the SF-36 health survey and resembles the SF -36 in the constructs and interpretation of scores. The psychometric properties were able to reproduce more than 90% of the variance in the physical and the mental subscales of the SF-36, and reproduced the profile of the eight dimensions of the SF-36. The SF-12 includes two main domains, the physical component summary (PCS) and the mental component summary (MCS), and eight subscales for assessing eight dimensions: physical functioning, role physical, social functioning, role- emotional, bodily pain, general health, vitality, and mental health [27]. We used the QualityMetric Health Outcomes Scoring Software 2 to calculate the PCS and the MCS scores. Scale scores were transformed into 0 (the worst) to 100 (the best). The higher the score indicates better the QoL. The SF-12 demonstrated good internal consistency reliability, Cronbach’s α was .91 for the elderly population in Chinese [28]. In this study, Cronbach’s α was .91.
State Anger Scale (SAS)
The SAS was developed by Chiu [29]. The questionnaire consists of 10 items. Scoring for each item ranges from 1 (not compliance with the emotional state of the moment at all) to 7 (full compliance with the emotional state of the moment). Higher scores indicate higher levels of anger. Cronbach’s α was .607 [29]. In this study, Cronbach’s α was .953.
Numerical Rating Scales (NRS) of pain
In this study, we wanted to know how well the T-WORC discriminates between subgroups of the study sample that differed in their pain scores. Overall shoulder pain was assessed according to an 11-point NRS. The NRS of pain was highly correlated with the VAS. It is a valid and reliable pain assessment tool. The NRS of pain scale range from 0 (no pain) to 10 (worst pain). A higher score indicates more severe pain [30].
Psychometric validation
Item analysis
Three methods were used for item analysis of 21-item of the T-WORC, including average inter-item correlations, item-scale correlations, and critical ratio.
Reliability
The reliability was assessed for internal consistency and test–retest reliability.
Validity
Construct validity was evaluated by examining convergent, divergent, known-group validity, and Exploratory Factor Analysis (EFA).
Convergent validity, Divergent validity, and Known-group validity
Convergent validity refers to the degree to which two measures of constructs that theoretically should be related are, indeed related [31]. Convergent validity was examined by hypothesis testing of the relations between T-WORC and QuickDASH, SF-12, and NRS of pain. Based on literature review that we hypothesized, the T-WORC would highly positively correlate with QuickDASH and NRS of pain; T-WORC would highly negatively correlate with SF-12 [15, 19, 21, 25]. Divergent validity refers to the degree to which two measures of constructs that theoretically should be unrelated are, in fact not related [31] or less related [32]. Specific disease-related QoL and anger are two different concepts. So, we hypothesized that the T-WORC would be less correlated with SAS. Known-group validity means that the instrument is administered to two or more groups that are expected to differ on the critical attribute due to known characteristics [24]. High pain scores and low pain scores of RCT patients are two different characteristics, so known-group validity was examined by comparing patients who had higher pain scores (upper 73%) with those who had lower pain scores (lower 27%) [33]. We hypothesized that the T-WORC score were significantly different between the high pain scores group and the low pain scores group.
Factorial Validity
An EFA was performed on the 21-item T-WORC using principal component analysis with oblique rotation (because all five subscales were correlated) to determine the factor structure of all the items. Additionally, the Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy and Bartlett’s test of sphericity were administered to check the suitability of data for factor analysis. The number of factors extracted was determined based on an eigenvalue > 1.0 and the scree plot. Factor loadings with an absolute value of .40 and above are included [24].
Ethical consideration
This was part of a large study. Ethical approval was obtained from the institutional review board of the hospital (2016-09-010B#1). All patients were fully informed and their consent were obtained before the study implementation. This study has no conflict of interests.
Data analysis
All statistical analyses were conducted with the Statistical Package for the Social Sciences for Windows, version 22.0, the significance of the p-value was set at a level of .05. Descriptive statistics, including means, standard deviations, and percentages were calculated for participants’ demographic and disease, the T-WORC, the QuickDASH, the SF-12, the SAS, and NRS of pain. The inter-item correlations between .30 and.70, and the corrected item-scale correlations of greater than 0.5 were considered appropriate to measure the target construct [24], and independent t-test was used for testing the critical ratio. The Cronbach’s α coefficients were calculated to estimate internal consistency. Test-retest reliability was calculated by intraclass correlation cofficients (ICCs). The value for Cronbach’s alpha of .75 or higher is considered acceptable for the instrument’s internal consistency reliability, and ICC > .75 indicated good test–retest reliability and stability [34]. To test for convergent and divergent validity, we calculated Pearson correlations between the T-WORC and the QuickDASH, the NRS of pain, the SF-12, and SAS. Known-groups validity was tested by comparing pain scores of high and low, an independent t-test was used for the analysis [33].