How long does a systematic review for a CER take?
6 Min Read
Methodology
The steps that take the most time in a literature review for a CER, the factors that extend timelines and how to shorten them without losing rigour.
By Bopha Suybeng, PharmD, founder of NOÉ. Twelve years of experience in the clinical evaluation of medical devices.
In short: there is no standard duration. A literature review for a CER can take from a few weeks to several months, depending on how broad the question is, the number of articles to screen, the number of reviewers and the level of traceability expected. The longest steps are almost always article screening and data extraction.
What the literature says about the duration of systematic reviews
For academic systematic reviews, an analysis of reviews registered in the PROSPERO registry (Borah et al., BMJ Open, 2017) found an average of more than a year between registration and publication. A review for a CER does not face the same publication constraints, but it shares the same demanding steps and an equally strong need for reproducibility.
The steps that take the most time
Title and abstract screening: it covers all search results, often several hundred or thousand records.
Full-text screening: articles have to be obtained, sometimes purchased, then read in full.
Quality appraisal: each included study is assessed and weighted.
Data extraction: populations, follow-up, performance and safety outcomes, adverse events.
Writing and alignment with the clinical evaluation plan and the state of the art.
Factors that extend timelines
A research question that is too broad or vague eligibility criteria.
Search strings that need reworking after a first pass.
Dual screening by two reviewers, with disagreements to resolve.
Back-and-forth with the notified body on the justification of exclusions.
A review redone from scratch at every CER update.
How to save time without losing rigour
Frame the question in the PICO(S) format before any search.
Write a complete protocol and have it approved upfront.
Record each decision when it is made, rather than after the fact.
Structure extraction with a grid defined in advance.
Plan for updates: keep search strings, dates and decisions so the review can be rerun.
Where a tool like NOÉ comes in
NOÉ mainly speeds up the repetitive steps: applying eligibility criteria across the whole corpus, keeping justifications, the PRISMA flow diagram, flagging the articles whose full text needs to be obtained, and data extraction. Clinical judgment, analysis and conclusions remain the expert's. For a worked example on a real case, see our article comparing a manual review with a review carried out in NOÉ.
