Types of studies

Randomized controlled trials (RCTs) and quasi‐randomized (method of allocating participants to a treatment which is not strictly random, for example by date of birth, hospital record number or alternation) controlled trials evaluating conservative interventions for treating exercise‐related musculotendinous, ligamentous and osseous groin pain were included.

Types of participants

People with a diagnosis of exercise‐related groin pain caused by pubic bone‐related dysfunction, enthesopathy or muscles strains, established by the primary studies, were included. Excluded were groin pain from any type of hernia, sacroiliac dysfunction, piriformis syndrome, nerve entrapments, lumbar spine, gynaecological or urological diseases, and hip pathologies.

Types of interventions

Included were all forms of conservative treatment for exercise‐related groin pain, such as strengthening of the muscles stabilising the pelvis and hip joints, stretching of the hip muscles, electrotherapy (for example, laser and ultrasound therapy), manual therapy, non‐steroidal anti‐inflammatory drugs (NSAIDs), steroidal injections or prolotherapy. The interventions could be single interventions or complex interventions, such as programmes incorporating different types of exercises or combinations of different interventions. Studies comparing conservative with surgical treatments were excluded.

Our main comparisons were:

• conservative intervention (single or complex intervention) versus placebo intervention;

• one conservative intervention (single or complex intervention) versus another conservative intervention (single or complex intervention).

All of the interventions could be set in the context of other interventions being provided to all trial participants.

Types of outcome measures

Electronic searches

We searched the Cochrane Bone, Joint and Muscle Trauma Group Specialised Register (December 2011); the Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library 2011, Issue 4); MEDLINE (1948 to November week 3 2011); EMBASE (1980 to Week 49 2011); CINAHL (1982 to December 2011); Latin American and Caribbean Health Sciences (LILACS) (1982 to December 2011); PEDro, the Physiotherapy Evidence Database (1929 to December 2011); SPORTDiscus (1985 to December 2011); and OTseeker, the Occupational Therapy Systematic Evaluation of Evidence Database (to December 2011). No language restrictions were applied.

The subject‐specific strategy was combined with the sensitivity‐maximizing version of the Cochrane highly sensitive search strategy (Lefebvre 2011) to identify randomized trials in MEDLINE. Complete search strategies are shown for CENTRAL, MEDLINE, EMBASE and LILACS in Appendix 1.

We searched Current Controlled Trials for ongoing studies.

Searching other resources

We contacted professional experts in this field and searched reference lists of papers and conference proceedings (2000 to 2011), such as of the American College of Sports Medicine Annual Meetings; American Physical Therapy Association; American Academy of Physical Medicine and Rehabilitation Annual Assembly; Canadian Physiotherapy Association Annual Congresses; Australian Physiotherapy Association Conferences.

Selection of studies

Two review authors (MOA and BNGS) independently selected and assessed potentially eligible articles for inclusion in the review. Any disagreements were discussed and resolved by consensus. Where necessary, a third author (MSP) acted as adjudicator. There was no blinding of authors, institution or journal of publication at any stage.

Data extraction and management

Two review authors (MOA and RBA) independently extracted data from selected articles using a standard extraction form. Data extracted included study methods, the characteristics of participants, interventions, outcome measures and results. When necessary, requests were sent to trial authors for additional information or data.

Data entry into RevMan was by three review authors (MOA, BNGS, RBA).

Assessment of risk of bias in included studies

As recommended by The Cochrane Collaboration (Higgins 2011), two review authors (MOA and RBA) independently assessed the risk of bias for each included study. Disagreements were resolved by a third review author (MSP). The following domains were evaluated for risk of bias: random sequence generation; allocation concealment; blinding of participants and personnel; blinding of outcome assessment; incomplete outcome data; selective outcome reporting; and other sources of bias. Risk of bias for the individual domains was assessed as either 'low risk of bias', 'high risk of bias' or 'unclear risk of bias'.

Measures of treatment effect

Dichotomous outcome data were expressed as risk ratios (RR), with 95% confidence intervals. For continuous outcomes, mean differences (MD) were used if the studies use the same instrument of evaluation. However, if the primary trials assessed the same variables using different scales, we planned to use the standardized mean difference (SMD). The estimated effects were reported together with the 95% confidence intervals (95% CI).

Unit of analysis issues

The unit of analysis was based on the individual patient. Due to the biological and epidemiological characteristics of exercise‐related groin pain, cluster randomized controlled trials and cross‐over randomized controlled trials were not expected.

Dealing with missing data

For missing data (for example, publication bias, outcome not measured, incomplete reporting, lack of intention‐to‐treat analysis, attrition from the study) we planned to adopt the following strategies.

• Whenever possible, we would contact the original investigators to request missing data.

• Make explicit the assumptions of any methods used to cope with missing data, for example, that the data were assumed to be missing at random, or that missing values were assumed to have a particular value such as a poor outcome.

• Perform sensitivity analyses to determine how sensitive the results were to reasonable changes in the assumptions that were made.

• Address the potential impact of missing data on the findings of the review in the Discussion section.

Assessment of heterogeneity

We planned to assess heterogeneity by visual inspection of the forest plot (analysis) along with consideration of the Chi² test for heterogeneity (with a level of significance of P < 0.1) and the I² statistic. We considered that there was probably substantial heterogeneity where I² > 50%.

Assessment of reporting biases

We planned to assess the likelihood of publication bias using funnel plots, if there were at least 10 trials contributing to a primary outcome.

Data synthesis

In the absence of significant clinical and statistical heterogeneity, we planned to use a fixed‐effect model and 95% confidence intervals for pooling the results of trials testing comparable interventions. However, if we had found diversity in clinical or methodological characteristics, or unexplained heterogeneity, we would have considered using a random‐effects model for analysis. Where possible, we planned to perform intention‐to‐treat analysis.

Subgroup analysis and investigation of heterogeneity

If data had been available, we planned to investigate heterogeneous results for specific patient groups, performing subgroup analyses according to gender, duration of symptoms, and activity (recreational versus professional athletes). We planned to investigate whether the results of subgroups were significantly different by inspecting the overlap of confidence intervals and performing the test for subgroup differences that is available in RevMan.

Sensitivity analysis

Where possible, we planned sensitivity analyses examining various aspects of study and review methodology, such as the inclusion or not of studies at high risk of bias (specifically from lack of allocation concealment, assessor blinding, and incomplete outcome data) and inclusion or not of trials only reported in abstracts. We also planned to conduct sensitivity analyses to look at the effects of missing data.