Types of studies

We will include:

• randomised controlled trials (RCTs), including cluster‐RCTs;

• quasi‐randomised trials;

• controlled before‐and‐after (CBA) studies; and

• interrupted time series (ITS) studies and repeated measures studies (RMS).

We will only include CBA, ITS and RMS studies if they fulfil the criteria of the Cochrane Effective Practice of Care (EPOC) Group: CBA studies must have at least two intervention sites and two control sites, ITS and RMS studies must have a defined point in time when the intervention has been started, and at least three data points before and three after it's start (EPOC 2013).

We will exclude historical controlled studies.

Types of participants

Studies are eligible for inclusion irrespective of participants' age, gender, socioeconomic status, ethnicity or nationality. For a study to be included, its participants must have been exposed to an intervention involving a sporting organisation, as either a setting, or a means of access to individuals. A study participant may have been an active or passive sports club member, coach, supporter, student, or any other individual.

Types of interventions

We will consider any intervention that is implemented within a sporting organisation or using a sporting organisation for access to a target population. Control interventions will be 'no intervention' or usual practice. Interventions may occur within a sporting organisation (e.g. implementation of a minimum age for participation in the Olympic Games) or outside the immediate boundaries of the setting (e.g. using the celebrity of a sports club or of an athlete to convey a health message) (Dvorak 2012; Irwin 2010). For the purpose of this review, we will define sporting organisations as “social entities involved in the sport industry that are goal‐directed, with a consciously structured activity system and a relatively identifiable boundary“ (Slack 1998). Sporting organisations may be local (e.g. local swimming club, gym), national (e.g. national football league), or international (e.g. International Association of Athletics Federations).

Eligible interventions include but are not limited to:

• mass media campaigns;

• information or educational sessions;

• management or organisational change strategies;

• policy changes, for example, targeting the sociocultural environment to encourage a specific group of people to participate;

• changes to traditional or existing programmes, for example, club or association‐initiated rule modification programmes;

• provision of activities beyond traditional or existing programmes, for example, 'Come and Try' initiatives, skill improvement programmes, volunteer encouragement programmes; and

• interventions using role models defined as “an individual who is perceived as exemplary, or worthy of imitation or who inspires individuals or groups of people through personal contact and relationship” (Payne 2003).

If feasible, we will group the interventions according to the target population as follows:

• interventions targeting the population outside the boundaries of a sporting organisation (e.g. general population or a selected group);

• interventions targeting members of a sporting organisation; and interventions targeting individuals in the setting of a sporting organisation that are not (yet) taking part.

We will exclude studies in which the individual effect of sporting organisations cannot be separated from the effect of other intervention components (e.g. multi‐component interventions), as well as studies implementing interventions across settings.

Further, we will exclude studies with interventions:

• designed specifically for the treatment (or as therapy) for specific medical conditions (e.g. rehabilitation programmes);

• designed specifically to increase paid active or paid non‐active participation;

• surrounding sports injury prevention; and aiming to reduce drug use related to sports (i.e. anti‐doping interventions). Not excluded are policies or practices aiming to reduce alcohol consumption or tobacco use in the setting of sporting organisations.

Types of outcome measures

Electronic searches

We will search the following databases.

• Cochrane Central Register of Controlled Trials (CENTRAL).

• MEDLINE and MEDLINE In‐Process and Other Non‐Indexed Citations (1946 to present).

• EMBASE (1974 to present).

• PsycINFO (1806 to present).

• CINAHL (1981 to present).

• SPORTDiscus (1973 to present).

• Dissertation Abstracts (1997 to present).

• Sociological Abstracts (1952 to present).

The search strategy in Appendix 1 has been developed for MEDLINE (Ovid) based on those of both previous Cochrane reviews (Priest 2008a; Priest 2008b). We combined the intervention search terms for policy interventions and non‐policy level health promotion interventions and complemented them with additional search terms for eligible interventions. We extended the list of search terms for sports and physical activity as well as for settings. Finally, we broadened the terms for study design. For the databases listed above, we will adapt the search strategy to their respective syntaxes.

We will search electronic databases of relevance to health promotion and public health, with a simplified search strategy (sporting organisation*) of the following.

• BiblioMap, the EPPI (Evidence for Policy and Practice Information)‐Centre database of health promotion research (http://eppi.ioe.ac.uk/webdatabases/Intro.aspx?ID=7).

• The Health Technology Assessment Database, accessible through the Cochrane Library (http://onlinelibrary.wiley.com/cochranelibrary/search).

• The Health Evidence Bulletins, Wales (http://hebw.cf.ac.uk/).

• The Effective Public Health Practice Project (http://www.ephpp.ca/systematicreviews.html).

• DECIPHer: Centre for the Development and Evaluation of Complex Interventions for Public Health Improvement (http://decipher.uk.net/).

• National Institute for Health and Clinical Excellence (http://www.nice.org.uk/).

• The Community Guide ‐ Guide to Community Preventive Services ‐ Systematic reviews and evidence‐based recommendations (http://www.thecommunityguide.org/index.html).

• C2‐SPECTR, the social, psychological, educational, and criminological trials register of the Campbell Collaboration (http://www.campbellcollaboration.org/lib/?go=monograph).

• Leisure Information Network website (http://lin.ca/) using the National Recreation Database (Canada).

• The World Health Organization (WHO) International Clinical Trials Registry (http://apps.who.int/trialsearch/).

• The registry ClinicalTrials.gov (https://clinicaltrials.gov/)

• The WHO European Database on Nutrition, Obesity and Physical activity (NOPA) (http://data.euro.who.int/nopa/).

We will also conduct an Internet search in English using the Google search engine (http://www.google.ch/) and Google Scholar (http://scholar.google.com/). We will use the following keywords or combinations thereof to locate potentially eligible studies: sports club, sporting organisation, health promotion, intervention, policy. Given the decrease in relevance in results listed by both search engines, we will only consider the first 500 hits for each.

Searching other resources

We will contact individuals as well as national and international bodies involved in sports (e.g. International Olympic Committee, European Athletics Association, Swiss Federal Office of Sport) to identify unpublished or internal reports and conference proceedings. Further, we will check the reference lists of all included studies and relevant background literature, and contact authors of included studies to identify additional published and unpublished studies.

Selection of studies

Two reviewers (AF, NP) will screen titles and abstracts of all search entries. We will retrieve full‐texts for search entries for which we cannot make a judgement of eligibility based on title and abstract alone. One reviewer (AF) will retrieve the full‐text of all potentially eligible studies, and two reviewers (AF, NP) will independently decide on their final eligibility against the inclusion criteria. If there is any disagreement or uncertainty, a third reviewer will be consulted (EvE). We will only include each study once, even if it has been reported in more than one publication. We will consider the largest or most recent data set and decide on a case‐by‐case basis. We will record the selection process in sufficient detail to complete a PRISMA flow diagram (Moher 2009), and 'Characteristics of excluded studies' table. We will not impose any language restrictions.

Data extraction and management

Two reviewers (AF, NP) will extract the data independently. A third reviewer (EvE) will resolve any discrepancies.

For the included studies, we will use a piloted and adapted version of the Cochrane Public Health data extraction template to extract data on:

• study characteristics: country, study design, funding source and contextual factors when available;

• sporting organisation characteristics: name, type of sport promoted, its geographic area of activity (local, national or international) and location;

• participant characteristics: age, gender, ethnicity, type of recruitment, relationship to the sporting organisation (active, passive or no member, supporter or relative of a member, other);

• intervention characteristics: name of the programme, content of the intervention, qualification and training of staff that deliver the intervention, mode of delivery (individual‐ or group‐level, interactive or passive message delivery), setting of intervention delivery, duration and intensity of intervention, components used to promote participation (such as incentives);

• outcome definitions and time points of outcome measurement. We will group outcomes in short‐term (< 6 months after intervention has been completed) and long‐term outcomes (≥ 6 months after intervention has been completed);

• outcomes data;

• dropout/adherence rates;

• factors related to implementation of the intervention: involvement of stakeholders, pilot‐testing of the intervention, modification of the primary intervention, intervention’s reach, completeness of implementation of the intervention and its maintenance after the evaluation period has ended; and financial costs of the intervention.

Further, we will use the Equity checklist of the Campbell and Cochrane Equity Methods Group to evaluate if the included studies did address health equity issues (Ueffing 2012), such as conducting analyses by socioeconomic or sociocultural subgroups such as income, occupation or ethnicity.

Assessment of risk of bias in included studies

Two reviewers (AF, NP) will independently assess the quality of each included study. We will assess the risk of bias using the Cochrane Effective Practice and Organisation of Care (EPOC) 'Risk of bias' tool (EPOC 2015). According to this tool we will assess the following items for randomised controlled trials, non‐randomised controlled trials and controlled before‐after studies: sequence generation, allocation concealment, blinding of outcome assessment, completeness of outcome data, selective outcome reporting, baseline characteristics, baseline outcome measurements, protection against contamination, and other sources of bias. For interrupted time series studies we will assess independence of intervention, blinding, selective reporting, other bias, completeness of outcome data, if the shape of effect was pre‐specified and if intervention was unlikely to affect data collection. We will categorise them as being at low, unclear, or high risk of bias. Disagreements or uncertainties will be settled by consensus, and, where needed, through discussion with a third reviewer (EvE). We will summarise the 'Risk of bias' assessment in a 'Risk of bias' table and graph.

Measures of treatment effect

We will present effects for binary outcomes using risk ratios (RRs) with 95% confidence intervals (CIs). We will present effects for continuous outcomes as the difference in mean values with 95% CI. For continous outcomes we will determine mean change from baseline data. The mean change will be determined by subtracting the final mean from the baseline mean. We will calculate the confidence interval of the mean change by using standard errors, confidence intervals, t values or P values that relate to the differences between means in two groups. If these measures are not available, we will use a correlation coefficient to impute a standard deviation (Higgins 2011). When continuous outcomes are measured with different scales, we will use the standardised mean difference (SMD) to measure the intervention effect.

Unit of analysis issues

We will collect data from studies that allocate individuals or groups to an experimental or control intervention. We will combine data from studies with a cluster design with other data only if the clustering has been accounted for. If it has not been accounted for in analyses, but the trial data are available from the authors, we will aim to re‐analyse relevant data. For cluster‐RCTs, we will adjust the results by using intra‐class correlation coefficients.

Dealing with missing data

We will assess missing data and dropouts in the included studies. We will report numbers, characteristics and reasons for dropout. We will contact authors of included studies for further information, if needed.

Assessment of heterogeneity

We are expecting substantial heterogeneity in the results of the studies due to differences in types of study designs, interventions, outcomes, and patient characteristics (clinical heterogeneity). Only if primary outcomes are the same and the characteristics of the studies are sufficiently similar, will we undertake meta‐analyses. For this purpose, we will evaluate the presence of heterogeneity by constructing forest plots and examining them for asymmetry. We will perform Chi² tests and use P ≤ 0.1 as a threshold to reject the null hypothesis of homogeneity. We will calculate the I² statistic to quantify statistical heterogeneity, with I² ≥ 75% indicating considerable heterogeneity. In this case (and according to Higgins 2011), we will refrain from pooling outcome data but will present a narrative summary of the findings instead.

Assessment of reporting biases

We will investigate the presence of reporting bias by plotting a funnel plot and visually assess it for asymmetry and outliers. Thus, we will evaluate the presence of reporting bias or differences in the results between smaller and larger studies. We will employ Egger’s test if there are sufficient studies that report standard errors for effect sizes (Egger 1997).

Data synthesis

The decision on whether to combine the studies quantitatively or describe them qualitatively will be made once we have extracted all relevant information from the included studies. We will examine the characteristics of the interventions and the comparability of the outcome measures. If the interventions or outcome measures are similar, we plan to synthesise the results in meta‐analyses. If data allows we will conduct three different meta‐analyses according to the study population (population outside the boundaries of a sporting organization, members of a sporting organization and individuals being in the setting of a sporting organization but not taking part in it ). We will conduct meta‐analyses only with data from RCTs and cluster‐RCTs. If other types of studies are included, we will summarise their outcome data qualitatively. We will use Review Manager 5 for meta‐analyses and employ the random‐effects model with weighing by inverse variance (RevMan 2014). If we cannot pool outcome data we will employ narrative synthesis techniques (descriptive data synthesis) (Popay 2006). In this case we will summarise the studies by grouping them according to the type of intervention. We will base all data analyses on the intention‐to‐treat principle.

When discussing results, we will make a distinction between statistically and clinically significant results. For non‐clinical outcomes (e.g. health behaviours), we will compare our results to findings in the literature to assess the significance of the changes compared to other studies.

Subgroup analysis and investigation of heterogeneity

To investigate possible sources of heterogeneity, we will conduct subgroup analyses in each of the three study population groups. If the available data allow it, we will examine subgroups defined by the following criteria:

• Type (intervention at population‐level versus intervention at group‐ or individual‐level) and duration of intervention.

• Type of sporting organisation (individual versus team sport, community‐based organisation versus national or international sporting organisations).

• Population subgroups: age groups, gender, ethnicity, socioeconomic status (measured with education, income or occupation, or a combination of them), or geographical location of the intervention.

• Implementation and contextual factors.

Sensitivity analysis

In a sensitivity analysis, we will explore the impact of methodological study quality (risk of bias) on study findings. We will repeat any meta‐analyses while excluding studies with high risk of bias. To explore whether the type of funding source has an effect on estimates of intervention effectiveness, we will repeat them, excluding commercially funded studies.