Select your language

The IVDR requires manufacturers to conduct and document a systematic literature search as an integral part of the performance evaluation. However, literature searches for IVDs often present more challenges than searches for other devices. Here, we highlight the key points to consider when identifying evidence on IVD devices.

Optimising the Search Scope: PICO and PIT Frameworks

The scope of a systematic review of medical literature is broken down into four main concepts: population, intervention, comparator, and outcomes (PICO). Additional concepts, such as study type, timing, and setting/user, are sometimes added. While it is possible to use the PICO approach for an IVD review, other frameworks may be more appropriate. Clinical evidence related to devices for diagnosis or screening is best analysed using the PIT framework (population, index test, target condition). For a companion diagnostic review, it may be beneficial to combine PICO and PIT to define the test and the intervention. The table below serves as a guide, providing practical examples of scope definitions for IVDs using the PIT and/or PICO approach. The device's intended purpose primarily determines the most suitable approach for a specific IVD.

Concepts Diagnostic example
(TcB for jaundice)
Prognostic example
(protease activity predicting venous leg ulcers healing)
Companion diagnostic example
(MGMT methylation temozolomide treatment for glioblastoma)
Population In term or preterm infants aged 0 to 28 days Venous leg ulcers patients Glioblastoma patients
Intervention NA[1] NA Temozolomide
Index test Noninvasive transcutaneous bilirubin  Protease activity MGMT promoter methylation
Comparator Total serum bilirubin NA Any
Outcomes NA Time to healing or proportion of healed ulcers Overall survival
Target condition Hyperbilirubinaemia NA MGMT methylation status
Setting/user (optional) Neonatal intensive care units, paediatric emergency units or wards Any Any
Timing (optional) Any Outcome: any follow‐upPrognostic factor: at baseline MGMT tested before treatment
Study type (optional) Cross‐sectional and prospective cohort studies  Longitudinal studies Longitudinal studies, nested case-controls, RCTs

 

Developing Effective Search Strategies: Challenges and Best Practices

Once the scope of the literature review has been defined, an appropriate search strategy needs to be developed. The standard approach for a systematic search strategy involves identifying search terms for each review scope concept, connecting these terms using the Boolean “OR” operator within each concept, and finally combining the different concepts with the Boolean operator “AND”. A common challenge with literature search strategies for IVDs is the inadequate description of relevant studies in abstracts, leading to indexing issues in databases.

The following approaches can be helpful in overcoming this issue:

  • Identify a few key articles and explore how these articles are reported and indexed.
  • Only include one or two concepts in a search protocol and focus on identifying all potentially relevant terms for the respective concept. With this approach, several search protocols must be developed to cover all concepts. A multi-stranded approach can be used to combine these different search protocols.
  • Do not rely on controlled vocabulary terms, like MeSH in PubMed.
  • Do not apply methodological search filters, e.g. for study type, to the final search set. The use of such filters in one of the strands of a multi-stranded search may be appropriate. However, always be aware of which study design may be appropriate to the IVD of interest – for example, a randomised controlled trial does not necessarily provide higher-level evidence for an IVD if the randomisation is related to the therapeutic intervention and not the diagnostic approach.

Examples of different approaches to the development of a search strategy are provided in the table below.

Search strategy Commonly used concepts Example (for TcB test for jaundice in newborns)
Single concept Index test[2] TcB OR bilicare OR Bilicheck OR BiliMed OR bilimeter OR JM‐103 OR JM‐105 OR bilirubinometer OR jaundice meter OR icterometer
Two concepts Index test AND Target condition Above AND (jaundice OR hyperbilirubinemia OR icterus)
Three concepts Index test AND Target condition AND Population[3] Above AND (babe OR baby OR infant OR newborn OR premature OR preterm OR NICU)
Multi-stranded:[4] (Index test AND Target condition) OR (Index test AND population) OR (Population AND Target condition) Combinations of the above

 

Comprehensive Search Strategies: Addressing Clinical Performance, Scientific Validity, and State of the Art

The performance evaluation of an IVD consists of three separate parts: the analytical performance report, the clinical performance report, and the scientific validity (SVA) report. In addition, the IVDR states that the performance evaluation plan should include a description of the state of the art (SOTA), which can be separated into two parts: the medical and the technical SOTA. Due to nature of IVDs, technical aspects of the SOTA must be highlighted. Both the SOTA and the SVA usually include multiple research questions, e.g., target condition (including its clinical relevance, diagnosis, and management), alternative methods, analyte, and IVD technology (e.g., technical/methodology guidelines). Depending on how well-established the analyte is, different parts of performance evaluation may overlap or diverge significantly in their scope. For example, well-established analytes will have a SOTA that supports the SVA. Whereas, this will rarely be the case for a novel analyte, where the clinical performance report may support the SVA.

Therefore, a single literature search can rarely cover all the parts of the evaluation, which may require multiple search strategies to address this.

 

Expanding IVD Literature Reviews: Beyond Bibliographic Databases to Identify Comprehensive Evidence

Like any systematic literature review, IVD reviews primarily rely on bibliographic databases. PubMed and/or Embase are most suitable for a comprehensive identification of relevant literature using a methodologically solid strategy. However, other sources also play a prominent role in the identification of evidence applicable to a specific IVD. Instructions for use of competitor devices and information on authorities’ websites can provide data on the performance of similar devices. Since the specific device used in a study is not commonly mentioned in the abstracts of IVD literature, strategies like citation searching, grey literature, and web search can prove crucial to identifying highly relevant studies missed in database searches.

 

Summary

Systematic literature searches for the performance evaluation of IVDs follow the same general principles as for other medical devices; however, a few key differences need to be considered. In contrast to interventional devices, the inclusion of comparators and clinical outcomes in search protocols is not always appropriate. Information that is relevant for the performance evaluation of an IVD is often not reported in abstracts, and incorrect indexing is observed more frequently, which needs to be considered when defining the search strategy. Additional sources should be consulted to ensure a comprehensive review of relevant clinical evidence.

Kathrin Abegg, DVM PhD
Team Leader Medical Writing ISS AG

Anna Neodo, PhD
Medical Writer ISS AG


[1] – NA indicates that usually this component will not need defining for this type of IVD.
[2] – Normally only suitable for well-defined and consistently described tests.
[3] – Only suitable if this can focus the search without eliminating key studies.
[4] – Several different concept combinations.

Got a question or need advice?

Do not hesitate to contact us if you have any questions or need advice.