Types of studies

All randomised controlled trials comparing the comparative effectiveness and/or safety of one of the interventions compared to the other intervention. We will exclude quasi‐randomised and non‐randomised studies. Cross‐over trials will be included, but only data from the first phase will be used.

Types of participants

Couples who have been trying to conceive for at least one year, the woman having at least one patent fallopian tube, an ovulatory cycle and no or mild endometriosis (American Fertility Society (AFS) criteria I) and the man having a prewash total motile sperm count > 3 * 10ˆ6.

Types of interventions

We will consider all trials where one of these interventions is compared.

• Expectant management.

• Timed intercourse.

• Ovarian stimulation using gonadotropins, aromatase‐inhibitors or anti‐estrogens.

• Intrauterine insemination (IUI) without ovarian stimulation.

• IUI with ovarian stimulation.

• In vitro fertilisation (IVF) with either a single embryo transfer, dual embryos transferred, in a modified natural cycle or combined with intracytoplasmic injection (ICSI).

The interventions of expectant management and timed intercourse will be combined, if no invasive techniques are used.

The reported interventions will be compared to each other or to no intervention (i.e. expectant management).

Types of outcome measures

Electronic searches

We will search the following electronic databases for relevant trials from inception onwards.

• The Cochrane Gynaecology and Fertility Group (CGF) specialised register of controlled trials (Procite platform) (Appendix 1).

• The Cochrane Central Register of Studies Online (CRSO Web platform) (Appendix 2).

• MEDLINE (Ovid platform) (Appendix 3).

• Embase (Ovid platform) (Appendix 4).

• PsycINFO (Ovid platform) (Appendix 5).

• CINAHL (Ebsco platform) (Appendix 6).

The MEDLINE search will be combined with the Cochrane highly sensitive search strategy for identifying randomized trials, which appears in the Cochrane Handbook for Systematic Reviews of Interventions (Version 5.1.0, chapter 6, 6.4.11). The Embase, PsycINFO and CINAHL searches will be combined with trial filters developed by the Scottish Intercollegiate Guidelines Network (www.sign.ac.uk/methodology/filters.html#random).

Other electronic sources of trials will include:

• trial registers for ongoing and registered trials:

  • www.clinicaltrials.gov (a service of the US National Institutes of Health);

  • www.who.int/trialsearch/Default.aspx (The World Health Organisation International Trials Registry Platform search portal);

• the Virtual Health Library Regional Portal (VHL) (bvsalud.org/portal/?lang=en) which includes LILACS; and

• PubMed and Google Scholar (for recent trials not yet indexed in the major databases).

Searching other resources

We will handsearch reference lists of relevant trials and systematic reviews retrieved by the search and contact experts in the field to obtain additional data. We will also handsearch relevant journals and conference abstracts that are not covered in the CGFG register, in liaison with the Information Specialist.

Selection of studies

Two investigators (RT and RW) will independently assess trial eligibility, according to the Criteria for considering studies for this review. We will resolve disagreements through discussion with a third investigator (MW). A PRISMA flow diagram will be drawn to show the results of the search and the number of included and excluded trials. The reasons for excluding any potentially‐eligible studies identified by the search from the (network) meta‐analysis will be documented.

Data extraction and management

For all included trials two authors (RT and RW) will independently extract data using a data abstraction form and summarise trial characteristics in tables. From each included study we will extract baseline characteristics of the couples (i.e. female age, duration of subfertility, body mass index, prior treatment), study settings, methods, the types of interventions (used dose, type of preparation, regimens, co‐interventions) and the outcomes. Where studies have multiple publications the authors will collate multiple reports of the same study under a single study identifier with multiple references. We will correspond with study investigators for further data on methods and results, as required.

Assessment of risk of bias in included studies

Two authors (RT and RW) will independently assess the risk of bias for each eligible study by using the Cochrane 'Risk of bias' assessment tool (Higgins 2011) which includes six domains: selection (random sequence generation and allocation concealment); performance (blinding of participants and personnel); detection (blinding of outcome assessors); attrition (incomplete outcome data); reporting (selective reporting); and other bias. Disagreements will be resolved by discussion with a third investigator (MW). We will describe all judgements fully and present the conclusions in the 'Risk of bias' table, which will be incorporated into the interpretations of review findings by means of sensitivity analyses. With respect to within‐trial selective reporting, where identified studies fail to report the primary outcome of live birth, but do report interim outcomes such as pregnancy, we will assess whether the interim values are similar to those reported in studies that also report live birth.

Measures of treatment effect

For dichotomous data (e.g. live birth rates), we will use the numbers of events in the control and intervention groups of each study to calculate Mantel‐Haenszel odds ratios (ORs). We will present 95% confidence intervals for all outcomes. Where data to calculate ORs are not available, we will utilise the most detailed numerical data available that may facilitate similar analyses of included studies (e.g. test statistics, P values). We will assess whether the estimates calculated in the review for individual studies are compatible in each case with the estimates reported in the study publications. When more than two studies compared the same treatments, a random‐effects pooled OR will be calculated. The random‐effects model incorporates the between study variability and is more conservative than the fixed‐effect model.

For each pairwise comparison we will present a 95% predictive interval; this can be interpreted as the 95% interval of the expected treatment effect in a new trial with this comparison (Salanti 2011). Furthermore we will calculate the probability that an intervention is ranked first, second and so on. We will display this ranking graphically for the primary and secondary outcomes using the surface under the cumulative ranking (SUCRA), where the SUCRA values can range from zero (i.e. the intervention is certain to be the worst) to one (i.e. the intervention is certain to be the best).

Unit of analysis issues

The primary analysis will be cumulative rates of each outcome per woman randomised. Data that do not allow valid analysis (e.g. 'per cycle' data) will be briefly summarised in an additional table and will not be meta‐analysed. Multiple births will be counted as one live birth event. Only first‐phase data from cross‐over trials will be included.

Dealing with missing data

We will analyse the data on an intention‐to‐treat basis as far as possible (i.e. including all randomised participants in the analysis, in the groups to which they were randomised). Attempts will be made to obtain missing data from the original trialists. Where data are unobtainable, we will undertake imputation of individual values only for the primary outcome of live birth or ongoing pregnancy: an event will be assumed not to have occurred in participants without a reported outcome. For other outcomes, we will analyse only the available data. Any imputation undertaken will be subjected to sensitivity analysis.

Assessment of heterogeneity

Assessment of reporting biases

In view of the difficulty of detecting and correcting for publication bias and other reporting biases, we will aim to minimise their potential impact by ensuring a comprehensive search for eligible studies and by being alert for duplication of data. If there are ten or more studies in an analysis, we will use a comparison‐adjusted funnel plot to explore the possibility of small study effects (a tendency for estimates of the intervention effect to be more beneficial in smaller studies) (Chaimani 2013).

Data synthesis

We will compare interventions using ORs with their respective 95% confidence intervals. When more than two studies compared the same treatments, a random‐effects summary OR will be calculated.

We will conduct a network meta‐analysis based on all investigated comparisons between treatments and the indirect analysis can be performed utilising all the possible pathways provided by the network. An indirect estimate of A versus B can be calculated by comparing direct comparisons of A versus C with trials of B versus C. In this way the OR for comparing A and B can be calculated using the following principle: ln(ORAvsB) = ln(ORAvsC) − ln(ORBvsC). The direct and indirect evidence will be combined for each comparison using the abovementioned analysis for direct and indirect comparisons. We will use STATA for the analyses.

Subgroup analysis and investigation of heterogeneity

We will assess subgroup differences by interaction tests. We will report the results of subgroup analysis quoting the Chi² statistic and P value, and the interaction test I² value. If we detect substantial heterogeneity, we will explore possible explanations in subgroup analyses (e.g. differing populations) and/or sensitivity analyses (e.g. differing risk of bias). We will take any statistical heterogeneity into account when interpreting the results, especially if there is any variation in the direction of effect.

Where data are available from at least two studies, we will conduct subgroup analyses for the primary outcomes only to determine the separate evidence within the following subgroups.

1. Younger women (=< 38 years) versus older women (> 38 years).

2. Treatment naïve couples versus couples who have received prior treatment.

3. Short duration of subfertility (<= 2 years) versus long duration of subfertility (> 2 years).

4. IVF with single embryo transfer versus IVF with dual or more embryo transfer.

5. IUI with follicle stimulating hormone (FSH) versus IUI with clomiphene citrate (CC).

Sensitivity analysis

We will conduct sensitivity analyses for the primary outcomes to determine whether the conclusions are robust to arbitrary decisions made regarding the eligibility and analysis. These analyses will include consideration of whether the review conclusions would have differed if:

1. eligibility had been restricted to studies with no domains at high risk of bias;

2. alternative imputation strategies had been implemented;

3. eligibility had varied by publication type (abstract versus full text); or

4. we had included only studies with the outcome live birth.