Performing Effective Literature Search – For EU MDR and Other Regions

A literature search for EU MDR is the systematic, documented search of published and unpublished sources for clinical data on your device, equivalent devices, and the state of the art. It is run to a predefined protocol so that it is thorough, objective, and reproducible. Other regulators ask for similar evidence in different ways. The FDA accepts published literature as part of the valid scientific evidence behind a 510(k), De Novo, or PMA; Australia’s TGA publishes its own clinical evidence guidelines; and China’s NMPA expects Chinese-language sources. Designing one search that satisfies every market you sell in is usually cheaper than running a separate search for each.

For any medical device, there must be sufficient clinical evidence to demonstrate conformity with the relevant general safety and performance requirements in Annex I of the MDR.

One of the most crucial components for a thorough and complete Clinical Evaluation is a comprehensive and systematic literature search which is performed to identify suitable clinical data that is not available from the manufacturer.

Why are literature searches important?

The search serves two purposes: establishing the device’s safety and performance, and establishing the state of the art it is measured against.

The literature search has two-fold importance. Not only does it establish the safety and performance of the device, but it also includes a review of the current knowledge and/or state-of-the-art (SOA).

Together, these two sections integrate into the complementary aspects of a Clinical Evaluation Report (CER), such as clinical data, worldwide post-market surveillance, and risk assessments, to conclude the clinical safety and performance of the device.

Sometimes, when the data from the manufacturer is insufficient, clinical data from the scientific literature of equivalent/similar devices will represent a major part of clinical evidence. Furthermore, for smaller manufacturers or manufacturers of Class I medical devices, the literature will be almost, if not all, of their clinical data.

Only after a series of stringent checks to assess its appropriateness, the appraised clinical data set identified is included in the clinical evaluation. Sales literature, with comparative information and pictures, can also assist in initial, comparative product evaluations.

Note that literature search and literature review are used interchangeably in this article. While not the same thing, for a clinical evaluation report they can be considered equivalent, as they are both an integral part of the process.

Literature Search Requirements Beyond the EU

The FDA, MHRA, TGA, and NMPA all use published literature as clinical evidence, but they differ in how central it is, which sources they expect, and how it must be documented.

United States (FDA)

The FDA does not require an EU-style clinical evaluation report. Instead, it decides safety and effectiveness on “valid scientific evidence” as defined in 21 CFR 860.7, which ranges from well-controlled investigations to well-documented case histories and reports of significant human experience. Published literature can contribute to that evidence for a 510(k), De Novo request, or PMA, particularly where it describes the performance of the predicate or similar devices. For safety data, FDA databases such as MAUDE (adverse events) and the medical device recall database are standard sources, and they are useful for EU PMS searches too.

United Kingdom (MHRA)

Great Britain currently accepts CE-marked devices under transitional arrangements, with MDR- and IVDR-compliant devices accepted until June 30, 2030. Between February and April 2026 the MHRA consulted on recognising MDR and IVDR CE marking indefinitely, so check the current position before planning a separate UK search. In most cases, a search that meets EU MDR expectations will also satisfy UK requirements.

Australia (TGA)

The TGA publishes clinical evidence guidelines for medical devices that set out what it expects from a literature review, including a documented search strategy and appraisal method. Manufacturers often add the TGA’s own adverse event database to their safety sources when the device is on the Australian market.

China (NMPA) and other regional sources

China’s NMPA clinical evaluation guidelines allow evaluation based on literature for comparable devices, and reviewers commonly expect Chinese-language databases such as CNKI and Wanfang to be searched alongside international ones. Other regional databases worth considering, depending on where you sell, include LILACS (Latin America and the Caribbean), KoreaMed (Korea), and J-STAGE (Japan).

Designing one search for several markets

Build the protocol around the strictest target market, which for most manufacturers is the EU. Add the regional databases and vigilance sources each other market expects, and document which market each source serves. Keep one set of search strings and one PRISMA flow diagram so the same evidence base supports every submission. CiteMed Literature supports multi-database searches and exports for MDR, IVDR, and FDA documentation.

Essential elements of the literature review

A literature review rests on a documented protocol, controlled data management with second review, and a clearly defined state of the art.

Literature search protocol

The literature search protocol should include the sources of data, the methodology planned for each search, and the exact search terms and parameters used.

It is important to determine the selection criteria to be applied to published literature and document the justification for these choices, along with strategies for addressing potential duplication of data from multiple sources.

For a full sample protocol structure, see Literature Search Protocol for EU MDR.

Data management practices

Manufacturers must also define data management practices such as quality control and the second review of extracted data by additional reviewers to ensure data integrity.

The rationale for an appraisal plan consisting of methods for appraisal of each of the data sources used and its relevance to their device must be provided. The literature search review report should also cover the analysis plan defining the methods for analyzing the data, such as data processing and transformation.

State-of-the-art

To determine the current SOA, manufacturers must refer to applicable standards and guidance documents in the corresponding medical device field. The use of information relating to the medical condition managed with the device, benchmark devices, and medical alternatives available to the target population would help in this regard.

Conducting a literature search, though, is fairly complex, with the current guidelines requiring that the search strategy be thorough, objective, reproducible, documented, and justified. Here, the use of trained literature search engines that optimize search results and result in significant time reduction for the manufacturer could prove extremely beneficial.

Formulating the right research question

A focused research question, usually built with PICO, translates directly into a search strategy and keeps results from being too broad or too narrow.

Perhaps the most critical part of planning your literature search and review is to formulate appropriate research questions.

If your research question or hypothesis is not focused and clear you will end up with results that are too broad or too narrow in scope, too much or too little literature, and literature that is not relevant to the research question.

You also want a research question that allows for the appropriate level of research and analysis on your chosen topic.

PICO method

It can be helpful to formulate the research question using the PICO methodology (P- Population, I- Intervention, C- Control or Comparison, O- Outcome). The PICO method is also one of the most common methodologies applied to the literature search. Choosing your research question with the PICO method in mind lets you directly transfer your question to your literature search strategy.

Other question frameworks include SPICE, PEO, and ECLIPSE. PRISMA and MOOSE are reporting guidelines rather than question frameworks: they describe how to report a systematic review (MOOSE for meta-analyses of observational studies), and PRISMA 2020 is the version to use for your flow diagram.

Literature search and review plan

The plan defines the scope (device data plus state of the art), inclusion and exclusion criteria, keywords, and a documented, repeatable systematic method.

Before the execution of the search itself, you should develop a literature search protocol that scopes out the plan, helps clarify possible issues and aligns with current guidelines.

The literature search protocol should include several elements.

Scope of the literature search and review

The scope of the literature search and review should reflect what kind of clinical data the search should cover.

The literature search should always include two categories of clinical data:

• Clinical data on the device in question, as well as similar or equivalent devices
• Data on the current state-of-the-art/knowledge related directly to the device and/or its equivalents, any available medical alternatives, as well as benchmark devices and technologies

The included literature should always reflect the intended use of the device and its clinical benefits.

Inclusion/exclusion criteria

To only include relevant and applicable literature in your literature review, you need to define which literature should be included and which should be excluded.

Common inclusion and exclusion criteria are:

• Language
• Date of publication
• Geographic location
• Study design (systematic reviews, meta-analyses, controlled trials, etc.)
• Type of publication
• Participants
• Peer-review

Keywords

You also want to make sure your literature search and review plan includes the right keywords and phrases relevant to the topic. To identify keywords, it is important to be conscious of the writing style and keywords of the different scientific publications.

You could also brainstorm a keyword map with words and phrases relevant to your device. Make sure to include alternate spellings, synonyms, and any other words related to your research.

Systematic review

The literature search protocol should demonstrate that your literature search is a systematic review of relevant literature, both favorable and unfavorable.

The literature search itself should be well-documented and detailed, which the search protocol should also reflect. It is important to thoroughly document your search process, search methods, key words, and relevant databases to ensure the search is repeatable and properly conducted.

It is extremely critical that manufacturers develop a robust literature search strategy that can be easily replicated and validated during subsequent updates to their CER.

Building an effective search strategy is the foundation of any compliant EU MDR literature review, which our step-by-step EU MDR literature review guide covers in full.

Identify the suitable databases

Combine scientific databases, internet sources, unpublished data, and less obvious sources such as trial registries and conference papers.

The source of the literature search database, such as scientific databases, internet searches, non-published data, and citations, determines the quality of data presented.

Scientific databases

Scientific online databases such as Pubmed, Cochrane central trial register, EMBASE, PDQ evidence, Trip, and CINAHL are commonly used for literature searches. Google Scholar, the manufacturer’s website, FDA, Cochrane database of systematic review, etc. are typically used for internet searches.

Often, non-published data is collated from instructions for use, implant registries information, etc.

Non-obvious sources

Don’t overlook sources that might not be obvious, but can still provide useful information. These sources can be literature-related, such as conference papers, ongoing research at universities, grey literature, and databases of pre-print publications. They can also be alternatives to actual literature, such as graduate and postdoctoral theses or clinical trial registries. Online forums and social media can flag possible safety signals for PMS, but they should not be used as clinical evidence.

Perform systematic literature screenings

Refine results with Boolean logic, MeSH terms, and filters, then screen at two levels and check each source’s validity and reliability.

Once you have your carefully chosen research question and keywords, you can begin to search and screen your data.

Refining search results

This can be done using approaches like Boolean search strategy and/or the use of Medical Subject Headings (MeSH) terms, you can get good results. The application of filters also helps to refine search results to meet expected criteria.

The search results are screened on at least two levels to assess relevance – title or abstract screening, based on inclusion and exclusion criteria, and a full-text screening of the included articles.

It is critical to determine if the data found is valid and relevant for the medical device under evaluation. MEDDEV 2.7/1 Rev 4 gives detailed guidance on appraising the methodological quality and scientific validity of data sets, including the distinction between pivotal and supportive data.

Source reliability

Another part of screening literature is to assess the quality of the source, as well as the relevance, applicability, and quality of the literature itself. When and where was it published, does the author have good credentials, is the article published in a peer-reviewed journal or has in another way been peer-reviewed, is it bias-free, etc. These are all excellent questions to ask before including the literature in your clinical evaluation, as you do not want to base your clinical evaluation on less than credible sources and research.

Review your initial research question

Once the search is done, check whether the research question still holds or needs refining.

Reviewing the initial research question after the literature search has been completed is often overlooked.

However, it makes good sense to go back to your research question to check if it needs to be revised or if it is still valid and relevant.

Most papers will include a section at the end about suggested future research or further research to be done on the topic, which is a good place to start to check your research question and whether it can be focused to address the topic in a more appropriate way.

Prepare the literature search review report

The report presents the screened results against each review question, covering safety, performance, benefits, side effects, and the state of the art.

This report presents results of the search which address aspects like safety, performance, benefits to the patient, side-effects, adverse reactions, state of the art, etc. Using search results that pass screening, literature review reports are formed for each of the review questions.

Our thoughts

Plan the protocol early, keep equivalence and state-of-the-art searches separate, test the protocol against known papers, and document everything.

A successful literature search, like every research endeavor, must have clearly defined protocols and effective and timely planning well before the searches are initiated.

Not only are the analytical methods and time frames important, but also the evaluation methodologies used to collect the data, and the specialized skill to determine its scientific merit.

Device equivalence

The literature search related to device equivalence should be treated and managed as a separate activity from the search of the current state of the art.

Protocol testing

It is necessary to test the protocol for the search before finalizing. For example, after identifying known papers and comparative sales literature relevant to the device being evaluated or the current state of the art, it is still necessary to test these search parameters to ensure that references to these known papers and sales literature appear in your search results.

To conclude, we must ensure that our literature search is extremely thorough and consistent, encompassing a breadth of search criteria. We must document every piece in detail, allowing independent verification and replication of results selecting literature objectively ensuring that the good, bad and justifiable are all included. Since every published study and even sales literature can be cited, these may lead to additional relevant data and perspectives.

At this final juncture, proper planning and staying within your device search parameters becomes crucial. It is too easy to quickly wander down a virtual rabbit hole and find it nearly impossible to retrace your path without wasting significant amounts of time.

Thanks for reading! If you have any more questions on Literature Search or CER, get in touch with the CiteMed team, or see how CiteMed Literature handles searches across EU MDR and other regions.

FAQ

What is a literature search for EU MDR?

It is the systematic, protocol-driven search for clinical data on the device, equivalent devices, and the state of the art that feeds the clinical evaluation under Article 61. It must be thorough, objective, reproducible, and documented well enough for a notified body to follow every step.

Does the FDA require a literature search like EU MDR does?

Not in the same form. The FDA has no CER requirement, but published literature can form part of the valid scientific evidence for a 510(k), De Novo, or PMA under 21 CFR 860.7. FDA’s MAUDE and recall databases are also standard sources for safety data in both US and EU submissions.

What are the global literature search requirements if I sell in several regions?

Each regulator has its own guidance, but the core expectations are similar: a documented protocol, multiple databases, clear eligibility criteria, and transparent appraisal. The practical difference is sources, such as Chinese-language databases for the NMPA or national adverse event databases for the TGA.

Can one literature search cover the EU and other regions?

Usually yes. Build the protocol to EU MDR standards, add the regional databases and vigilance sources each other market expects, and document which source serves which market, so a single evidence base supports every submission.

Which databases should a literature search for other regions include?

Start with PubMed, Embase, and the Cochrane Library, then add region-specific sources as needed: CNKI or Wanfang for China, LILACS for Latin America, KoreaMed for Korea, J-STAGE for Japan, plus each market’s adverse event database.

Is a literature search the same as a literature review?

The search finds the records; the review screens, appraises, and interprets them. For the full review process, see our EU MDR literature review guide.