Types of studies

Randomised controlled trials (RCTs) irrespective of the randomisation method and blinding procedure used.

Types of participants

Children up to age 16 years with rAOM, defined as three or more episodes in six months, or four in one year with one episode in the last six months (Goycoolea 1991). We will include all children whether or not there is evidence of a middle ear effusion at the time of study enrollment.

Types of interventions

Types of outcome measures

We will analyse the following outcomes in the review, but we will not use them as a basis for including or excluding studies.

Electronic searches

The TSC will search:

• the Cochrane ENT Trials Register (searched via the Cochrane Register of Studies);

• the Cochrane Register of Studies Online (search to date);

• Ovid MEDLINE (1946 to date):

  • Ovid MEDLINE (In‐Process & Other Non‐Indexed Citations);

  • PubMed (as a top up to searches in Ovid MEDLINE);

• Ovid EMBASE (1974 to date);

• Ovid CAB Abstracts (1910 to 2014 week 43);

• EBSCO CINAHL (1982 to date);

• LILACS, lilacs.bvsalud.org;

• KoreaMed (via Google Scholar);

• IndMed, www.indmed.nic.in;

• PakMediNet, www.pakmedinet.com;

• Web of Knowledge, Web of Science (1945 to date);

• ClinicalTrials.gov (via the Cochrane Register of Studies);

• ICTRP, www.who.int/ictrp;

• Google Scholar, scholar.google.co.uk;

• Google, www.google.com.

The subject strategies for databases will be modelled on the search strategy designed for CENTRAL (Appendix 1). Where appropriate, these will be combined with subject strategy adaptations of the highly sensitive search strategy designed by Cochrane for identifying randomised controlled trials and controlled clinical trials (as described in the Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0, Box 6.4.b. (Handbook 2011).

Searching other resources

We will scan the reference lists of identified publications for additional trials and contact trial authors if necessary. In addition, the Trials Search Co‐ordinator will search PubMed, TRIPdatabase, The Cochrane Library and Google to retrieve existing systematic reviews relevant to this systematic review, so that we can scan their reference lists for additional trials. We will search for conference abstracts using the Cochrane ENT Trials Register and EMBASE.

Selection of studies

Two review authors (LL and PTM) will independently screen titles and abstracts obtained from the database searches and the reference lists of relevant systematic reviews to assess their potential relevance for reviewing the full text. The same review authors will independently review the full text of potentially relevant articles against the inclusion and exclusion criteria. We will resolve any disagreements by discussion with a third review author (RPV).

Data extraction and management

Two review authors (LL and PTM) will independently extract data from the included studies using standardised data extraction forms. We will extract the following data from each study:

• Trial characteristics: setting, design, method of data analysis.

• Participants: study population, number of children in each group, participant characteristics such as age and gender.

• Interventions: type of surgery including pre‐operative, intra‐operative and postoperative treatment.

• Outcomes: primary and secondary outcomes recorded, time points, adverse effects and complications related to the intervention and comparators.

• Aspects of methodology relating to risk of bias (see below).

We will also extract the following summary statistics for each trial and each outcome:

• For continuous data: the mean values, standard deviations and number of patients for each treatment group. Where endpoint data are not available, we will extract the values for change from baseline.

• For binary data: the numbers of participants experiencing an event and the number of patients assessed at the time point.

• For ordinal scale data: if the data appear to be approximately normally distributed or if the analysis that the investigators performed suggests parametric tests were appropriate, then we will treat the outcome measures as continuous data. Alternatively, if data are available, we may convert them into binary data.

We have prespecified the time points of interest for the outcomes in this review. While studies may report data at multiple time points, we will only extract the longest available data within the time points of interest. For example, for 'intermediate‐term' follow‐up periods, our time point is defined as 'three to six months' post‐randomisation. If a study has reported data at three, four and six months, we will only extract and analyse the data for the six‐month follow‐up.

We will contact the authors of the original publication in case of missing data and we will resolve any discrepancies by discussion. Where a study has more than one publication, we will retrieve all publications to ensure complete extraction of data. If extracted data are suitable for pooled analysis two authors (PTM and RVP) will conduct this.

Assessment of risk of bias in included studies

Three review authors (AGMS, RPV and DAN) will independently assess the methodological quality of the included studies and resolve any disagreements by majority opinion. Guided by the Cochrane Handbook for Systematic Reviews of Interventions (Handbook 2011), we will take the following items into consideration:

• random sequence generation (selection bias);

• allocation concealment (selection bias);

• blinding of participants and personnel (performance bias);

• blinding of outcome assessment (detection bias);

• incomplete outcome data (attrition bias);

• selective reporting (reporting); and

• other sources of bias.

We will present the results of the 'Risk of bias' assessment in a 'Risk of bias' graph and summary figure.

Measures of treatment effect

We will express the pooled measures of treatment effect for dichotomous outcomes as risk ratio (RR) with accompanying 95% confidence intervals (CI). For the key outcomes that we will present in the 'Summary of findings' table, we will also express the results as absolute numbers based on the pooled results and compared to the assumed risk. We may also calculate the number needed to treat to benefit (NNTB) using the pooled results.

We will express continuous outcome variables either as a mean difference (MD) with 95% CIs, if reported on the same scale, or as a standardised mean difference (SMD) with 95% CIs, if different continuous scales were used.

Unit of analysis issues

This review will not use data from phase II of cross‐over studies or from studies where the patient is not the unit of randomisation, i.e. cluster‐randomised trials or studies where 'ears' (right versus left) were randomised.

Dealing with missing data

We will contact trial authors in order to obtain additional information in case of missing data. We will only analyse the available data based on the intention‐to‐treat (ITT) principle. For continuous outcomes, we will calculate missing statistics, such as standard deviations (SDs), from other available statistics (e.g. P values) according the methods described in Chapter 7 of the Cochrane Handbook for Systematic Reviews of Interventions (Handbook 2011).

Apart from imputations for missing standard deviations, we will conduct no other imputations. We will extract and analyse all data using the available case analysis method.

Assessment of heterogeneity

We will primarily assess the level of clinical diversity between trials by reviewing them for potential differences in the types of participants recruited, interventions used and outcomes measured. Secondly, we will assess statistical heterogeneity for each outcome by using the Chi² test, with a significance level set at P value < 0.10, and the I² statistic, with I² values over 50% suggesting substantial heterogeneity (Higgins 2003).

Assessment of reporting biases

We will search the internet, ClinicalTrials.gov (http://clinicaltrials.gov/) and the World Health Organization (WHO) International Clinical Trials Registry Platform (ICTRP) (http://www.who.int/trialsearch) for available study protocols to determine whether outcomes reported were pre‐defined and whether all outcomes listed in the study protocol were reported in the trial publications. If there are sufficient trials, we propose a more formal method of assessing reporting bias, i.e. by using funnel plots

Data synthesis

We will analyse the data according to the intention‐to‐treat principle, i.e. by analysing all participants in the groups to which they were originally randomised. Where data are missing or there was loss to follow‐up, we will conduct an available case analysis.

We will consider pooling the results for meta‐analysis using Review Manager 5.3 (RevMan 2014), in the absence of significant clinical diversity, i.e. where we judge participants, interventions, comparisons and outcomes across trials to be sufficiently similar. We will calculate treatment differences with the Mantel‐Haenszel method using a fixed‐effect model where no substantial statistical heterogeneity is present (I² < 50%). If substantial statistical heterogeneity is detected but unresolved by sensitivity analysis and pre‐specified subgroup analyses, we will calculate treatment differences with the Mantel‐Haenszel method using a random‐effects (DerSimonian and Laird) model to provide a more conservative effect estimate. For dichotomous outcomes, we will calculate the NNTB using the results of the meta‐analysis (which itself uses risk ratio) based on the average risk of the control groups in the included studies ('study population') (Handbook 2011).

Subgroup analysis and investigation of heterogeneity

If possible, we will perform pre‐planned subgroup analyses even if statistical heterogeneity is not observed. We have planned these analyses as the factors indicated are suspected to be potential effect modifiers. They include:

• type of surgery (grommets only versus grommets and concurrent adenoidectomy);

• presence of middle ear effusion at randomisation or at the time of surgery (yes versus no).

In addition to the subgroups above, we will conduct the following subgroup analysis in the presence of statistical heterogeneity:

• age (below two years of age versus two years and older);

• type of grommet (short‐term versus intermediate/long‐term).

Sensitivity analysis

To assess the robustness of the review findings, we will perform a sensitivity analysis in which RCTs with a high risk of bias are excluded. We will define high risk of bias as high risk of allocation concealment bias or attrition bias (overall loss to follow‐up of more than 20% or differential follow‐up observed, or both). In further sensitivity analysis, we propose to calculate the effects of grommets by excluding any AOM recurrences that occurred between the date of randomisation and the date of surgery.