Types of studies

We will include randomised controlled trials (RCTs), which may be single‐blind or unblinded.

Types of participants

People (adults and children of either gender) who are prescribed drugs known to cause delayed type hypersensitivity reactions. These include antiepileptic drugs, antiretrovirals, antigout drugs, and antibiotics such as beta‐lactams (penicillin, amoxicillin, piperacillin, cephalosporins) and sulphonamides (sulphamethoxazole and trimethoprim).

Types of interventions

We will consider genetic testing for any genetic variants associated with hypersensitivity reactions using all available techniques to determine individual genotypes. The intervention is a randomly allocated genetic test; if the test is positive, a drug that can cause hypersensitivity is avoided.

Types of outcome measures

We will base core outcome measures on several papers describing clinical classification of drug‐induced skin reactions, including the recently published paper entitled 'Phenotype standardisation for immune‐mediated drug‐induced skin injury' (Pirmohamed 2011), as well as earlier papers by the RegiSCAR (European Registry of Severe Cutaneous Adverse Reactions) consortium (Bouvresse 2012; Kardaun 2013; Sekula 2011).

We will assess clinically defined hypersensitivity reaction, immunologically confirmed hypersensitivity reaction (if skin patch testing or lymphocyte proliferation assay data are available), and long‐term sequelae (including ophthalmologic, cutaneous, or liver damage, etc).

We have provided a full list of clinically relevant outcomes and distinction between primary and secondary outcomes in Appendix 1.

Electronic searches

We will search the following databases for relevant trials:

• the Cochrane Skin Group Specialised Register;

• the Cochrane Central Register of Controlled Trials (CENTRAL) in The Cochrane Library;

• MEDLINE via OVID (from 1946);

• EMBASE via OVID (from 1974); and

• LILACS (Latin American and Caribbean Health Science Information database, from 1982).

We have devised a draft search strategy for randomised controlled trials (RCTs) for MEDLINE (OVID), which is displayed in Appendix 2. This will be used as the basis for search strategies for the other databases listed.

Searching other resources

Selection of studies

Two review authors (AA and AJ) will independently assess studies for inclusion; they will independently screen all the titles and abstracts of publications identified by the searches to assess their eligibility. We will assess the full text of eligible citations for inclusion. We will exclude publications that do not meet the criteria at this stage and prepare a table of 'Characteristics of excluded studies' to clearly differentiate between those studies that are not at all relevant and those that may not fulfil the criteria for inclusion but may be considered relevant by some readers. We will reach consensus on the selection of trials and the final list of studies.

Data extraction and management

Two authors (AA and AJ) will independently extract data from included studies and resolve disagreements by discussion. If consensus is not reached, they will consult a third author (MP). We will collect the following information on study characteristics and methods: study design; inclusion and exclusion criteria; setting; country; language of publication; ethnicity of participants; control population (exposed to the culprit drug or healthy population); description of genotyping techniques used; genotyping quality control, which will include deviation from Hardy‐Weinberg equilibrium (Hardy‐Weinberg equilibrium is a crucial concept in population genetics; it predicts how gene frequencies will be inherited from generation to generation and is used as a measure of quality of genetic tests) and genotype call rate; age; gender; concomitant medications; time from exposure to culprit drug to skin reaction; type of skin reaction; location of skin lesion; duration; treatment; sequelae; other manifestations indicating systemic involvement; and laboratory tests.

Assessment of risk of bias in included studies

All review authors will assess the risk of bias of three studies as a pilot to ensure we are using consistent methods. AA and AJ will then independently assess the risk of bias in each trial according to the approaches described in theCochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). Specifically, we will use The Cochrane Collaboration's tool for assessing risk of bias (Table 8.5.a in the Cochrane Handbook for Systematic Reviews of Interventions), which is based on seven domains (sequence, generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective outcome reporting, and other issues). The tool allows for the risk of bias to be assessed as 'low', 'high', or 'unclear' (indicating lack of information or uncertainty over the potential for bias). An additional author (MP) will assess any disagreements.

Measures of treatment effect

We will use statistical methods in accordance with theCochrane Handbook for Systematic Reviews of Interventions (Higgins 2011) to measure treatment effect. We will use mean difference (MD) with 95% confidence interval (CI) for continuous data or indeed standardised mean difference (SMD) with 95% CI where all studies report an outcome using similar scales. We will use risk ratio (RR) with 95% CI for dichotomous data.

Unit of analysis issues

We will take care to avoid a unit of analysis error due to repeated observations on participants, multiple treatments, or re‐occurring events. We will also consider issues in cluster randomised trials, such as recruitment bias, baseline comparability of clusters, and number of clusters, and make sure that appropriate statistical methods are used that take into account weighting, etc.

Dealing with missing data

We will consider the possible different types of missing data. We will deal with missing studies and the associated risk of bias by assessing for publication bias, whilst we will deal with missing outcomes and the associated risk of bias by assessing for selective reporting (see Assessment of reporting biases section).

In the event of missing summary data, we will contact the study author if data required to calculate outcomes of interest are missing. If no further data are available, we will not include the study in the meta‐analysis. However, we will report the limited results from the study in narrative form in the results section and consider whether they are consistent with the results of the meta‐analysis for that outcome. Also, where appropriate, we will make assumptions about the missing data (e.g. assuming all missing values to have a particular value, e.g. an adverse event) and conduct sensitivity analyses to test how sensitive the analyses are to our assumptions. We will address the potential implications of the missing data in the Discussion section.

Assessment of heterogeneity

We will assess the extent of heterogeneity using the I² statistic.

We will use the following thresholds for the interpretation of the I² statistic:

• 0% to 40% = might not be important;

• 30% to 60% = moderate heterogeneity;

• 50% to 90% = may represent substantial heterogeneity; and

• 75% to 100% = considerable heterogeneity.

If we do a meta‐analysis, we shall report the I² statistic and interpret the two data together.

Assessment of reporting biases

We will assess biases, including publication bias, 'time‐lag' bias, outcome reporting bias, language bias, and citation bias. We will use a funnel plot, Begg test, and Egger test to evaluate publication bias (Begg 1989; Egger 1998). We will only carry out tests for funnel plot asymmetry when there are at least 10 studies included in the meta‐analysis, because when there are fewer studies, the power of the tests is too low to distinguish chance from real asymmetry.

Data synthesis

If there is no significant clinical heterogeneity, we will synthesise the results in meta‐analyses. We will synthesise data according to type of intervention (e.g. genotype test). We will use a random‐effects model.

Where events are rare, a random‐effects approach may be inappropriate. Where events are rare, extra care will be taken to adopt appropriate methods of meta‐analysis as recommended in the Cochrane Handbook for Systematic Reviews of Interventions (section 16.9), since many meta‐analysis methods are suboptimal where events are rare through results being biased, confidence intervals being too wide, or power being too low. Choice of method will be guided by control group risk, likely treatment effect size, and consideration of balance in numbers of treated and control participants in the constituent studies. Where the control groups differ, e.g. they are drawn from a healthy population or from a population of people treated with the culprit drug but without any adverse effects, we will conduct separate analysis. We will use Review Manager to undertake the meta‐analyses.

Subgroup analysis and investigation of heterogeneity

Where there is substantial heterogeneity, we will explore the causes by way of subgroup analyses. Indeed, some HLA genetic variants are associated with a high risk of drug‐induced skin reactions, so interventions would only be given to the appropriate subpopulation.

We will consider the following:

• participant factors (age, ethnicity, ADR classification, and comparability of participant groups); and

• trial design issues (genotyping methodology and quality control, blinding, drugs included, and drug dosage and duration of use).

Sensitivity analysis

We will evaluate the robustness of the results of the meta‐analyses by removing trials of low methodological quality as defined by their risk of bias.