How to Start a Systematic Review as a Medical Student
A realistic step-by-step plan for medical students starting a first systematic review: question, protocol, screening, risk of bias and reporting.
A systematic review answers a focused question by finding, selecting and appraising all the relevant studies using a pre-planned, reproducible method. It is excellent training and can be a strong publication, but it takes more time and teamwork than many students expect. Plan for several months, not several weeks.
Step 1: Build a team
You need at least two people to screen studies and extract data independently, a senior or guide with subject knowledge, and ideally access to a librarian or someone experienced in literature searching. If you plan a meta-analysis, involve someone comfortable with statistics early.
Step 2: Frame a focused question
Use PICO for intervention questions:
- Population: adults with type 2 diabetes
- Intervention: structured yoga programme
- Comparison: usual care
- Outcome: change in HbA1c
Narrow questions are more manageable and more useful than broad ones.
Step 3: Check that the review has not been done
Search PubMed, the Cochrane Library and PROSPERO (the international register of systematic review protocols). If a recent, good-quality review exists, your review may still be useful if it covers new studies or a different population, but you must explain why.
Step 4: Write and register a protocol
The protocol states your question, eligibility criteria, databases, search strategy, screening process, data to extract, risk-of-bias tools and planned synthesis. Registering it on PROSPERO (where eligible) before you start screening improves transparency and reduces the risk of changing methods after seeing results. The PRISMA-P guidance describes what a protocol should contain.
Step 5: Search comprehensively
- Search at least two major databases, commonly PubMed/MEDLINE plus others such as Embase, Cochrane CENTRAL or Scopus, depending on access.
- Combine free-text words and subject headings (MeSH in PubMed) with Boolean operators (AND, OR).
- Check reference lists of included studies and trial registers.
- Record the exact search strings and dates; you will report them.
Step 6: Screen in two stages
- Titles and abstracts, then
- Full texts against your eligibility criteria.
Two reviewers should screen independently and resolve disagreements by discussion or a third person. Record the numbers at each stage for the PRISMA flow diagram.
Step 7: Extract data and assess risk of bias
Use a pre-tested extraction form (study design, participants, intervention, outcomes, results). Assess risk of bias with a recognised tool, such as RoB 2 for randomised trials. Choose an appropriate tool for observational studies and name it in your protocol.
Step 8: Synthesise and report
Synthesise results narratively, in tables, and, if studies are similar enough, with a meta-analysis. Report the review using the PRISMA 2020 checklist and flow diagram.
A realistic timeline
| Stage | Typical time |
|---|---|
| Question, team, protocol, registration | 3–6 weeks |
| Searching and de-duplication | 2–3 weeks |
| Screening (two reviewers) | 4–8 weeks |
| Data extraction and risk of bias | 4–6 weeks |
| Analysis and writing | 4–8 weeks |
Times vary widely with the number of studies found.
Common mistakes
- A question so broad that thousands of records need screening.
- Searching only one database or only free full texts.
- Screening alone, without a second reviewer.
- Changing inclusion criteria after seeing the studies.
- Pooling studies that are too different to combine meaningfully.
Key takeaways
- Start with a focused PICO question and a team of at least two reviewers.
- Check PROSPERO and existing reviews, then write and register a protocol.
- Search widely, screen in duplicate and record every number.
- Report with PRISMA 2020.
References
- Higgins JPT, Thomas J, Chandler J, et al., eds. Cochrane Handbook for Systematic Reviews of Interventions. Cochrane. training.cochrane.org/handbook
- Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71. doi:10.1136/bmj.n71
- PROSPERO: International prospective register of systematic reviews. crd.york.ac.uk/prospero
- Sterne JAC, Savović J, Page MJ, et al. RoB 2: a revised tool for assessing risk of bias in randomised trials. BMJ. 2019;366:l4898.
This article is for education. Always follow the requirements of your university, guide, ethics committee and target journal.