Shoulder pain is a very common condition in the general population and may affect up to 66.7% of the population over a lifetime.1 Rotator cuff related shoulder pain, which includes rotator cuff tears, is one of the most frequent diagnoses.2–4 Current evidence shows that surgical approaches are not superior to conservative approaches, and that exercise-based treatment should be considered as a first-line option.5,6

The effectiveness of exercise-based interventions in individuals with rotator cuff tears has been previously reported.5–7 It is therefore timely to assess the description and implementation fidelity of clinical trials investigating exercise-based interventions for rotator cuff tears. Incomplete reporting of interventions may prevent clinicians and researchers to promptly replicate these interventions in clinical practice and future trials.

The Template for Intervention Description and Replication (TIDieR) can be used to improve the description of interventions and their implementation. The TIDieR checklist was created in 2014 to improve reporting of interventions in clinical trials.8,9 This checklist contains 12 items that cover information about how a treatment is delivered, the equipment and the level of training needed, and the dose and intensity used in the program, among others.10 In addition to providing a detailed description of an intervention, implementation fidelity of the intervention is another aspect that may improve reliability and internal validity of the study, as it involves the knowledge of the fidelity between what was prescribed compared to what was performed as an intervention in the trial.11 If studies do not provide enough and adequate details about the intervention, the findings may be under or overestimated, resulting in misleading interpretations and hindering reproducibility for researchers and clinicians.12 The National Institutes of Health Behaviour Change Consortium (NIHBCC)13 was created to assess the implementation fidelity and improve the clarity and quality of reporting interventions. This tool evaluates the full treatment implementation model that consists of treatment delivery, receipt, and enactment.14,15 Ideally, clinicians and researchers should be able to replicate evidence-based exercises to treat individuals with shoulder pain and rotator cuff tear. Therefore, it is important that studies not only have low risk of bias, but also provide information about materials, infrastructure, dosage, possible modifications made in the treatment, and appropriate implementation fidelity.

The objective of this study was to assess the description and implementation fidelity of exercise-based interventions in clinical trials for individuals with rotator cuff tears. We expected to highlight possible gaps for clinicians and researchers.

Our database search identified 4400 articles. Thirteen studies fulfilled the criteria for inclusion with a total of 900 participants. A flow chart with the selection process is shown in Fig. 1. The characteristics of the included studies are provided in Table 1. Two follow-up studies reporting secondary analyses for different time points from the same trial in separate publications,26–29 were merged and presented in this review as a single study.30,31

Fig. 1.

Flow chart of the selection process.

(0.39MB).

Among the included studies, the oldest article was published in 2010, and the most recent ones were published in 2021. Most studies included individuals with partial-thickness tears rather than full-thickness rotator cuff tears, and the mechanism of injury varied but was mostly not reported. Several approaches were used in the management of individuals with rotator cuff tears. In addition to exercises, the treatment also involved injection therapy, manual therapy, electrotherapy, and shoulder surgical procedures. The outcome measures used to assess effectiveness were highly variable. Shoulder function and pain were evaluated in all studies (n = 13), but using different tools. Muscle strength (n = 6),30,31,35,39,40,42 range of motion (n = 8),30–32,35,37,39,40,42 radiological outcomes (n = 3),29,35,40 pressure pain threshold (n = 1),33 and satisfaction (n = 2)33,36 were also evaluated in some studies. Psychosocial (n = 3)30,35,38 and psychological (n = 2)33,37 factors were assessed in a few studies. The mean PEDro score was, on average, 6.2/10 (standard deviation [SD]: 1.3) indicating a good methodological quality of the articles included (Table 1).

The TIDieR scores are provided in Table 2. The average score across all studies on the TIDieR was 10.4/24 (SD: 2.6) points, ranging from 6 to 15 points. The percentage regarding each TIDieR item checklist varied from 7.6% (lowest value achieved) to 100% (highest value achieved) and are presented in Table 3. All studies (n = 13) provided information about the first item (brief name that provided the name or a phrase that described the intervention). Few studies reported possible adaptations in the management (n = 4)26,27,30,40 and adherence (n = 2)32,35 to the program (Table 3).

Table 2.

Template for Intervention Description and Replication (TIDieR) score, score of each item and total score, for the included studies.

Score for item 1 to 12 and overall TIDieR score. Green, item was reported (2 points); yellow, item was partially reported (1 point); red, item was not reported (0 point).

Regarding implementation fidelity across the NIHBCC scale, the overall fidelity score across all the studies (n = 13) of each domain ranged from 0.5 to 51.2%. The highest score was for treatment design domain that reached 51.2%, and the lowest scores were in the domains of training of providers and treatment enactment, which scored 0.5% and 3.8%, respectively. The treatment delivery and treatment enactment also achieved low final scores, 20.8% and 3.8%, respectively. The overall score of all studies with the five domains taken together was 27.4%. All information about NIHBCC score is presented in Table 4 with the score of each article.

This review evaluated the description and implementation fidelity of exercise-based interventions included in clinical trials for individuals with rotator cuff tears. Our findings show these clinical trials showed good methodological quality, but the description and fidelity of the intervention were poorly reported. Clinicians and researchers may face challenges when attempting to replicate and implement the interventions due to insufficient available information.

Most of the included clinical trials were published in 2018 or after. Even after the publication of the TIDieR checklist in 2014, few items were properly reported in more recent studies, making it difficult to replicate the intervention. Interventions performed in individuals with rotator cuff tear were highly variable. Studies compared exercises versus surgery,30,31,35,39,40 different exercises modalities36 or exercises associated with manual therapy33,37 or electrotherapy38 and also injectable therapies.32,34,41,42

The average quality of the included studies was considered “good” based on an overall mean score greater than 6 points on the PEDro scale.43 While the PEDro scale is a valid tool to assess the methodological quality of clinical trials,44 and the studies are of good quality, the lack of information hinders the implementation of the intervention. Full treatment description would allow replication of the intervention,8 while incomplete information creates a barrier in the implementation of potential effective treatments.10 The results of this study show a lack of information about interventions because the low total TIDieR scores indicate a lack of information even in those studies recently published.32–39,45 Missing information included treatment duration, dose, intensity, intervention planning, and modifications, as well as patient adherence to the program.

Clinicians treating individuals with shoulder pain and rotator cuff tears, researchers conducting trials with this population and also for policymakers interested in implementing evidence-based interventions may face challenges when trying to use interventions included in this review. Patients may not adhere to exercise interventions or opt for other approaches such as surgical procedure, medications, or injectable therapies due to a lack of clarity about, for instance, exercise progression or adaptations. Our findings are in agreement with other studies involving physical therapy treatment in patients with hip osteoarthritis,22 femoroacetabular impingement syndrome,46 subacromial pain syndrome,12 hamstring strain,21 Achilles tendon rupture,47 and preoperative patient program for orthopaedic surgery48 where researchers also need to improve the quality of reporting interventions.

In general, the evaluation of fidelity of studies involving individuals with rotator cuff tear was low. As in previous studies,12,25 the lowest scored domain was training of providers, meaning that no detailed information about how the training was conducted, the assessment of skill acquisition, and whether any specific characteristics were encouraged or avoided during the sessions. Non-significant results of a treatment can potentially be attributed to ineffectiveness of an intervention or also by providers with inadequate training and monitoring.24 In addition, the patient's skills during the exercises should be assessed and informed. It is also important to minimize the chances of the patient not performing or performing the exercise incorrectly when not supervised.

A clinical trial is often considered a high-impact publication. However, detailing the implementation process is essential to understand possible barriers and facilitators of an intervention program. Over the years, authors have been striving to minimize bias and improve the methodological quality of clinical trials.49 However, improving the quality of information is also important, and can impact the ability to replicate the intervention, especially in clinical practice. Although it is easier to think about the importance of designing and delivering the treatment, it is also necessary to train professionals, and to understand how the patient receives and executes the intervention. Therefore, future studies need to improve the quality of reporting interventions, and scientific journals should request authors to report each step of the intervention more clearly, facilitating its replication and implementation. It is possible that improving the understanding and execution of the exercise program leads to greater adherence to treatment. Through good fidelity delivery, it is possible to improve evidence-based practice and assist the clinicians in decision making.50

This study has some limitations. Some of the items of the NIHBCC may have influenced the final score. For instance, the item regarding the method to ensure dose equivalence between intervention groups was rated as "not applicable" in all articles, which removes this item from the final score in the treatment design domain. In contrast, for studies comparing similar exercises without the addition of other modalities or treatments this item was relevant and included in the final score. In addition, it is possible that the authors may have provided an item, as for example, advice or education, but did not mention that in the manuscript. There are no methods to weight or infer if it was done unless it was explicitly mentioned, resulting in a reduction of the total score.

Exercise-based treatment in individuals with rotator cuff tears presents poor description and fidelity of the interventions. More information about the infrastructure, adaptations during the treatment and adherence are needed to improve the description of the treatment. In addition, information about provider training is essential. Future studies should improve the treatment description and fidelity in attempt to facilitate implementation by clinicians, researchers, and policymakers and replication of the study.

LPR, DPRS, PRC and FAS designed the study. LPR and FCM performed the study selection and data extraction. All authors made significant contributions to the scientific knowledge and development of the review, including approval of the final manuscript.

All data relevant to the study are included in the article or are available as supplementary material.

This project was funded by Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES), Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) (144436/2019-1) and Erasmus+ (63350-MOB-00143). The funders had no role in study design, data analysis, decision to publish, or preparation of the manuscript.