Types of studies

We will include randomised controlled trials and quasi‐randomised trials (i.e. where the method of allocation is known but is not considered strictly random, such as alternation, date of birth, and medical record number). We will also include cluster‐randomised trials.

Types of participants

We will include trials that studied participants with patellar tendinopathy, as defined in the trials. The tendinopathy can be caused by an acute traumatic injury, or chronic onset. Study participants can be either athletes or non‐athletes. We will exclude trials that included combined injuries (e.g. participants diagnosed with patellar tendinopathy and patellofemoral pain syndrome).

Types of interventions

We will include trials reporting strengthening exercise interventions. These exercises must aim to strengthen muscles of the lower limb. The interventions may include any contraction types; i.e. concentric, eccentric, eccentric‐concentric and isometric exercises. These exercises are often conducted using squats, leg press or leg extension position. We will exclude trials where the intervention is based upon stretching exercises in isolation as well as running‐/cycling‐based exercises, Tai Chi and dancing.

We will accept trials with any content, duration, frequency or intensity of exercise. We will include trials regardless of whether exercises are delivered in group or individual classes, whether they are land‐ or water‐based and whether exercises are supervised or not.

We will extract data on co‐interventions (if available). We will not exclude any trial due to co‐interventions.

Types of outcome measures

We will include clinically relevant measures that could be considered patient‐centred. Physiological and biomechanical variables (for example, range of motion, motor control, muscle endurance) will not be considered for this review.

Electronic searches

We will search for randomised controlled trials from the following electronic databases without restrictions of language or date of publication: Cochrane Central Register of Controlled Trials (the Cochrane Library current issue), MEDLINE (1950 to present) and Embase (1980 to present). The MEDLINE search strategy will also include Cochrane's highly sensitive search strategy for randomised controlled trials (Appendix 1).

We will also conduct a search of ClinicalTrials.gov (www.ClinicalTrials.gov) and the World Healh Organization (WHO) trials portal (www.who.int/ictrp/en/).

Searching other resources

We will also search reference lists of eligible papers and previous published systematic reviews related to patellar tendinopathy.

Selection of studies

Two review authors (ADL and LCHJ) will independently screen titles and abstracts for potentially eligible studies. Full‐text papers will be used to determine the final inclusion in the review. Disagreements between review authors will be resolved through discussion or by arbitration by a third review author (LOPC). We will include full‐text papers regardless of the date of publication and written in any language. Papers written in English, Portuguese, Spanish, Italian and Dutch will be included as the review team includes authors able to read these languages. All remaining papers that are written in languages other than these will also be included, after we seek the help of translators.

Data extraction and management

Data extraction will be completed independently by two review authors (ADL and LCHJ) from each of the eligible papers using a standardised data extraction form containing the following information.

1. Bibliometric data: authors, year of publication, language.

2. Study characteristics: study design, sample size, description of the sample, country, recruitment method, funding.

3. Characteristics of the participants: gender, age, duration of symptoms, severity of the condition at baseline, other baseline clinical characteristics, inclusion and exclusion criteria.

4. Description of the exercise interventions, including: type of exercise, exercise content, number of sessions, duration of each session of treatment, length of programme, supervision, group or individual, setting, comparisons, and co‐interventions.

5. Outcomes; major and minor outcomes specified will be collected and time points reported, including a description of the measurement tool used for continuous outcomes (scale of tool and direction of effect).

6. Study results on outcomes of interest: number of participants in the analyses, mean and standard deviation per treatment group for continuous outcomes, and number of events and number of participants per treatment group for dichotomous outcomes.

7. Notes: funding for trial, and notable declarations of interest of trial authors.

If data on more than one pain outcome are reported in a trial, we will extract data for the measure that is highest on the following hierarchy: pain overall; pain at rest, pain with activity, unspecified. From possible measures of pain, we will preferentially extract data from either Visual Analogue Scale (VAS) over Numerical Rating Scale (NMRS) over others (e.g. McGill Pain Questionnaire).

If data on more than one physical function scale are reported in a trial, we will extract data according to the following hierarchy: 1) the most prevalent measure, 2) second‐most prevalent measure and so on. We will also use validated over unvalidated instruments as a hierarchy method.

If data on more than one quality of life instrument are reported, we will preferentially extract data from the SF‐36 over others.

If both final values and change from baseline values are reported for the same outcome, we will select change scores, but if only post scores are available from most studies, then we will use post scores to enable meta‐analysis including the largest possible number of studies. We will perform a meta‐analysis even if we have a mix of change and post scores.

We will select unadjusted values over adjusted values but include adjusted if unadjusted not reported.

We will select data from intention‐to‐treat (ITT) analysis first, then per‐protocol, if ITT is not reported.

We will select closest to: immediately post treatment, six months, 12 months.

Assessment of risk of bias in included studies

The 'Risk of bias' assessment of included studies will be performed independently by two review authors (ADL and LCHJ) using the'Rrisk of bias' assessment tool recommended by Cochrane (Higgins 2017). Disagreements between review authors will be resolved by discussion, or arbitration by a third review author (LOPC). Each of the items of the 'Risk of bias' assessment will be scored as “high”, “low” or “unclear”.

The assessment will be done independently two review authors (ADL and LCHJ) and any disagreements will be resolved by discussion, or arbitration by a third review author (LOPC).

Measures of treatment effect

If continuous outcome measures including pain, disability and quality of life are measured with the same scale across studies, we will present the mean difference (MD) and 95% confidence intervals (CIs). If different scales are reported across trials, we will present the standardised mean difference (SMD) and 95% CIs. SMD will be back‐translated to a typical scale (e.g. 0 to 10 for pain) by multiplying the SMD by a typical among‐person standard deviation (e.g. the standard deviation of the control group at baseline from the most representative trial) (Schünemann 2017a).

For dichotomous outcomes, including adverse events and return to work or sport, we will present risk ratio (RR) and 95% CI.

In the 'Comments' column of the 'Summary of findings' table, we will provide the absolute per cent difference, the relative per cent change from baseline, and for outcomes that show a clinically important difference between treatment groups, the number needed‐to‐treat for an additional beneficial outcome (NNTB), or number needed‐to‐treat for an additional harmful outcome (NNTH).

For dichotomous outcomes, we will calculate the NNTB or NNTH from the control group event rate and the relative risk using the Visual Rx NNT calculator (Cates 2008). The NNTB for continuous measures will be calculated using the Wells calculator (available at the CMSG Editorial office, musculoskeletal.cochrane.org/). A minimal clinical important difference of 1.5 points on 10‐point continuous visual analogue (VAS) pain scale; 10 points on 100‐point function and quality of life scales (van der Roer 2006) will be used for input into the Wells calculator.

For dichotomous outcomes, we will calculate the absolute risk difference using the risk difference statistic (RD) in RevMan and the result will be expressed as a percentage. For continuous outcomes, we will calculate the absolute benefit as the improvement in the intervention group minus the improvement in the control group (MD), in the original units and expressed as a percentage.

The relative percent change for dichotomous data will be calculated as the risk ratio ‐ 1 and expressed as a percentage. For continuous outcomes, we will calculate the relative difference in the change from baseline as the absolute benefit divided by the baseline mean of the control group, expressed as a percentage.

Unit of analysis issues

We do not expect to retrieve cluster‐randomised trials or cross‐over trials for this population, however repeated observations are expected in the included trials. To address the issue of repeated observations, we will collapse the analysis into short‐term (immediately after the intervention), medium term (six months after randomisation), and long term (i.e. 12 months or more) effects. In the case of trials that collected data at several time points within each category, we will use the data from the closest time point.

Where multiple trial arms are reported in a single trial, we will include only the relevant arms. If two comparisons (e.g. exercise A versus placebo and exercise B versus placebo) are combined in the same meta‐analysis, we will halve the control group to avoid double‐counting.

Dealing with missing data

Firstly, review authors will contact study authors by email requesting any necessary data that is not reported in the manuscript. In cases where data are reported as a median and interquartile range (IQR), it will be assumed that the median is equivalent to the mean and the width of the IQR is equivalent to 1.35 times the standard deviation (SD) (Higgins 2011b). Data can be also estimated from graphs in the case that this information is not presented in tables or text. If any information regarding SDs is missing, we will calculate them from CIs or standard errors (SEs), if available, of the same study. Finally, if no measure of variability is presented anywhere in the text, we will estimate the SD from another trial in the meta‐analysis; we will preferentially use the trial with lowest risk of bias to estimate this SD.

Assessment of heterogeneity

Assessment of heterogeneity will be based upon visual inspections of the forest plot (e.g. overlapping confidence intervals) and more formally on the Chi² test and the I² statistic recommended by the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011b).

As recommended in the Chapter 9, Cochrane Handbook for Systematic Reviews of Interventions (Deeks 2017), the interpretation of an I² value of 0% to 40% might 'not be important'; 30% to 60% may represent 'moderate' heterogeneity; 50% to 90% may represent 'substantial' heterogeneity; and 75% to 100% represents 'considerable' heterogeneity. As noted in the Cochrane Handbook for Systematic Reviews of Interventions, we will keep in mind that the importance of an I² depends on (i) magnitude and direction of effects and (ii) strength of evidence for heterogeneity.

The Chi² test will be interpreted where a P value ≤ 0.10 will indicate evidence of statistical heterogeneity.

If we identify substantial heterogeneity, we will report it and investigate possible causes by following the recommendations in section 9.6 of the Cochrane Handbook for Systematic Reviews of Interventions .

We will combine results in a meta‐analysis using a random‐effects model if an I² is < 50%. If substantial heterogeneity is present (i.e. I² > 50%), we will not combine the results but instead present them as a narrative synthesis.

Assessment of reporting biases

We will generate funnel plots (if we retrieve at least 10 trials) in order to determine possible reporting biases, i.e. small‐study effects (Sterne 2017).

To assess outcome reporting bias, we will check trial protocols against published reports. For studies published after 1st July 2005, we will screen the Clinical Trial Register at the International Clinical Trials Registry Platform of the World Health Organization (http://apps.who.int/trialssearch) and clinicaltrials.gov for the a priori trial protocol.

Data synthesis

We will combine the results from individual trials if possible via meta‐analysis. This pooling of the data (if applicable) will be dependent on the level of diversity of the retrieved studies and if the trials are similar enough in terms of participants and interventions. Results will be combined in a meta‐analysis using a random‐effects model. If an I² > 90% heterogeneity is present, the results will not be combined but presented as a narrative synthesis.

Subgroup analysis and investigation of heterogeneity

We will stratify some of the analysis and create separate comparisons based upon a number of subgroups if needed (Higgins 2011b), these will include the following.

1. Duration of symptoms: if there are sufficient continuous data from at least 10 studies, we will consider meta‐regression to assess if symptom duration modifies the effect of the intervention.

2. Population (i.e. athletes and non‐athletes (defined by the authors of the manuscripts); adults (aged 18 or over), adolescents (aged 13 to 17 year old) children (aged 12 or younger).

These subgroup analyses will be performed only for major outcomes (e.g. pain intensity).

Sensitivity analysis

We plan to perform sensitivity analyses, regardless of expected number of studies. We plan to include studies at low risk of selection and detection bias in a sensitivity analysis, using the pain outcome.