The planning stage of a systematic review is essential in avoiding the 5 most common mistakes in conducting systematic reviews and increasing the likelihood of future publication. The steps outlined on this page are often conducted concurrently.
A well-built clinical question is the cornerstone of evidence-based health care and has been incorporated into the production of systematic reviews. Formulating a question is the first and most essential step in preparing a protocol and will inform the whole review process, including:
A question formulation tool or framework can be useful in breaking down a topic area into its key components. There are different tools for different types of review
You should choose the most appropriate tool for your type of question. If your topic isn’t easily structured using one of these tools, you can remove elements or adapt the headings to make it work for your question. The key aspect of formulating a question is that you can take a topic and separate it into its component parts.
Example Questions
Here are 2 examples of how we might break down 2 different questions on the same topic of delayed antibiotic prescription:
1. How do delayed antibiotic prescriptions for respiratory infections affect patient & service outcomes compared to immediate /no prescription?
Population = Patients with respiratory infections
Intervention = Delayed antibiotic prescription
Comparison = Immediate or no prescription
Outcomes = time to recovery, repeat GP appointment, emergency hospitalisation, patient satisfaction...
2. What are the barriers and facilitators to implementing delayed antibiotic prescription to patients with respiratory infections attending primary care?
Population = Patients with respiratory infections
Concept = Barriers / Facilitators to implementing delayed antibiotic prescription
Context = Primary Care
Once you have defined your question you can start the process by conducting scoping searches. These are often simple searches for the key elements of your question. Using the example question above, we might search for:
Delayed prescribing AND antibiotics AND respiratory tract infections
Step 1: Locating existing systematic reviews
Identifying existing reviews or review protocols has three main purposes:
To verify that your question hasn't already been answered – published review
To ascertain whether researchers are in the process of conducting a review on the same question – review protocol
To identify related systematic reviews that must be accessed so you can review the reference lists for identifying relevant primary studies.
Useful databases for identifying systematic reviews in health care
Step 2: Locating key studies
If you have existing knowledge of the topic area, you will already be aware of key studies that would provide background information for the review or be included in your final analysis. If not, you might want to conduct initial searches of a key database for your subject area (e.g. PubMed/Medline or Embase for healthcare) to inform the development of your protocol and search strategies. This will help with:
Refining the scope of your review – if you have too many or too few results on your initial searches, you might want to look again at your question
Informing the development of a search strategy – by reading relevant articles you’ll develop an understanding of the variability in terminology which will need to be incorporated into your final search
Writing the contextual / background information for your protocol
Once you've formulated your question and established the need for a new review on your topic, you will need to start developing a protocol to guide the conduct of your review. This will cover inclusion/exclusion criteria, screening methods, risk of bias assessment and data analysis.
PRISMA-P provides guidance for the information that needs to be reported in a protocol for a systematic review. Systematic review protocol templates are available from PROSPERO. Scoping review templates are available from the Joanna Briggs Institute and OSF websites.
It is good practice to prospectively register your protocol and in many cases a requirement for future publication of the review. You may want to explore further why prospective registration of systematic reviews make sense.
Traditionally, PROSPERO has been the main repository for prospectively registering systematic review protocols in health care. In recent times, the development of alternative evidence reviews (e.g. scoping reviews) and the increase in preprint archives and collaborative open research platforms has increased the options available to researchers for prospective registration. In this case, you may need to further explore where to prospectively register a systematic review
As the lead on a systematic review project, you will need to assemble a team of people around you that can provide methodological and topic support for your review. The size of the team around you will depend on the complexity of the question and potential volume of research that will need to be screened, quality assessed and synthesised. You will need people who can:
Develop, peer review and conduct searches
Double up on screening studies, extracting data and quality assessment
Provide methodological expertise in statistical, thematic or realist analysis
Supply topic / clinical expertise