A good question is the first step in any review. Your systematic review must be based on a research question that:

1. Is focused and answerable
2. Adds new research on a topic
3. Considers the right study designs

Focused & Answerable

Use a model question framework such as PICO to identify key concepts and clearly state your question. In addition, gathering a gold set of exemplar articles from preliminary literature searching will provide examples of the literature you plan to synthesize and proof that your question is answerable.

Visit formulating a research question in the Research Toolkit for detailed guidance on PICO, question frameworks and building a gold set.

This process helps you identify concepts and terms that will be used to develop a search strategy.

Adds new research on a topic

Conduct a preliminary literature search to ensure you are filling a gap in the evidence. For projects involving evidence synthesis such as systematic reviews and meta-analysis the aim is to create new knowledge by increasing the power of the analysis. 

If a review does not create new knowledge - you may inadvertently spend a great deal of time achieving what Chevret, Ferguson, and Bellomo (2018) concluded…  “Are systematic reviews and meta-analyses still useful research? No".

Study designs

Consider which study designs are best matched to your question. Different types of clinical questions are best answered by the following study designs.

• Therapy - determining the effect of interventions = RCT, Systematic reviews, Meta analysis.
• Prognosis - likely outcome for a patient population = RCT, Cohort study, Case controlled study.
• Harm - outcomes from an activity of exposure = RCT, Cohort study, Case controlled study.
• Diagnosis - accuracy of diagnostic tests = Cross sectional analytic study, blind comparison to a gold standard.
• Prevention - reducing chance of disease by identifying risk factors = RCT, Systematic review.

Inclusion and exclusion criteria

Inclusion and exclusion criteria sets boundaries for your review by determining which studies will be selected for evidence synthesis. This eligibility cannot be decided as you go. A key feature of systematic reviews is pre-defined inclusion and exclusion criteria.

Inclusion criteria indicates what you want to investigate
Exclusion criteria indicates what does not need to be included

This criteria is determined after developing the research question and before the final search is conducted, although it should be informed by test literature searches and background research.

Clear eligibility criteria will make it easier for the team to work through the screening process.

It is useful to build a 'gold set' of relevant papers as you plan your review. Also known as a 'sample set', a gold set refers to a collection of exemplar articles that are highly relevant to your review question, ideally with sound study design. 

Collecting a gold set 

The papers in your gold set may come from a variety of sources: 

• Key papers recommended by subject experts, supervisors, or peers 
• Articles located via reference checking of key papers provided by subject experts or supervisors 
• Results of preliminary 'scoping searches' from databases such as MEDLINE, Embase, or PubMed
• Citation searching to locate further relevant references cited in, or by, key papers. Databases including Scopus and Web of Science, and search engines such as Google Scholar, include information about articles which have cited a particular paper. Look for "Cited by" or "Times Cited" features. 

Using your gold set 

• As they are highly relevant studies, it is common for papers from your gold set to be cited within the protocol for your review.
• Use the papers in your gold set to identify relevant search terms that can be included in your own search strategy. 
• As you develop your search strategy, you can test its effectiveness by checking whether it retrieves the papers in your gold set, along with other relevant references on your topic. This is an indication that your search strategy is sound. 

Before you start a systematic review, you need a protocol document that outlines how the research will be conducted. It is an important planning document that will guide work throughout the project. In essence, a protocol:

• Sets rules for the review and provides clarity for team members
• Outlines team work and research methodology
• Pre-defines the objectives of the review
• Promotes transparency and mitigates against author bias

What to include in your protocol

The protocol document must include:

• Description of a clear and answerable research question that will be investigated
• The inclusion and exclusion criteria for selecting studies
• Eligibility criteria for potential studies 
• A search strategy from at least one database that will be translated to other databases and is reproducible
• Critical appraisal and/or quality assessment tools that will be used
• Information about how bias will be assessed
• Extraction tools and guidelines that will be followed

Online templates for system review protocols include the JBI Systematic Review Title Registration Form and PROSPERO Systematic Review Protocol & Support Template.

The following reporting standards from Cochrane, Joanna Briggs Insitute (JBI) and PRISMA are important to consider.

Cochrane Handbook, Chapter 1, section 1.5
Provides guidance on protocol development, expectations, data management and quality assurance.

JBI Manual for Evidence Synthesis, Chapter 1, section 1.3
Registering a review title and protocol.

PRISMA for systematic review protocols (PRISMA-P)
Statement paper, explanation and elaboration paper and key documents including the checklist.

Making your protocol publicly available

Protocols should be made publicly available in a registry or journal before you undertake your systematic review. In fact many journals require registered protocols to consider systematic review manuscripts. At a minimum, journals will ask where the protocol is registered. 

Registries:

• PROSPERO - International prospective register of systematic reviews
• Open Science Framework (OSF)
• Figshare

Some journals also publish systematic review protocols. Check the instructions to authors in the journal/s you plan to publish with. They may have directions about protocol registration. Or you can submit to one of the below journals:

• BMJ Open
• JBI Evidence Synthesis
• Systematic Reviews BioMed Central journal