Before placing an in vitro diagnostic (IVD) on the EU market, the manufacturer must demonstrate compliance with all applicable requirements of the Regulation (EU) 2017/746 on in vitro diagnostic medical devices (IVDR). This includes the requirement to demonstrate that their diagnostic tests are scientifically valid — in other words to prove that the measured analyte is associated with the intended clinical condition or physiological state. While the (patho)physiological role of some analytes is well known and routinely used in the diagnostic workup of patients, others can be completely novel, and their diagnostic value is yet to be established. Understanding what scientific validity entails and how to demonstrate it is therefore essential for successful market access. An insufficiently substantiated scientific validity argument can result in significant delays or even rejection during conformity assessment, regardless of the device’s overall design quality.

Performance Evaluation: How Scientific Validity Fits In

The IVD performance evaluation under the IVDR consists of three components: the analytical performance report, the clinical performance report, and the scientific validity (SVA) report. Each of these plays a crucial role in fulfilling the performance evaluation requirements set out in the IVDR. Conformity with the applicable General Safety and Performance Requirements must be supported by evidence from all three areas to show the product is safe and achieves the intended clinical benefit.

In other words, the scientific validity report demonstrates that a particular analyte is sufficiently suitable to support the identification of a clinical condition or physiological state. This requirement also applies to devices that were already placed on the market under the IVDD.

Read on to take a deeper look at the possible sources and types of evidence to demonstrate the scientific validity of an analyte. Depending on how well-established the analyte is, different parts of the 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. Read more about systematic literature searches,  SOTA, and the link to performance and clinical data collection in another blog post here.

Sources: Where to find Evidence

The IVDR and the MDCG 2022-2 guidance document on general principles of clinical evidence for IVDs suggest possible sources of evidence for the SVA:

Scientific (peer-reviewed) literature: This can include any scientific articles e.g., found in scientific publication databases such as PubMed.

Consensus expert opinions and positions from relevant professional organizations: Guidelines, recommendations, and position statements can be found on the websites of well-established international and national health authorities (e.g., World Health Organization, National Institute for Health and Care Excellence), as well as professional organizations (e.g., European Society for Molecular Oncology, European Glaucoma Society). Some guidelines are published in scientific journals and can be identified during scientific literature databases search.

Data from devices measuring the same analyte or marker: This information can come from scientific literature on such devices, as well as other sources such as promotional or instructional material (e.g. instructions for use), clinical trial registries, summary of safety and performance (SSP), authorities registries and databases (e.g., FDA 510(k) database), textbooks, laboratory websites, and external quality assessment schemes (EQAS). The devices do not have to necessarily measure the analyte by the same method.  

Other relevant published data (on the device under evaluation): In addition to scientific literature, other published evidence on the device under evaluation can be harnessed, e.g., clinical trial registries, textbooks, EQAS, information from laboratory websites etc.

Proof of concept and clinical performance studies: Evidence from the IVD manufacturer’s own studies can also support the SVA.

Evidence from the first four sources can be identified during systematic literature review covering not only scientific publication databases, but also the websites of relevant healthcare organizations, similar device manufacturers, authorities, clinical trial registries etc.

Types of SVA Evidence

The SVA can rely on various types of evidence. The amount and quality of the available evidence will depend on how established the analyte-condition association is:

Clinical practice guidelines:^[1] well-established healthcare organizations recommend using the analyte in the context of the intended condition/state (e.g., for diagnosis, screening, surveillance etc.).

Health technology assessments (HTA): analyte test(s) for the intended condition is systematically evaluated in term of effectiveness, cost-effectiveness, and broader impact by well-established HTA organizations.

Textbooks: certain analyte-condition associations are so well-established that they form part of the basic education for medical students worldwide. Established IVD examination procedures may also be published in reference textbooks for laboratory medicine.

Systematic reviews / meta-analyses: quantitative and qualitative analyses of clinical evidence on analyte testing in the context of the intended condition / state.

Impact on patient outcomes and management: Studies investigating the impact of medical interventions based on results of a diagnostic test, either on patient outcomes or on the health system may become available for some analytes.^[2] Examples are impact of fecal occult blood testing in cancer screening programs (vs. standard care) on patient mortality, impact of calprotectin testing on the reduction of unnecessary endoscopies. Although the association of the analyte with the condition is demonstrated indirectly, such studies can help establish both the validity and importance of the analyte testing for in the healthcare system and represent a compelling source of clinical evidence.

Clinical performance of the analyte tests: data that shows that analyte tests adequately perform their function (e.g., diagnosis, prognosis, screening). IVDR explicitly mentions classic diagnostic accuracy parameters: sensitivity, specificity, positive and negative predictive values, and likelihood ratios. Depending on the type of IVD, other performance parameters may be also or more relevant.

Other measures of association of analyte and intended condition: if this association can be quantified by other types of analysis (e.g., simple correlations).

Levels of analyte in population with vs without intended condition or state. 

Nonclinical studies: e.g., animal, modeling, in vitro, in silico that supports the role of the analyte in the intended condition / state.

Building an SVA Strategy

As a key component of the performance evaluation plan, the SVA strategy must be tailored to the clinical context of the analyte. The guidance document MDCG 2022-2 on clinical evidence for IVDs distinguishes between analytes with specific indications, whose measurement serves specific diagnostic purposes, and analytes that may be relevant for multiple clinical conditions and are consequently intended to assess physiological status rather than a specific diagnostic indication. This difference in clinical context should be reflected in the SVA strategy and may impact the selection of sources and types of clinical evidence.

Apart from the clinical context, the SVA strategy mainly depends on how established the link is between the analyte and condition. For IVDs with well-known analyte(s), it is often challenging to deal with the deluge of available evidence. It is then best to focus on high-level and recent evidence (e.g., last 5-10 years) such clinical guidelines and systematic reviews and meta-analyses. Additional searches can be performed to check for any new studies that would not yet be included in the identified guidelines or systematic reviews. When using this approach, it is important to consider the literature search date indicated in the method section of the guidelines or reviews and not their publication date, since it can take a year or more for a guideline or review to get published. If the analyte has been well-established for several decades and is not under active research, the SVA may be also covered by medical textbooks.

The situation differs markedly for novel analytes or novel analyte-condition pairing. In such cases, often only very low-level evidence is available (e.g., animal studies, low quality / level clinical data) and manufacturers may need to conduct their own clinical study to support the SVA.

Another complication has now emerged with the rise of omics and machine-learning or artificial intelligence IVDs, which involve hundreds or even millions of analytes and conditions, to the point that it becomes unfeasible to consider the scientific validity of each analyte and condition separately. Such devices often rely on databases where a curation process evaluates the strength of the analyte-condition association. A transparent and systematic curation process with clear validity rating thus supports the SVA. Some of these IVDs have also been tested to evaluate their impact on patient management and outcomes, e.g., how precision oncology using whole-exome sequencing can impact clinical care of cancer patients.

More Than Compliance: The Strategic Value of Scientific Validity

Scientific validity should not be assumed, even for well-known analytes. While manufacturers are typically experts in their respective fields, maintaining objectivity can be challenging, particularly when a strong commitment to an innovative concept drives development effort. Limited time and high workloads may also constrain the ability to remain current with the rapidly evolving scientific literature. However, the process of establishing scientific validity is not a mere regulatory formality. A structured, evidence-based assessment of the analyte's association with the intended clinical condition or physiological state can provide critical insights during the development of the device. It enables a more precise understanding of the target condition, patient population, and relevant biomarkers or technologies. This, in turn, supports clearer risk identification and a more accurate definition of the intended purpose and may even reveal opportunities to expand the device's application to additional clinical contexts, analytes, or specimen types.

Do you have any questions about this article or do you need support? Contact us for a non-binding consultation.

Anna Neodo, PhD
Senior Clinical & Scientific Affairs Manager - E-Mail

^[1] Clinical guidelines, HTAs, textbooks, and systematic reviews are listed here in no particular order, because most or all of them can be available for a well-established analyte and can be published in any order.

^[2] Such studies can include randomized controlled trials (RCT) or non-randomized interventional studies.