The systematic literature review is the methodological foundation of most CERs. It makes it possible to identify and evaluate all available clinical data on the device or on comparable devices. The word “systematic” is not decorative: it is what distinguishes a regulatory-acceptable literature review from a subjective selection of favourable articles.
Why “systematic” is the central criterion
A non-systematic review — arbitrary selection of articles, non-reproducible search, undocumented omission of unfavourable studies — exposes you to two serious regulatory risks.
The first is selection bias: if only favourable studies are retained, the conclusion of the CER is distorted. Notified bodies verify that the search strategy was broad enough to identify unfavourable studies, and that exclusions are justified by criteria defined in advance and not by their results.
The second is even more serious: the deliberate omission of unfavourable safety data may be characterised as regulatory fraud. A CER that omits publications documenting adverse effects on comparable devices is a file that does not meet the requirements of Article 61.
The databases to search
PubMed / MEDLINE. Essential for international medical literature. Free, exhaustive in the biomedical field. The search is carried out via the PubMed interface or via tools such as Rayyan for managing the results.
EMBASE. A complement to PubMed with better coverage of European journals and medical conference proceedings. Paid, but often accessible through academic institutions.
Cochrane Library. Systematic reviews and meta-analyses. A reference for high-level evidence on medical interventions.
Regulatory databases. EUDAMED (vigilance and post-market surveillance data), national databases of the competent authorities, notified bodies’ assessment reports available publicly.
Specialised databases according to the type of device. IEEE Xplore for active devices and medical software, CINAHL for devices used in nursing care, oncology databases for devices used in cancer care.
Traceability of the search strategy
The search strategy must be documented in a complete and reproducible manner: search terms used with Boolean operators, databases searched, search dates, inclusion and exclusion criteria for studies, number of articles identified at each filtering stage (identification, screening, eligibility, inclusion). This process is often represented in the form of a PRISMA flow diagram.
A CER without a documented search strategy is systematically returned by notified bodies with a request for supplementary information. The absence of this documentation is considered an absence of systematic review, whatever the articles ultimately cited.
Handling unfavourable articles
Studies that document adverse effects, performance limitations, or incidents on comparable devices must not be excluded from the review. They must be analysed, and their conclusions must be integrated into the conclusion of the CER. If these data alter the benefit/risk ratio, this must be documented and justified — and the PMCF plan must take account of it.