Performing Effective Literature Searches
For MDR and Other Regulations
Any medical device must have adequate clinical evidence to demonstrate compliance with relevant, essential safety and performance requirements. A thorough and systematic Literature Search is one of the most important components of a full and complete Clinical Evaluation. An SLR helps find appropriate clinical data that isn’t available from the manufacturer and demonstrates its clinical safety and performance.
The Literature Search establishes the safety and performance of the device, and also includes a review of the current knowledge and/or state of the art (SOTA). These two sections work together to help the device evidence its clinical safety and performance.
When limited data from the manufacturer is available, clinical data from scientific literature of equivalent/similar products can be a substantial portion of conformity evidence.
Vital Components
We should include all the sources of data, the methodology intended for each search, and the specific search words and criteria used n the Literature Search protocol. It’s crucial to determine how to apply selection criteria to published literature and justify why you’re doing so, as well as strategies for dealing with data duplication from multiple sources.
To maintain data integrity, manufacturers must define data management methods, such as quality control and a second evaluation of extracted data by additional reviewers. It is necessary to provide a rationale for an appraisal plan that includes methods for evaluating each of the data sources used and their relevance to their device. The Literature Search review report should also cover the analysis plan and define the methods for analysing the data, such as data processing and transformation.
Manufacturers must consult applicable standards and guidance documents in the corresponding medical device field to determine the current SOTA. Using information on the medical condition managed with the device, benchmark devices, and medical alternatives available to the target population would be beneficial.
The literature search, however, is complex, and the current guidelines require the search strategy to be thorough, objective, comprehensive, and repeatable, as well as documented and justified. The use of trained literature search engines that improve search results and save the manufacturer time could be advantageous.
Plan a Literature Search and Review Plan
Manufacturers should define a literature search protocol that scopes out the plan, clarifies any problems, and conforms with existing guidelines before beginning the search.
You must set a literature search plan with several searching approaches out before beginning the search. One of the most common methodologies is the PICO (P- Population, I- Intervention, C- Control or Comparison, O- Outcome). Other examples are SPIDER, SPICE, PEO and ECLIPSE. PRISMA statement is used as a reporting structure for your report. Where the approasial criteria can come from MEDDEV2.7.1rev4 or IMDRF Clinical Evaluation guidance, alternatively you can use CASP, MOOSE, CONSORT, STROBE, STARD, GRADE, NOS, PROBLAST, QADRAS, LEGEND or SPIRIT where appropriate.
It’s vital for manufacturers to build a solid literature search technique that we can readily repeat and validate throughout future CER revision.
Identify the Most Suitable Literature Search Databases
The Literature Search database determines the quality of the data. Thus scientific databases are required, and not ‘grey’ databases such as internet searches, non-published data, and citations. CER’s commonly use scientific databases such as Cochrane, EMBASE, Pubmed, CHINAHL, SCOPUS, Cochrane Library or PROQUEST databases for literature search.
Separately, vigilance databases from competent authorities provide real world evidence (RWE) in safety and performance, such as device adverse events and recalls. In future, EUDAMED will be the database for EU MDR but it doesn’t have the critical volume of data other sites do yet. Therefore, searches on competent authority websites such as the FDA, BfArM, Sante, MHRA, HPRA etc are being used.
Start Systematic Literature Review
For the best results, you should use a medical librarian or experienced SLR reviewer to carefully develop your search syntax. The syntax is made up from chosen keywords that have been vetted for reach and strategies, such as Boolean operators. This is designed to ensure you find articles relevant to your device. You should also use filters to refine results to meet the expected criteria. We should check various searches for the most relevant and best reach before signing off on the search strategy.
Conduct your literature search and transfer all your results into your inclusion and appraisal workflow. Articles will need to be screened one by one to determine their relevance. Where there is doubt, a second round of screening needs to be done based on inclusion and exclusion criteria, which will consider the full text article. Once this is done, you now have your articles to review for their suitability and contribution.
Suitability appraisal explores whether it is the correct type of device, the same Intended purpose of the device, and for what population and indication it was used in the article. It’s vital to determine if the data you’ve obtained is accurate and relevant to the medical device being evaluated.
Contribution appraisal looks at the quality of the data and design of the study. The MEDDEV guidelines provide guidance on evaluating data sets’ consistency, methodological quality, and scientific validity.
The appraisal criteria does not cover the risk of bias, so you need to include it in your SLR process. Bias has a big impact on the reliability and credibility of the data.
Prepare the Systematic Literature Review (SLR) Report
Each review must produce an SLR report. This can either be a stand-alone document (recommended) or reported in the CER.
The report will use the results that have passed through each reviewer’s checks of the inclusion criteria and appraisals.
The SLR needs to contain a summary of each article, such as the type of study (RCT, Meta-analysis, cohort ext), type of device, indication and population. This summary is usually in a table format and can also cover points such as safety and performance, state of the art, benefits to the patient, along with side-effects and adverse reactions, etc. By putting the information in this format, it will also be easier to use in the CER to present evidence of GSPR conformity.
Having a robust and comprehensive SLR template in place which outlines all the information and highlights key areas such as the process, databases searched, criteria, findings and conclusion is essential.
CLIN-r+ Recommendations
To be successful, a literature search must be conducted by an experienced medical device or IVD systematic literature reviewer. There must be clearly defined protocols and well planned database search syntax the searches are conducted and the results placed into the SLR process. It is important to set out analytical and evaluation methods to be used, and ensure you have the specialised skills as stated in the MEDDEV 2.7.1rev4. All CER’s and their SLR reports must be conducted by staff that meet the MEDDEV2.7.1rev4 requirements to be accepted by auditors.
The SLR process must be conducted in a systematic way with evidence to show the process has been adhered to. Staying within the Clinical Evaluation Plan’s device search parameters is crucial.
It is important to note that we should keep the literature search separate from the State of the Art (SOTA) search. E.g., once we have identified known papers and sales literature comparative to the device, it is still necessary to further test the search parameters and ensure that references to the papers and sales literature appear in the SLR search results.
Overall, the literature search must be exceedingly thorough, consistent, and cover a wide range of search parameters. We must objectively document the findings in detail, and include good, bad and justifiable results. Additionally, we must ensure that it can be easily replicated and independently verified.
If you have further questions and would like to discuss how we can support you in the systematic literature review process, get in touch.
