2.1 Inclusion and exclusion criteria

Patients in group 1 had a clinical suspicion of a symptomatic rotator cuff tear. A MRI or ultrasound was preoperatively made to classify the lesion (partial or full thickness). If two or more of the following signs were present, a rotator cuff tear was suspected: Neer’s or Hawkins-Kennedy impingement sign, positive Jobe test (supraspinatus), painful arc sign, positive infraspinatus resistance test, positive Neer impingement test (subacromial injection with lidocaine), positive lift-off/belly press test (subscapularis resistance test), positive drop-arm test. Patients in group 2 were diagnosed as having non-ruptured DRC. They had a clinical suspicion of DRC, according to the same signs as described for group 1, however, no rupture of the rotator cuff on MRI or ultrasound. This group included rotator cuff pathology like rotator cuff calcifying tendonitis, tendinosis and bursitis. Patients in group 3 were identified as having shoulder instability. Three or more of the following criteria should have been present: history of sense of instability, previous glenohumeral dislocation, positive relocation test, positive apprehension test, glenohumeral translation, positive sulcus sign, positive hyperabduction test, positive jerktest, labral lesion (MRI), capsular lesion or laxity (MRI). A more detailed description of the clinical tests can be found in the Appendix section of the protocol article². Patients who were eligible for more than one group, e.g., both instability and impingement, were excluded from the study. Patients who had difficulty understanding the Dutch language, had additional shoulder pathology such as adhesive capsulitis, AC-pathology, osteoarthritis or previous surgery on the affected shoulder were also excluded.

2.2 Visits

Patients were seen 3 times for this study. The two pre-treatment visits and a 6 months post-treatment visit were scheduled with an independent, dedicated physiotherapist. During these visits patients were asked to filled out the WORC and their shoulder hindrance. During the first pre-treatment visit only study assessments were performed. The second pre-treatment visit was scheduled 2–3 weeks after the first visit and was used to perform assessments to allow for a test-retest assessment and either conservative treatment was started or the indication for surgery was made.

2.3 Treatment

Patients in group 1 underwent an arthroscopic rotator cuff repair. If during surgery it turned out that the cuff was irreparable due to size, atrophy or fatty degeneration, the patient was perioperatively excluded from the study. Patients in group 2 (DRC) were conservatively treated by injection, physiotherapy or activity adjustment. If conservative treatment failed, arthroscopic subacromial decompression was performed. When calcific tendonitis was present the calcium deposit was removed. In case a cuff tear was found a repair was performed, however, the patient was then excluded from the study. Patients in group 3 (instability) received conservative treatment or surgery (labral or capsular repair).

2.4 Clinical outcome

The WORC is a disease-specific Health Related Quality of Life questionnaire specifically for patients with complaints of the rotator cuff.¹ It consists of 21 items each scored on a 100-mm scale (ranging from 0 best to 100 worst). Items are divided over 5 subdomains: physical symptoms, sports and recreation, work, lifestyle and emotions. Shoulder hindrance was scores on a single item in which patients were asked to rate their shoulder hindrance on an 11-point numeric rating scale from 0 (no hindrance) to 10 points (extreme hindrance).

2.5 Statistics

All data was analysed with the statistical program SPSS (Statistical Package for the Social Sciences, Chicago, IL, Version 22.0). Floor and ceiling effects were studied as the presents of these effects is an indication for limited content validity.³ Floor and ceiling effects were considered to be present when 15% or more achieved respectively the lowest ±10% or highest ±10% possible score and thus patients cannot adequately be distinguished from each other.⁴

2.6 Responsiveness

Responsiveness is a measure of the ability of a questionnaire to detect clinically important changed over time. To determine minimal important changes (MIC) a visual method integrating an anchor-based and a distribution-based approach was used.⁵ This was performed for the total group as well as for the three groups separately. Furthermore, the MIC values were calculated for patients with a low WORC baseline score (WORC < 35) and a high WORC baseline score (WORC ≥ 35).

Change scores for the WORC were calculated by subtracting pre-treatment scores (first assessment) from post-treatment scores. Consequently, positive change scores indicate an improvement in WORC, while negative scores indicate deterioration. The shoulder hindrance score was used as an external criterion to distinguish patients who improved from those who did not. Patients were categorized into three groups according to the shoulder hindrance score’: (1) importantly deteriorated (increase in shoulder hindrance score’ >2); (2) not importantly changed; and (3) importantly improved (a decline in shoulder hindrance score’ >2). The limit of more than 2 points was chosen based on a previous study.⁶ Correlation (Spearman’s rho) between the anchor (shoulder hindrance) and WORC change score was calculated, as an adequate anchor should at least have a correlation of 0.50.⁵ The distribution of the change in WORC score was plotted and two cut-off points of the MIC were determined: the Receiver Operating Characteristic (ROC) cut-off point and the 95% limit cut-off point.⁵ The area under the curve (AUC) was also determined. An AUC between 0.7 and 0.8 was considered as acceptable, between 0.8 and 0.9 as excellent and larger than 0.9 was considered outstanding.⁷

2.7 Minimal detectable change

The Standard Error of Measurement (SEMagreement) was based on the test-retest data and used to calculate the minimal detectable change (MDC), defined as 1.96 * √2 * SEM.⁸ The MDC can be explained as the amount of change below which there is more than a 95% chance that no actual change has occurred. To make comparisons possible between the results of this study and the literature, the Standardized Response Mean (SRM) is presented as well. The SRM was calculated as the mean change in WORC score divided by the standard deviation of the change. An SRM of 0.2 was defined as small effect, 0.5 as moderate effects and 0.8 as large effects.⁹

3.1 Patients

Of the 111 consecutive patients, six patients were not eligible: 2 patients had osteoarthritis of the glenohumeral joint, 1 patient had osteoarthritis of the AC-joint, 1 patient had an irreparable cuff lesion, 1 patient had multiple shoulder pathologies and 1 patient did not want to participate. This resulted in a total of 105 patients of which 37 in group 1, 38 in group 2 and 30 in group 3. Fourteen patients were excluded during the study or were lost to follow-up: 6 patients from the surgical group preferred conservative treatment, 1 patient transferred from group 1 to group 2 because there was no cuff lesion during surgery, 1 patient had not complied with the study visits, 1 patient was free of symptoms before start of treatment and 6 patients were lost to follow-up without apparent reasons. Therefore, 90 patients were included in the analysis (Table 1).

3.2 Interventions

In all patients of group 1, an arthroscopic rotator cuff repair was performed. The number of torn tendons varied; 1 tendon in 20 (67%) patients, 2 tendons in 3 (10%) patients, 3 tendons in 6 (20%) patients and 4 tendons in 1 (3%) patient. In all patients the supraspinatus tendon was involved. Of group 2 in 9 patients an arthroscopic decompression was performed, 9 received physiotherapy, 5 received injection with lidocaine and cortisone, 10 had both physiotherapy and injection, and 1 patient only got advise how to prevent overuse. In 15 patients of group 3 an arthroscopic stabilization was performed by Bankart repair, and in 2 by a combination of Bankart repair and remplissage. Nine patients of group 3 received physiotherapy.

3.3 Clinical scores

In Table 2, the mean values and standard deviations of the WORC and shoulder hindrance score are presented of the two baseline visits, the post-treatment visit and the difference between the first baseline visit and post-treatment visit. Following a cuff repair (group 1), patients had an average increase of 41.1 points on the WORC. DRC and instability patients improved respectively 26.0 and 33.8 points on the WORC. Shoulder hindrance decreased between 3 and 5 points (Table 2).

At baseline there were no floor- or ceiling effects. However, 6 months post-treatment there were considerable ceiling effects for the WORC total score and all WORC subdomains (Table 3). The largest ceiling effects were seen in the instability group varying from 52% till 80%. For the rotator cuff repair and DRC patients, the ceiling effects ranged respectively from 23% till 43% and 22% till 56%.

3.4 Responsiveness

Spearman’s rho correlation between the WORC change score of the total group and the shoulder hindrance score (anchor) was 0.55 (p < 0.001). Correlations of group 1, 2 and 3 separately were respectively r = 0.51 (p = 0.004), r = 0.55 (p = 0.001) and r = 0.43 (p = 0.033).

Only one patient (DRC) fell in the importantly deteriorated category. It was therefore not possible to determine the MIC for deterioration and this patient was excluded from further analysis. For all three groups, the MIC values were larger than the minimal detectable change (margin of error). MDC was smallest for patients undergoing a rotator cuff repair (Table 4). The 95% limit cut-off resulted in larger MIC values than the ROC cut-off. Therefore, using the 95% limit cut-off method, more change is needed before it can be considered a clinical important change. The AUC was outstanding for the cuff repair group, excellent for the DRC group and acceptable for the instability group. All SRM values were above 0.8 indicating large effects.

In Fig. 1 the distribution of the patients in the ‘no important change’ and ‘important change’ groups are plotted against the change in WORC score. The 2 MIC values are also presented in the graphs. These figures illustrate that patients that importantly improved on the anchor, showed more improvement in WORC score. The farther the peaks of the two groups are apart, the better the responsiveness of the outcome measure.

Close

Fig. 1 Distribution of the patients in the ‘no important change’ and ‘important change’ groups expressed in percentage of changes in scores on the WORC for the total group and the three group separately. The ROC cut-off and 95% limit cut-off values are also presented in the graphs.

Export to PPT

Despite the small sample sizes an attempt was made to distinguish between patients with a low and a high WORC at baseline (Table 5). The MIC values of the rotator cuff repair group with low baseline scores could not be calculated, because 6 months post-treatment, all of the patients of this group were importantly changed. The same accounts for the instability patients with a low baseline. Only one patient was not importantly changed and thus the 95% limit cut-off which used the standard deviation of the mean change could not be calculated. In general, the MIC values for patients with a high baseline are smaller than for patients with a low baseline, meaning that patients with a low baseline need to change more before it is considered clinically important change.

4.1 Minimal important change

Minimal Important Change (MIC) was calculated using multiple methods. Ideally the results of the different methods lay close to each other in which case the choice of method has no effect on outcome.⁵ This is the case in the rotator cuff repair group (34.0 vs 35.3) but not in the DRC group (22.9 vs 41.9) or instability group (31.8 vs 46.0). If there is a large difference in MIC, the largest value should be chosen as MIC.⁵ Therefore, patients undergoing a rotator cuff repair need to improve more than 35 points on the WORC to be considered clinical importantly improved. DRC patients and instability patients need to show substantially more improvement on the WORC score to be considered clinical importantly improved, namely 42 and 46 points.

4.2 Baseline impairment

One can imagine that an individual MIC can depend on the impairment at baseline. For this reason a sub analysis was performed calculating the MIC in which patients were divided based on a baseline WORC score <35 and ≥35. Not all calculations could be made because no “not clinically changed” patients existed in the cuff repair group with low baseline score. Considering the DRC group and the instability group, there was an increase in MIC values for patients with a low baseline WORC score. Therefore, patients with a low WORC score at baseline need a larger improvement in WORC score to make it clinically important, although we should be reticent because of the low numbers.

4.3 Comparison of literature

Determining the responsiveness of the WORC has been done before, however, not in this specific patient group of rotator cuff repair. In other studies the responsiveness of the WORC was tested on patients with some kind of DRC^11–14; a specific group with patients undergoing rotator cuff repair was lacking. Another difference with existing studies is the way that responsiveness was measured. In most studies the responsiveness of the WORC was investigated by calculating effect size (ES) and standardized response mean (SRM) which are distribution-based approaches.^11–14 Disadvantage of such an approach is that these are measures of the magnitude of the change score and they do not provide a good indication of the importance of the observed change. An anchor-based approach as external criterion is therefore promoted to determine important improvement or deterioration. The disadvantage of solely an anchor-based approach is that it does not take into account the variability of the instrument. Therefore a visual method in which an anchor-based MIC distribution method is used, combining the two approaches, is recommended and also used in this current study.⁵

Another study in which the two approaches were combined is the study of Ekeberg and colleagues.¹⁵ They compared the responsiveness of the WORC with the Shoulder Pain and Disability Index (SPADI) and Oxford Shoulder Score (OSS) in patients with DRC. Patients were included based on physical examination and treated by injections. They concluded that in that patient group the WORC was not more responsive compared to the SPADI. We did not expect that conclusion since the WORC is developed for a more specific patient group than the SPADI. For that reason, we also calculated the SRM in order to compare with Ekeberg et al., and found a SRM in the improved DRC group of 1.82 which is comparable with Ekeberg et al. (SRM 1.69). However in the cuff repair group we found a SRM of 3.05, which is substantially higher. Therefore we are convinced that their conclusion considering the responsiveness of the WORC compared to the SPADI does not account for patients undergoing rotator cuff repair.