A systematic review is a structured, transparent method for answering a focused research question by locating, appraising, and synthesizing all relevant studies. For students and early-career researchers, the process can feel daunting, but breaking it into three core stages—protocol, search, and study selection—makes it manageable. This guide explains each step in practical terms, with examples and tools you can use to conduct your own systematic review with confidence.
A systematic review is not just a literature review with a longer reading list. It is a rigorous research method that follows an explicit, pre-planned set of steps to minimize bias. Unlike a narrative review, which may select studies based on the author’s interests, a systematic review aims to capture all relevant evidence that meets clear eligibility criteria.
The protocol is your roadmap. Before you search for anything, you must write down exactly what you plan to do, why, and how. This does more than keep you organized; it helps others understand that your review was planned and not improvised along the way.
A helpful way to structure your question is the PICO framework. The table below shows how this works with a clinical example.
| PICO element | Example in a systematic review question |
|---|---|
| Population | College students who report mild to moderate anxiety |
| Intervention | A structured aerobic exercise program lasting at least 8 weeks |
| Comparator | No intervention or usual campus counseling services |
| Outcome | Change in self-reported anxiety scores from baseline to post-intervention |
Searching for evidence is not about typing a few keywords into Google Scholar. A high-quality systematic review depends on a balanced and sensitive search that finds all potentially relevant studies. You need to combine multiple databases, controlled vocabulary, and Boolean logic to build a reliable strategy.
For example, a simple PubMed search string for the PICO question above might look like this:
"aerobic exercise" OR "physical activity" OR "exercise intervention" AND "college students" OR "university students" AND anxiety OR "anxiety symptoms"
Test your search strategy by checking whether it retrieves a few known relevant papers. If not, adjust your terms before proceeding.
Study selection turns your search results into a final set of included studies. You will screen records in stages: first by title and abstract, then by full text. To keep this process fair and reproducible, your eligibility criteria must be specific and applied consistently.
Your eligibility criteria should be specific enough to apply consistently, but not so narrow that you miss relevant studies.
Handling hundreds or thousands of citations in spreadsheets is possible but painful. Fortunately, several free and paid online tools are designed for systematic review workflows. These tools let you import references, remove duplicates, and work through screening in a structured way.
Popular tools include Rayyan and Covidence, but you can also use a shared spreadsheet if your review is small. The key is to keep a clean record of every decision you make.
Once you have your final list of included studies, you need to collect their key characteristics and outcomes in a structured way. This data extraction stage is critical because your analysis will rely on complete and consistent data from every study.
Even with careful planning, small mistakes can weaken a systematic review. Knowing the common pitfalls in advance helps you stay on track and produce a more reliable final product.
A systematic review is only as good as its search. If you miss key terms, you miss key studies.
Conducting a systematic review is a demanding but highly rewarding process. By preparing a clear protocol, building a careful search strategy, and selecting studies in a reproducible way, you can produce evidence that is genuinely useful to researchers and decision-makers. Remember that the process is not linear; you may need to revisit your search terms, refine your criteria, or adjust your extraction form as you learn more.
A systematic review protocol is a written document that explains the rationale and planned methods for the review. It includes the research question, search strategy, eligibility criteria, data extraction plan, and quality assessment methods. Registering the protocol in a public repository also helps prevent duplication and unplanned changes.
It depends on the topic, the number of included studies, and your experience. A small review done by a student can take several months, while a large review with a team might take over a year. Most of the time is spent searching, screening, and extracting data, so good organization is essential.
At least three databases are recommended for a traditional systematic review. For many health and social science topics, this means PubMed or Cochrane Library, Web of Science or Scopus, and a subject-specific database like CINAHL or PsycINFO. Searching grey literature sources is also recommended to avoid publication bias.
If your search returns very few studies, you may need to broaden your question or reconsider your inclusion criteria. It can also be useful to check how other researchers have defined similar concepts and to search for unpublished studies in clinical trial registries and academic theses. Reporting a thorough search in an empty review is still valuable, but you should explain the limited evidence clearly.
Use AND to combine different concepts, OR to combine synonyms within a concept, and NOT only when necessary, because it can exclude useful records. Write down your question first, then break it into concepts. For each concept, list synonyms and variant spellings, and link them with OR. Finally, combine the concept groups with AND.
Inclusion criteria define the characteristics a study must have to be included in your review, such as participants, intervention, and outcome. Exclusion criteria are conditions that disqualify a study even if it seems relevant, such as wrong language, full text unavailable, or duplicate publication. Both should be based on your research question and stated clearly in your protocol.
It depends on your research question and available resources. Excluding non-English studies can introduce language bias, but translating studies can be resource-intensive. Before deciding, think about whether the topic is likely to have important evidence in other languages. Many reviews restrict to English and state this limitation explicitly.
Grey literature refers to reports, dissertations, conference papers, government documents, and other materials that are not published in traditional peer-reviewed journals. Searching grey literature can reduce publication bias because studies with null or negative findings are sometimes never published in journals. Examples include trial registries, institutional repositories, and conference abstracts.
Start with your title and abstract screen using a strict checklist derived from your eligibility criteria. Do not read the full text unless a record’s relevance is unclear. Work in batches, use a screening tool, and track every decision. Having a second reviewer helps, but even one reviewer can stay consistent by regularly reflecting on the criteria.
A PRISMA flow diagram is a visual summary of the study selection process in a systematic review. It shows how many records were identified, duplicates removed, records screened, full texts sought, and studies included or excluded. It is a standard part of most systematic review reports and makes the review process transparent.
