Best Practices for Conducting Literature Searches in IVDR Performance Evaluation

Just as in clinical evaluation, literature searches are critical in the performance evaluation process under the IVDR. Scientific (peer-reviewed) literature is established as an acceptable source to demonstrate the scientific validity and clinical performance of IVDs under Part A, Annex XIII of the IVDR. Other acceptable sources of data that may be identified in literature searches include consensus expert opinions/positions from relevant professional associations (scientific validity) and published experience gained by routine diagnostic testing (clinical performance). 

Importance of identifying “state-of-the-art” in performance evaluation under the IVDR

Another area in which literature search can play a critical role, but is frequently overlooked by manufacturers, is in identifying the “state-of-the-art”. This is highly relevant when identifying the evidence necessary to demonstrate conformity with GSPRs 1, 4, and 9, all within the scope of GSPRs to be considered in performance evaluation. ‘State-of-the-art’, while not defined in the IVDR, is defined in MDCG 2022-2 as the “Developed stage of current technical capability and/or accepted clinical practice in regard to products, processes and patient management, based on the relevant consolidated findings of science, technology, and experience.” While many manufacturers think of compliance with international standards and guidance documents (such as those recommended in BS ISO 16142-2:2017) as being sufficient to determine the state-of-the-art, these are not the only sources. Clinical or professional association guidelines (such as those published by the International Federation of Clinical Chemistry and Laboratory Medicine (IFCC)) may also be considered to represent the state-of-the-art. It is the manufacturer’s responsibility to demonstrate that state-of-the-art methods have been applied. 

It would therefore be prudent for manufacturers to consider including searches and queries in their literature searches that address scientific validity, clinical performance, AND state-of-the-art. 

Best practices for IVDR performance evaluation literature searches

Taking the above into consideration, the following represents what may be considered best practices for performance evaluation literature searches: 

1. Process standardization

The best literature searches follow a pre-established, well-defined, and consistent methodology. Manufacturers should consider establishing a literature search as a process within the framework of their quality management system (QMS). This may reside as a sub-process under the performance evaluation process, although literature search also plays an important role in post-market surveillance (PMS), and post-market performance follow-up (PMPF), and can be a powerful tool in the risk management process for the identification of reasonable foreseeable hazards, it could also be established as either a standalone work instruction and/or instruction-driven form (protocol and report) templates cross-referenced by the relevant procedures. 

2. Identification and understanding of relevant literature data sources

An important consideration is identifying those sources of published data to be searched as several literature databases exist, each with their respective advantages and disadvantages. When selecting data sources, time should be spent on testing their functionality, particularly regarding the limits available for selection in order to optimize searches and avoid scenarios where significant volumes of “excluded” literature are returned for search queries. Commonly used literature databases include Embase, PubMed, the Cochrane Library, and UK National Institute for Health and Care Excellence. However, relevant professional association databases should also be considered, particularly as sources of data on the state-of-the-art. 

Notified Bodies also encourage manufacturers to gather all relevant clinical evidence in order to understand IVD risks and benefits. Therefore, it is strongly recommended that the following data sources be considered in literature searches: 

• • Scientific (peer-reviewed) literature databases
  •  Professional association guidance databases (or similar)
  •  Clinical trial registries (e.g. ClinicalTrials.gov, WHO ICTRP, EU Clinical Trials Register)
  •  Regulatory agency vigilance databases (e.g. TGA DAEN / SARA, US FDA MAUDE / TPLC)
  •  “Other” sources (e.g. literature included in performance evaluation for similar devices)

3. Well-defined inclusion and exclusion criteria 

Clear and well-defined inclusion/exclusion criteria should be established for each of the three searches focuses (i.e. scientific validity, clinical performance, and state-of-the-art). When defining state-of-the-art exclusion criteria, time limit requirements should be established. The standard time limit used for state-of-the-art searches is 10 years, however, for devices subject to significant levels of innovation this could be reduced to 5 years. 

4. Literature search terms and query structures

A framework should be established to assist in the identification of relevant search terms which form the basis of the queries to be executed in each data source. In the case of literature, one of the most common frameworks utilized is PICO which stands for: 

• •  Patient/Population/Problem
  •  Intervention (cause, factor, treatment, etc.)
  •  Comparison
  • Outome

Constructing effective search queries for literature searches in medical device evaluation

Separate PICO statements should be compiled for each of the three searches focuses. Where possible, Medical Subject Headings (MeSH), which is the National Library of Medicine (NLM) controlled vocabulary thesaurus used for PubMed article indexing, should be used. For vigilance database searches, applicable European Medical Device Nomenclature (EMDN) descriptions or US FDA Product Codes are preferable so that data on similar devices is captured. When constructing search queries for the purposes of clinical performance they should always include the subject and similar device trade names, where such exist. 

5. Ensure IVDR performance evaluation literature search reproducibility 

The level of detail in the literature search protocol and report under the IVDR should be sufficiently detailed to ensure that when any search query is run, identical results are obtained in order to demonstrate the reproducibility of the search. In the case of literature, this includes listing every single article identified in the search with clear identification of inclusion and exclusion and, in the case of the latter, justification for exclusion based upon the established inclusion/exclusion criteria. 

6. Establishment of data appraisal criteria 

The final consideration regards the establishment and application of data appraisal criteria and applies to all types of included data. These should be clearly established up-front (e.g. as part of the literature protocol and/or performance evaluation plan/report) and include criteria for all three search focuses. In the case of literature, the Oxford Centre for Evidence-Based Medicine (OCEBM) Levels of Evidence are one of the most commonly applied criteria, however, manufacturers are free to determine the criteria to be applied. 

Learn more about IVDR performance evaluations

• • Planning Performance Evaluation Under the IVDR
  • Understanding the Specific Evaluation Documentation Required for IVD Performance Under Article 56 of the IVDR
  • Navigating IVDR Compliance: Applicability of Scientific Validity for IVDs

Contact MedEnvoy

If you have any additional questions regarding IVDR performance evaluation or require relevant training/consulting services, get in touch. 

• • Frequently asked questions