Introduction
Hip-related pain is an umbrella term encompassing pain arising from non-arthritic hip joint pathologies in three categories: 1) femoroacetabular impingement syndrome (FAIS), 2) acetabular dysplasia and/or instability, and 3) hip joint problems without distinct morphology (such as labral and chondral lesions).1 Hip-related pain is associated with significant burden in young and middle-aged adults, leading to poor function and low quality of life.2
Exercise therapy for musculoskeletal pain is an effective, low-cost intervention with few adverse events,3,4 and is recommended in high quality clinical practice guidelines for diagnoses such as osteoarthritis, low back pain, neck pain and rotator cuff disorders.5 Exercise therapy is also suggested as a key component of treatment for hip-related pain, whether or not surgical intervention is undertaken.6,7 Physical therapist-led interventions that mainly include exercise therapy, have moderate positive effects compared to sham/control interventions.8 However, the optimal content and delivery of exercise therapy is unclear.9
To establish best practice, the details of exercise therapy interventions must be described. Complete reporting of the details of an intervention is an important aspect of study quality. Incomplete reporting of the intervention details within a study limits the ability to inform future research,10 and lowers the clinical applicability of the research findings.11 With the aim to increase the reporting completeness of complex interventions, the CONSORT statement extension for non-pharmacological trials,10 and the Template for intervention description and replication (TIDieR) guidelines,12 were developed. The Consensus for Exercise Reporting Template (CERT) was developed to guide and facilitate reporting completeness of exercise therapy interventions.13 The CERT checklist can also be used to evaluate completeness of reporting of exercise therapy protocols. It is unknown whether reporting completeness of exercise therapy interventions for hip-related pain has improved since the publication of the CERT guidelines (i.e., 2016). It is also unclear whether any relationship exists between reporting completeness of exercise therapy interventions and other factors related to study quality, such as risk of bias.
Some studies have examined the reporting completeness of exercise therapy interventions for pain around the hip and groin.14–16 Systematic reviews have described incomplete reporting in studies using exercise therapy as treatment for extra-articular groin pain (i.e. adductor-related, inguinal-related or pubic-related groin pain)15 and hip OA.14 A recent scoping review examined specific exercises for FAIS, and how these relate to proposed pathomechanics, in people treated with a non-operative approach.16 The authors used CERT as a secondary measure and a proxy for study quality, but did not report any detailed CERT synthesis or description of the intervention content.16
To the authors knowledge, there are no systematic reviews examining the completeness of the reporting of exercise therapy interventions with/without concurrent surgical intervention for people with hip-related pain. Such knowledge can inform understanding of efficacy of exercise therapy for this patient population, as well as improve future research. Therefore, the main aim of this study was to assess the reporting completeness of exercise therapy protocols for people with hip-related pain. In addition, the aim was also to provide a summary of the content of the exercise therapy protocols included in the study, as well as compare CERT scores between i) studies published before and after publication of CERT, and ii) studies with different levels of risk of bias.
Materials and Methods
This systematic review, adhering to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement, was preregistered in the PROSPERO database (CRD42020154139).
Literature search and study selection
Literature search
A systematic search in the MEDLINE, CINAHL, and Cochrane databases was conducted by a research librarian for research studies from earliest available to 31 October 2019 and updated February 19, 2021. The following key words were used:
Population-based terms: ‘hip’, ‘hip pain’, ‘femoroacetabular impingement’, ‘non arthritic hip pain’, ‘acetabular labrum’, ‘dysplas*’, ‘hip chondral’, ‘hip instability’
Intervention-based terms: ‘exercis*’, ‘training’, ‘physiotherap*’, ‘therap*’, ‘physical therap*’, ‘rehabilitat*’, ‘manual therap*’, ‘mobilization’, ‘patient education’, ‘conservative’, ‘non operative’
The search strategy was adapted to the different databases (Appendix A). Reference lists of included studies were screened for further relevant studies.
Study selection
Studies were eligible for inclusion if they were randomized controlled trials, cohort studies, case control studies, or published study protocols that included:
-
People with non-arthritic hip-related pain, as defined by the International Hip-related Pain Research Network, Zurich 20181
-
People with persistent pain (>3 months duration)
-
Aged 18 to 50 years
-
A description of exercise therapy, including details such as treatment modality, prescription, type and/or duration.
Review studies and clinical commentaries were not eligible for inclusion. Studies were also excluded if they included:
-
People with verified osteoarthritis (Tönnis grade >1)
-
People with total hip replacement
-
People with acute hip injury, such as a fracture of the neck of the femur
-
People with extra-articular pain, such as adductor-related groin pain
-
Interventions including treatment (e.g., therapeutic injection therapy or manual therapy) without exercise therapy
After the initial search performed by a research librarian, all records were imported into the Covidence software, and duplicates were removed. Two researchers (AE, AP) conducted independent screening of titles and abstracts and eligible studies were read in full text (Figure 1). Any disagreements on inclusion were resolved in a consensus meeting, with a third researcher (EA) acting as deciding vote if necessary.
Data extraction and assessment
Descriptive details of included studies were extracted by a single researcher and included author, publication year, patient population, sample size, participant age, intervention type (exercise therapy alone or in combination with surgery), outcome measures and results.
The CERT checklist17 was used to extract and assess the reporting completeness of the included studies. The CERT consists of 16 items over seven domains; what (materials); who (provider); how (delivery); where (location); when, how much (dosage); tailoring (what, how); and how well (compliance/planned and actual). Each item was scored as a 0 (not described), 1 (described) or NA (not applicable). The score ranges from 0 to 19 with higher numbers indicating better description. If any studies compared exercise therapy to surgery and exercise therapy, the treatment protocol for the group receiving exercise therapy only was examined. If a study had multiple exercise therapy treatment groups, the treatment protocol hypothesized to be superior was selected for analysis. Data from each study and any related sources (i.e., appendices, supplemental material, published study protocols, development descriptions and feasibility studies) was independently extracted and assessed by two researchers with experience treating hip-related pain with exercise therapy. The Explanation and Elaboration statement to the CERT was used to guide scoring.17 The reason for any items being considered ‘not described’, were recorded. The details and location of each item response was recorded for each study (Appendix B). To evaluate if completeness of reporting has improved since the publication of CERT (December 2016), studies published 2019 or later were considered likely to have had access to the tool during planning and conducting of their study. Data related to risk of bias was also independently extracted and assessed by two researchers (AE, AP) using the Cochrane risk of bias tool version 2 (RoB 2).18 Any disagreements on CERT or RoB 2 scores were resolved in a consensus meeting, with a third researcher (EA) acting as deciding vote if necessary.
Data analysis
Cohen’s kappa was used to measure agreement between raters on the CERT score and RoB 2 tool, and median and inter-quartile ranges (IQR) were used to describe the data. The item responses from the included studies were synthesized by a single researcher, to provide an overview of the contents of the interventions. Mann-Whitney U test was used to compare the CERT scores of studies published before and after the publication of the CERT checklist, and studies with different levels of bias.
Results
Study selection
In total, 5444 records were identified, and 234 studies were screened in full text. While 52 studies used exercise therapy as part of their intervention, only 23 studies19–41 (44%) reported any details. The remaining 29 studies were not included in the CERT synthesis (Figure 1). Twenty-five of these 29 studies were surgical trials, using exercise therapy as part of post-operative rehabilitation (Appendix C).
Study characteristics and CERT scores are described for the 23 studies that could be included in the synthesis.
Study characteristics
Study design
Of the 23 studies, three were randomized controlled trials (RCTs),23,27,30 four pilot RCTs,19,25,29,34 four RCT study protocols,21,36,37,41 five prospective case series,20,22,24,26,32 one feasibility study,28 and six retrospective case series.31,33,35,38–40 Sample size ranged from 15 to 348. The studies used exercise therapy alone,19–29,40,41 or in combination with hip surgery.30–39
Participants
The studies reported mean ages ranging from 23-43 years and included participants with a diagnosis of FAIS,19–25,27,29,30,34–38 dysplasia,26,28,41 FAIS and borderline dysplasia,33,39 hip-related pain,31,40 and chondrolabral pathology.32
Outcome measures and results
The studies included in the synthesis reported effects of interventions on hip-specific and/or generic patient-reported outcome measures, as well as tests of physical function (Table 1).
CERT score synthesis
None of the studies referred to CERT in their methods section. CERT scores ranged between 1 and 17, with a median of 12 (IQR 5-14). Five21,25,28,30,36 studies reported on 15 or more CERT items (75%), 1420,21,23–30,34,36,37,41 reported on 10 (50%) or more and 18 studies19–21,23–30,32–34,36,37,40,41 reported on five (25%) or more items. No study had the maximum score of 19. The most reported items were tailoring (14a and 14b), which was reported in 2019–34,36,37,40,41 of 23 studies. The lowest score was observed for motivation strategies (item 6) and starting level (item 15), which was reported in two23,37 and three21,28,41 studies, respectively. Details regarding scores are provided in Table 2, and protocol content in Appendix B. The agreement for CERT scores between the two raters was K=0.72, representing a substantial agreement.
CERT items synthesis
What (item 1)
Eighteen (78%) studies19,21–30,32–34,36–39 described equipment used. Commonly used materials were resistance bands, weight cuffs, stationary bicycles, and unstable surfaces. While most studies described the equipment used (e.g., leg press), the specific type of equipment (e.g. model of machine) or resistance level was rarely described.
Who (item 2)
Six (26%) studies21,25,29,30,34,36 sufficiently described the title and qualifications of the prescriber. Of the studies that did not report this item, eleven studies20,23,24,26–28,32,37,38,40,41 described the professional title (primarily physical therapists) but not qualifications or experience, while six studies19,22,31,33,35,39 provided no details about the prescribers.
How (item 3-11)
Individual/group (Item 3)
Eleven (48%) studies20,21,23–27,29,30,34,37 provided detail whether exercise therapy was conducted in individual or group settings, with all 11 studies using individual training sessions.
Supervised/unsupervised (Item 4)
Sixteen (70%) studies20,21,23–30,34,36–38,40,41 reported on supervision. Fourteen of these20,21,23–25,27,29,30,34,36–38,40,41 used a combination of supervised (ranging from 5-24 sessions) and non-supervised exercise. One study28 had supervision on every training session (20 sessions), and one study26 had no supervision.
Adherence (Item 5)
Twelve (52%) studies20,21,23–26,28,30,34,36,37,41 reported adherence tracking. Eight studies20,21,23–25,34,36,37 used a training diary or app, two studies21,30 a Likert scale, two studies19,26 used verbal confirmation of adherence, and three studies28,30,41 recorded the number of attended sessions.
Motivation (Item 6)
Two (9%) studies23,37 reported any motivation strategies used. This consisted mainly of education on the importance of adherence to the exercise therapy.
Progression (Items 7a&b)
Eight (35%) studies24,28,30,32–34,36,41 reported criteria for progression of exercise. The criteria varied, with studies reporting the use of time frames,41 pain free range of motion and ambulation,32–34 pain free exercise execution,30,32–34 a rate of perceived exertion24,36 and/or VAS pain cut-off,36 being able to complete >2 repetitions above prescribed on last set,28,41 and/or force production limb symmetry index32–34 as markers for progression.
Ten (43%) studies21,24–26,28–30,36,41 described how the program was progressed. Most studies used concurrent means of progression, including: i) increased exercise volume (repetitions and/or sets performed),21,25,26,41 ii) increased intensity, targeting heavier loads,25,28,29,36,41 iii) progression from isolated muscle exercises to more complex motions, such as compound functional movements, single leg work or unstable surfaces,21,24,25,29,30,36 iv) addition of more exercises over time,27 v) faster loading rates, such as plyometric training.25,36
Exercises (Item 8)
Seventeen (74%) studies19,21,23–30,32–34,36,37,40,41 reported the exercises used. These included isolated non-weightbearing exercises (such as side lying hip abduction),19,21,23–30,32–34,36,37,41 isometric trunk training (such as planks),19,24,25,27,29,33,36 compound lower extremity exercises (such as squats and lunges),21,23–25,27–30,32–34,36,37,40,41 cardiovascular training (with exercise bikes, elliptical machines or running),25,30,32–34,36 stretching/mobility,21,27,29,30,32–34,37 and/or plyometrics (jumping and landing drills, running progressions).25,32–34,36
Home component (Item 9)
Fourteen (61%) studies20,21,23–26,28–30,34,36,37,40,41 reported on any home component to their exercise program. Twelve studies20,21,23–25,29,30,34,36,37,40,41 primarily provided participants with a home-based program, with supervised sessions to check exercise technique and progression. One study26 used a home program only as the intervention, and one study28 used no home component.
Non-exercise component (Item 10)
Seventeen (74%) studies19–23,25,27–30,33–39 described any non-exercise component. Manual therapy (such as soft-tissue treatment and/or mobilizations) was performed in 13 studies21,23,25,27,29,30,33–39 and 8 studies19–23,25,30,34 used patient education, commonly concerning hip anatomy and activity modification.
Adverse events (Item 11)
Twelve (52%) studies20,21,23–26,28–30,36,37,41 described adverse events related to their exercise intervention. No serious adverse events related to exercises were reported, though 4 studies23,24,28,30 reported participants experiencing muscle soreness and a transient increase in pain after exercise therapy. One study reported approximately 25% of patients dropping out of the intervention due to increases in pain or fatigue related to the exercises.26
Where (item 12)
Seventeen (74%) studies20,21,23–30,34,36–41 included descriptions of the study setting, with exercise therapy interventions mostly being performed at outpatient physical therapy clinics.
When/how much (item 13)
Fourteen (61%) studies19–21,24–30,34,36,37,41 reported on intervention dosage. The duration ranged from three weeks to six months, and frequency ranged between daily training to three sessions weekly. Six studies21,24,28,30,36,41 provided dosage anchored against a measure of intensity, such as rate of perceived exertion or a percentage of repetition maximum (RM).28,36,41
Tailoring (items 14-15)
Twenty (87%) studies19–34,36,37,40,41 reported whether the program was tailored to the individual, of which 15 used an individualized approach20,21,23–25,27,29–34,36,37,40 and five a generic program.19,22,26,28,41 The treating physical therapist tailored the program based on the patient impairment, pain-free range of motion surgical procedure, and desired activity levels and sport-specific demands. Three studies (13%)21,28,41 reported the patients’ starting level, two of which RM-based starting levels,28,41 and one where the treating physical therapist adapted the starting dose based on patient presentation.21
How well (item 16 a & b)
Eight (35%) studies21,23,25,29,30,34,36,37,41 reported on intervention fidelity. To increase fidelity, physical therapists delivering the intervention were given written instruction21,23,30,34,36 and physical training21,23,25,29,30,36 in application of the protocol. Two studies used follow up sessions with the researchers.23,36 The authors of two studies were also treating clinicians.25,37 Seven studies (54%)20,23–25,27,28,30 reported whether the interventions were delivered according to plan, primarily using reports of adherence and attended sessions to describe the applied intervention. The included RCT study protocols21,36,37,41 were not applicable for this item as the intervention had not been completed.
Studies published before and after CERT
Studies published before 201921–31,33–37 had higher (better) CERT scores (n=16, median 14, IQR 7.25-15) compared to those published 2019 or later19,20,32,38–41 (n=7, median 6, IQR 3-11) (p=0.034).
Risk of bias
Risk of bias was high in 14 studies,19,20,22,24,26–28,31–33,35,38–40 some concerns in 8 studies21,23,25,29,30,34,37,41 and low in one study36 (Table 1, Appendix D). Studies with some concerns or low risk of bias were analyzed together (n=9), and had higher CERT scores (median 14, IQR 14-16) compared to those with high risk of bias (n=14) (median 6, IQR 3-11) (p<0.001). Agreement for risk of bias was substantial (K=0.69).
Discussion
Fifty-two studies used exercise therapy as part of their intervention to treat hip-related pain, but 29 studies provided no details beyond mentioning the use of exercise therapy and could not be included in the CERT synthesis. Of the 23 studies included in the synthesis, the median CERT score was 12 (IQR 5-14) and none reached the maximum score of 19. The results suggest that studies using exercise therapy to treat hip-related pain did not report protocols in sufficient detail to allow replication in future studies or clinical practice.
In line with the results of the present study, previous systematic reviews have reported median CERT scores ranging from 5-15 in exercise therapy studies for diagnoses such as rotator cuff disorders,42 achilles rupture,43 low back pain,44 and hip OA.14 In the present study, the most described item was tailoring (item 14a and 14b, described by 87%), while the least reported items were motivation strategies (item 6, 9%) and starting level (item 15, 13%). Motivation is a key factor in adherence to rehabilitation,45 and behavior change related to physical activity.46 A lack of describing motivational strategies could imply this aspect has not been considered, which could in turn limit the effectiveness of otherwise well designed and described protocols. People with hip-related pain is a heterogenous group with varying levels of functional impairments.2,47 Therefore, the appropriate starting level may be unique to the individual, and a lack of criteria description may lead to a starting level that is too challenging for some, and not sufficiently demanding for others. Also, less than half of the studies described progression criteria (item 7a, 35%), or how progression was performed (item 7b, 43%), and 14 of 23 (61%) studies had descriptions of dosage, such as repetitions, sets and frequency (item 13). This is in accordance with previous systematic reviews where the commonly unreported items were motivation strategies, starting levels, progression criteria and fidelity.14,42–44 Based on our results and similar reviews, important aspects of exercise therapy protocols are consistently unreported in the literature. Improvement of reporting completeness may better our understanding of exercise therapy for this patient population, as well as allow for replication in further studies and implementation into clinical practice.
As the optimal exercise therapy for people with hip-related pain is currently unknown, a clinical focus on impairments related to the disorder, such as reduced hip muscle strength, has been suggested.7 Without complete reporting in research trials, the strategies to best address these impairments are unclear. Also, results from trials comparing exercise therapy to other interventions, such as hip arthroscopy, need to be analyzed with completeness of reporting in mind, as exercise therapy may encompass a wide range of treatment strategies.
In the present study, CERT scores were significantly lower (worse) in studies published 2019 or later compared to those published earlier. One reason for this result could be that only six studies were published in the later time frame, whereof three were retrospective. Complete reporting requires thorough planning which is less likely achieved in a retrospective compared to a prospective design. Although the goal of reporting guidelines, such as CERT, is to improve reporting completeness, a recent systematic review using TIDieR found that reporting on physical therapy interventions had not meaningfully improved in a sample of trials from 2000 and 2018.48 This, and the fact that no studies in our review referred to CERT in their methods section, indicate of a lack of implementation of guidelines by the research community. In our review, studies with lower risk of bias had significantly higher CERT scores than those with higher risk of bias. While the RoB 2 tool was designed to assess risk of bias in RCTs, the authors used it as a secondary measure for all included studies to provide a broad picture of the risk of bias, although this tool may not be the most appropriate for all study designs. While CERT and the RoB 2 measure different constructs, their association may reflect that a more thoroughly designed and planned study address factors related to risk of bias as well as intervention transparency. Increasing the overall scientific rigor of the published literature by raising awareness and implementation of relevant guidelines, and addressing important aspects of risk of bias, may also affect reporting completeness.
CERT has been suggested to be used as a checklist when planning an exercise therapy protocol, and as a measure of reporting completeness in systematic reviews on exercise therapy.17 There are some challenges in using CERT in systematic reviews. First, there is no consensus on what is to be considered a good or sufficient score. Charlton et al used a classification of CERT scores based on percentage of items described; high (>75% of items reported), moderate (60-74%) and poor (<60%) levels of reporting standard.15 However, these values were chosen based on cut-offs from the Downs and Black scale, and not on recommendations or evaluation of data specific to CERT. Second, the maximum score of 19 was not met by any studies included in the current study or in any of the other reviews using CERT,15,16,42–44 apart from one study in the systematic review on hip osteoarthritis.14 Further research into what constitutes a realistic and relevant score might be needed. Third, the use of a composite overall score may not be appropriate, as some items may be of greater importance to allow replication. For example, the type and goal of exercises performed (for example: isolated hip muscle strength, functional performance), the dosage prescribed (volume, intensity), progression and adherence might be considered the basis of an exercise therapy protocol, with other items providing additional details. Further research on the relative importance of items or domains in the CERT may enhance its interpretability, and potentially lead to weighting of different items.
Strengths of this study include adherence to established guidelines and good agreement between raters. Both raters were experienced in using exercise therapy for hip-related pain, which may be valuable for understanding the nuance of the protocols. Also, the inclusion of a comprehensive description of the exercise therapy protocol content may serve as a summary to researchers and clinicians. There are limitations to the current study. To be included in the CERT synthesis, the inclusion criteria required some form of exercise therapy description. For this reason, most studies (29 of 52) that had used exercise therapy could not be included in the CERT synthesis, due to no description of their intervention. As such, these results may paint an overly positive picture of the state of reporting completeness for this patient population and should be viewed in combination with the other 29 studies. It should be noted that 25 of the 29 studies not included in the CERT synthesis were surgical studies using exercise therapy as part of post-operative care, highlighting a lack of reporting in these trials.
Conclusions
Less than half of the studies (23 out of 52) using exercise therapy to treat hip-related pain reported sufficient details to be included in the CERT synthesis. Of the 23 studies included in the synthesis, the median CERT score was 12 (IQR 5-15), with no study reaching the maximum CERT score of 19. Interestingly, studies with less risk of bias had better CERT scores. Furthermore, the publication of CERT did not yield better scores. Taken together this seems to suggest that study rigor may drive better overall exercise intervention reporting. The lack of reporting completeness of exercise therapy protocols makes it difficult to replicate interventions in future research as well as in clinical practice, and to draw conclusions on efficacy and dose-response to such interventions.
Conflict of Interest
The authors declare no conflict of interest.